## Preface and Executive Summary — IMF TA mission to NRBT (January 27–February 7, 2025)

## Source details

**Canonical URL:** [Preface and Executive Summary — IMF TA mission to NRBT (January 27–February 7, 2025)](https://www.imf.org/-/media/files/publications/tar/2025/english/tarea2025065-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/tar/2025/english/tarea2025065-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/tar/2025/english/tarea2025065-print-pdf.pdf.json)

---

### Mission purpose, team, and scope
- At the request of the National Reserve Bank of Tonga (NRBT), MCM conducted a virtual TA mission from January 27–February 7, 2025 to assist modernization of the NRBT monetary policy framework and operations.
- Mission team: Roger McLeod (MCM), Meguy Kuete Ngougning (MCM), Thiago Christiano Silva (MCM), Bernard J. Laurens (Short-term Expert).
- Core focus: implement a more market-based monetary operational framework by:
  - developing the appropriate mix of reserve requirements and effective monetary operations;
  - improving collateral policy for monetary policy operations;
  - comprehensively strengthening NRBT communications.
- Virtual follow-up TA will be provided to assist implementation.

### Context and core problem
- Exchange rate stability is the de jure and de facto primary objective; the pa’anga is pegged to a basket of currencies.
- Persistently high excess liquidity (from unsterilized FX reserve accumulation) impedes effective monetary policy transmission.
- The interbank market has been inactive for over a decade.
- Key operational components for market-based monetary policy are absent.

### Strategic approach and mission emphasis
- Two complementary modernization workstreams recommended:
  - (i) develop analytical capacity to assess monetary policy stance;
  - (ii) develop operational capacity to align short-term money market rates with the desired stance.
- Mission focused on operational capacity (workstream ii) to set conditions for emergence of an interbank market and enable NRBT to "be in the driver’s seat".

### Key operational recommendations (high level)
- Shift to a conventional Interest Rate Corridor (IRC) with the policy rate at the center (mid-corridor framework).
- Activate regular open market operations (OMOs) — weekly seven-day Main Liquidity Management Operations (MLMOs) issuing NRBT Notes at the policy rate — to align short-term interbank rates with the policy rate.
- If liquidity becomes deficit, auction seven-day repos properly collateralized.
- Standing facilities (marginal lending facility—MLF; marginal deposit facility—MDF) should be freely accessible to banks on an overnight basis; MLF must be properly collateralized.

### Sequencing and tender design
- Initial operation of MLMO through fixed rate, full allotment tenders given current capacity constraints.
- Move to variable rates, fixed amount tenders once forecasting and market understanding improve.
- Adopt a wide IRC at the start to stimulate interbank trading; engage banks to identify remaining obstacles (information availability, legal framework for collateralized lending).

### Liquidity monitoring and Treasury coordination
- Develop a weekly short-term liquidity monitoring and forecasting framework based on daily NRBT balance of accounts.
- NRBT should recommend to Treasury:
  - stop placing time deposits with commercial banks;
  - engage in a study to assess feasibility of a Treasury Single Account (TSA).
- NRBT should engage Treasury for regular (weekly) exchange of short-term cash flow forecasts.
- Operationalize an Emergency Liquidity Assistance (ELA) framework.

### Role and design of the SRD
- Reorganize SRD to support the IRC: SRD to serve exclusively to adjust structural liquidity; policy rate signals monetary stance.
- Technical adjustments to consider:
  - (i) consider some remuneration of the SRD;
  - (ii) consider allowing banks to meet the SRD on average over the reserve maintenance period (RMP) at an advanced stage;
  - (iii) link SRD non-compliance penalty to the policy rate.
- Two-stage sequencing:
  - Stage 1: mid-rate IRC with fixed rate and full allotment tenders for OMOs.
  - Stage 2: variable-rate, fixed allotment OMOs; SRD averaging over RMP; fine-tuning OMOs.

### Collateral and communications
- Initial restriction of collateral eligibility to public sector securities appropriate.
- Remove preference order among monetary policy collateral; develop broader collateral policy: eligibility, valuation techniques, haircuts capturing market, liquidity, default, and maturity risks.
- Communications: assessed institutional framework using NLP; recommended institutional policy and measures to improve readability, focus, and forward-looking content.

---

### Key Findings and Analytical Inputs

### NRBT objectives, peg, and FX regime
- De jure and de facto objective: maintain internal and external price stability; internal price stability: inflation below a reference rate currently set at five percent.
- Pa’anga pegged to a basket: US dollar, New Zealand dollar, Fijian dollar, Australian dollar; ±five percent monthly adjustment limit; US dollar used as intervention currency.
- NRBT attempts to maintain foreign reserves above three to four months of imports cover.
- Capital flow management and exchange controls in place; remittances are largest capital flow component.

### NRBT balance sheet and liquidity dynamics (selected figures in percent of GDP)
- Total Assets: 78.94 (2023), 41.78 (2013), 18.79 (2003)
- Foreign Assets — Short/Long Term Investments: 69.77 (2023), 33.88 (2013), 7.02 (2003)
- Foreign Assets — Accrued Interest: 1.61 (2023), 0.74 (2013), - (2003)
- Foreign Assets — IMF: 0.89 (2023), 0.66 (2013), 1.16 (2003)
- Currency in circulation: 9.67 (2023), 5.46 (2013), 3.83 (2003)
- Statutory Reserve Deposits: 10.11 (2023), 2.30 (2013), 4.75 (2003)
- Demand Deposits: 47.66 (2023), 26.52 (2013), 8.75 (2003)
- Payable to Government: 1.17 (2023), 0.39 (2013), - (2003)
- Other Liabilities: 5.59 (2023), 4.65 (2013), 0.28 (2003)
- Paid up Capital: 0.41 (2023), 0.71 (2013), 0.23 (2003)
- General Reserves: 1.86 (2023), 1.42 (2013), 0.29 (2003)
- Revaluation Reserve Account: 0.42 (2023), 0.34 (2013), 0.67 (2003)

### Monetary policy stance and instruments
- NRBT policy rate de facto: zero percent (floor since July 2016).
- Board resolutions (April 19, 2024): NRBT Policy Rate: zero percent; NRBT Notes Rate: two percent; Repo Rate: fixed at four percent.
- Current IRC elements (Table 3 summary preserved):
  - Repo Operations: 4% interest rate; seven-day or longer maturity; collateral: NRBT Notes; Government of Tonga bills and bonds.
  - NRBT Notes: 2% interest rate; issuance at NRBT discretion; undisclosed to market; no issuance so far.
  - Deposits with NRBT: 0% interest rate; at banks’ discretion.

### SRD history and design features
- SRD ratio timeline:
  - 15 percent (pre-2006);
  - 12.5 percent effective March 3, 2006;
  - 10 percent effective April 2, 2007;
  - five percent effective August 4, 2009;
  - 10 percent effective July 2017;
  - 15 percent effective April 2023.
- SRD measured monthly at month end; no averaging; all SRD in local currency.

### FX composite model and basket optimization
- Currencies considered: USD, AUD, NZD, FJD, CNY.
- FX composite model used monthly data from 2006 to 2024; optimized weights produced a more stable currency-basket index over the sample period relative to current weights.
- Pass-through reduced-form equation and variance-minimization objective applied (equations and optimization constraints preserved in source).

---

### Operational Recommendations — IRC, OMOs, SRD, and Liquidity Forecasting

### Interest Rate Corridor (IRC) and Main Liquidity Management Operation (MLMO)
- Recommendation 2 (NRBT): Introduce a conventional mid-rate IRC with fixed rate and full allotment tenders. ¶57
  - Announce a policy rate to communicate monetary stance and target short-term interbank market rates. ¶57
  - Commit to use OMOs (auctions of NRBT Notes) to steer interbank rates. ¶57
  - Standing facilities (MLF, MDF) pre-set; banks can borrow or deposit at NRBT at initiative. ¶57
- Specific operational design (Table 7 preserved):
  - Main liquidity management operation-MLMO: Issuance of NRBT Notes / Repo operations — Seven-day / Weekly / Fixed-rate, full allotment or variable rate fixed amount tenders.
  - Fine-tuning OMOs: Overnight deposits/repo operations — Overnight / As needed / Policy rate serves as maximum/minimum bid rate.
  - MLF: Overnight repo operations — Overnight / IRC Upper bound (policy rate + 100/200 pbs). Collateral: NRBT Notes, Government of Tonga bills and bonds.
  - MDF: Overnight deposits — Overnight / IRC Lower bound (policy rate – 100/200 pbs).

- Operational sequencing:
  - Initially operate weekly MLMO through fixed-rate, full allotment tenders (accept all bids; volume determined by banks) given limited forecasting capacity.
  - Once liquidity monitoring and forecasting capacity improves, transition to variable-rate, fixed-amount tenders. ¶62

### Standing facilities and corridor width
- SFs (MDF and MLF) must be freely accessible daily through overnight operations; MLF access contingent on collateral availability.
- NRBT must not limit MDF volumes to avoid distorting interbank rates.
- Recommendation 5: Adopt a wide IRC at the start to encourage interbank trading and discourage overuse of SFs. ¶65

### Short-term liquidity monitoring and Treasury coordination
- Recommendation 6 (NRBT): Strengthen liquidity monitoring and forecasting by developing capacity to analyze and forecast the autonomous factors of liquidity on a weekly basis. ¶68
- Recommendation 7 (NRBT): NRBT should recommend to Treasury to stop placing time deposits in commercial banks and engage in a study to assess feasibility of a TSA framework. ¶73
  - Treasury account snapshot:
    - One time deposit in a commercial bank for an amount of 41 million.
    - Aggregated balances of Treasury current accounts in commercial banks about 100 million.
    - Treasury accounts with NRBT (including project accounts and main operating account) represent about 175 million.
- Recommendation 9: Engage in discussions with Treasury for exchange of cash flow forecasts; currently Treasury does not provide short-term cash flow forecasts to NRBT, though weekly information exists and Treasury officials expressed willingness to share. ¶75
- Recommendation 8: Develop operational guidelines for activation of an ELA framework. ¶74

### SRD redesign and administration
- Recommendation 10 (NRBT): Rely on SRD to adjust structural liquidity position.
  - Consider remunerating the SRD (ideally close to policy rate; in short-term a lower rate or no remuneration acceptable given large excess liquidity). ¶81
  - Consider allowing SRD averaging over RMP at advanced stage. ¶85
  - Link SRD non-compliance penalty to NRBT Policy Rate and set penalty above MLF but avoid excessive penalties that cause hoarding. ¶85
- Entrust SRD administration to NRBT unit in charge of monetary policy implementation; clear communication that SRD changes are structural, not policy stance signals.

---

### Collateral Framework, Valuation, and Haircuts

### Current framework and gaps
- Legal provisions (Section 38; Section 48A) empower NRBT to extend loans and repurchase government securities but are less explicit on broader asset acceptance and mobilization modalities.
- Current repo eligibility (preference order): NRBT Notes, Tonga Government Bonds, Tonga Government Treasury Bills. Valuation practices: NRBT notes at discount; T-bills at settlement price; coupon-paying securities at face value.
- Gaps: No haircuts, concentration limits, or overcollateralization; no collateral policy for ELA; mobilization procedures not standardized; limited secondary market pricing.

### Valuation methodology and recommended approach
- Collateral should be marked to market; in absence of market prices, develop theoretical pricing using issuance-implied yields and interpolation.
- Use issuance-implied yields and NRBT Note rate to interpolate yields; price bonds as sum of discounted cash flows; Nelson-Siegel methodology recommended to calibrate term structure curves and discount factors.
- Government bond issuance profile (2024):
  - Average maturity at issuance: five years.
  - Maturities issued in 2024: two bonds with maturity of three years, one with maturity of four years, and two with maturity of five years.
  - Implied yields to maturity for 2024: 2.5 percent, 2.5 percent, and three percent for maturities of three, four, and five years respectively.

### Haircut calibration and risk equivalence
- Haircuts calibrated to achieve risk equivalence across asset classes; key components: liquidity risk (time-to-liquidation), market risk (price volatility), credit risk (default probability), and additional risks (exchange-rate, valuation, residual maturity, mobilization).
- Where market pricing is sparse, use relative assessment of standalone risk components and add-ons.
- Box F.1 illustrative VaR-based example yields haircuts: { 2.33%; 6.60% } for two hypothetical assets under prescribed parameter values.
- Recommendation 11: Remove preference order for monetary policy collateral and standardize mobilization procedures for assets. ¶97
- Recommendation 12: Develop sound valuation methodologies to mark collateral to market. ¶98–¶99
- Recommendation 13: Coordinate with Government to issue bonds at maturities that help calibrate a term-structure sovereign yield curve. ¶100
- Recommendation 14: Develop and implement risk mitigation measures, including granular haircuts capturing market, liquidity, and credit risk. ¶97

---

### Communications — Assessment, Metrics, and Recommendations

### Institutional gaps and current instruments
- NRBT publications and frequencies (Table 10 summary preserved): Monetary Policy Decision/Statement (Semiannual), Press Release (Semiannual), Annual Report (Annual), Quarterly Bulletin (Quarterly), Monthly Economic Update (Monthly), Monetary Policy Data (Monthly), other periodic releases.
- Gaps:
  - No institutional communications policy documenting objectives, messages, audiences, channels, spokespersons, impact assessment, or crisis communication.
  - No Financial Stability Report in production (one being developed).
  - Communications function understaffed; one communications officer in Risk and Compliance performing media liaison tasks; recommendation to create Communications division.

### Textual and NLP-based assessment — key empirical findings
- Lexical/readability:
  - Monetary policy decisions have shortened since 2019 but readability declined since 2021; by 2024 monetary policy decisions reached a readability score below 40 (classification: “difficult”).
- Syntactical:
  - Sentence structures simpler than peer average but have become more convoluted since 2022.
- Semantical and topic analysis:
  - Monetary policy is the most prevalent topic over 25 years; Tonga’s monetary policy decisions discuss more topics than peers, leading to diluted messages.
  - Forward-lookingness was low until 2022; improvement needed to anticipate external shocks.
- Textual metrics:
  - Net Confidence Index: communication leaned toward confidence-building after 2014; rise from 2014 to 2017; remains confidence-building but to a lesser degree recently.
  - Straightforwardness Index: generally unidirectional signals but declined after 2022 due to mixed messages.
  - Explanation Index: Tonga’s monetary policy decisions devote five explanatory sentences per stance-based sentence.

### Communication recommendations (priority and actions)
- Recommendation 15: Develop a formalized institutional policy for communications with complementary frameworks (objectives, principles, spokespersons, messages, audiences, channels, tools, formats, effectiveness assessment). ¶109–¶110
- Recommendation 16: Develop a Financial Stability Report. ¶111
- Recommendation 17: Clearly communicate, following international best practice, the implementation of the proposed monetary operational framework. ¶116
  - Publicly explain stage-wise measures to absorb excess liquidity and induce interbank lending to steer interbank rates toward the policy rate.
- Recommendation 18: Expand outreach and adopt multitiered communication tailored to audience technical depth. ¶117–¶118
- Recommendation 19: Enhance effectiveness by:
  - (i) improving readability (reduce wording complexity);
  - (ii) producing more focused monetary policy statements (reduce topic diversification);
  - (iii) integrating more forward-looking elements. ¶135

### Implementation support and organizational changes
- Key complementary frameworks:
  - Develop a media monitoring function and database of communications activities and owned channels.
  - Implement periodic surveys to assess trust, mandate awareness, inflation target awareness, and confidence in the central bank.
  - Create a Communications division (priority to add one staff to existing officer to form a two-person division).
  - Conduct crisis communication simulations.
- Monitoring: develop regular assessment tool for communications effectiveness using media monitoring and NLP techniques.

---

### Implementation Stages, TA Program, and Expected Outcomes

### Two-stage implementation approach (summary)
- Stage 1 (priority actions):
  - Transition to new IRC framework based on weekly MLMO: weekly issuance of seven-day NRBT Notes at policy rate via fixed-rate, full allotment tenders.
  - Consider increasing SRD ratio and some SRD remuneration.
  - Link SRD noncompliance penalty to policy rate.
  - Develop capacity to assess policy rate level; start weekly liquidity monitoring and forecasting; rely on current collateral framework with haircuts; develop communication strategy.
  - SRD averaging over RMP generally excluded at this stage.
- Stage 2 (subsequent actions):
  - Weekly issuance of NRBT Notes at variable rate, fixed allotment tenders.
  - Introduce averaging provisions over RMP.
  - Remunerate SRD at opportunity cost.
  - Liquidity forecasting framework fully operational.
  - Implement modernized collateral framework and fine-tuning OMOs.
  - Communication strategy fully operational.

### Proposed TA program
- Virtual TA from 2025 Q2 to 2026 Q2 to:
  - Set up a short-term liquidity monitoring and forecasting framework and estimate demand for reserves.
  - Develop collateral framework for liquidity-providing monetary operations (valuation, haircut, margin requirements).
- Follow-up virtual TA for implementation support and calibration of OMOs and collateral arrangements.

### Expected outcomes if implemented
- By end of Stage 1: NRBT transitioned to new IRC framework based on regular MLMO through fixed-rate, full-allotment tenders; improved liquidity forecasting capacity anchored on monitoring autonomous factors; predictable market-based placement of surplus liquidity and short-term funding among banks.
- Completion of Stage 2: move to variable-rate, fixed-amount MLMOs, RMP averaging, fine-tuning OMOs, remunerated SRD at opportunity cost, and a modern collateral framework that strengthens monetary operations and transmission.

---

*Italic: Preface and Executive Summary, IMF Technical Assistance Report — Monetary and Capital Markets Department TA mission to NRBT (January 27–February 7, 2025).*

### Preface ................................................................................................................

### Preface

### Mission purpose and team
- At the request of the National Reserve Bank of Tonga (NRBT), a Monetary and Capital Markets (MCM) Department conducted a virtual Technical Assistance (TA) mission from January 27–February 7, 2025, to assist the authorities in modernizing the monetary policy framework and operations of the NRBT.
- Mission team:
  - Roger McLeod (MCM), mission leader
  - Meguy Kuete Ngougning (MCM)
  - Thiago Christiano Silva (MCM)
  - Bernard J. Laurens (Short-term Expert)
- Purpose: assist with implementing a modernized and more market-based monetary operational framework, focusing on:
  - developing the appropriate mix of reserve requirements and effective monetary operations to enable the NRBT to employ market-based mechanisms to influence market conditions and mitigate vulnerabilities to the peg;
  - improving the collateral policy for monetary policy operations;
  - comprehensively strengthening NRBT communications.
- Virtual follow-up TA will be provided to assist implementation.

### Stakeholder engagement and review
- Mission meetings included NRBT officials:
  - Tatafu Moeaki, Governor;
  - Ms. Ungatea Latu, Deputy Governor;
  - Daniel Taumoepeau, Assistant Governor Policy;
  - Mrs. Mefilina Tohi, Chief Manager Economics and Research;
  - Ms. Nanuma Fakatava, Chief Manager Financial Markets;
  - Mrs. Kueni Fifita, Chief Manager Financial Systems;
  - Mrs. Vasi Taufaloa, Senior Manager Finance.
- The mission also met with representatives from the NRBT Board, the treasury department of the Government of Tonga, and several commercial banks.
- The report incorporates comments from the NRBT and those received through the IMF’s internal review process.

### Executive Summary

### Context and core problem
- Exchange rate stability is the de jure and de facto primary objective of the NRBT with the Tongan pa’anga pegged to a basket of currencies.
- The NRBT employs capital and exchange controls to support this regime.
- Persistently high levels of excess liquidity (emanating from unsterilized FX reserve accumulation) continue to impede effective monetary policy transmission.
- The interbank market has remained inactive for over a decade.
- Key components for an effective operational framework are still absent.

### Mission focus (building blocks)
- The mission focused on:
  - (i) the design of the currency basket;
  - (ii) the Interest Rate Corridor (IRC);
  - (iii) the design of the Statutory Reserve Deposit (SRD) instrument;
  - (iv) the design of a short-term liquidity monitoring and forecasting framework;
  - (v) the collateral policies to support liquidity-providing monetary operations;
  - (vi) NRBT communications.

### Strategic approach
- The mission recommends the NRBT start a gradual process of monetary policy modernization via two complementary workstreams:
  - (i) develop analytical capacity to assess monetary policy stance;
  - (ii) develop operational capacity to determine and assess actions in the money market to align short-term money market rates with the desired monetary policy stance.
- The mission focused on operational capacity (workstream ii) to set conditions for the emergence of an interbank market, enabling the NRBT to "be in the driver’s seat" in setting short-term money market interest rates through liquidity management.

### Key operational recommendations
- Shift to a conventional Interest Rate Corridor (IRC) with the policy rate at the center (mid-corridor framework).
- Activate regular open market operations (OMOs) to align short-term interbank market interest rates with the policy rate.
- Rely on Main Liquidity Management Operations (MLMOs) to mop up excess liquidity via auctioning seven-day NBRT Notes.
- If banking system liquidity becomes deficit, auction seven-day repos properly collateralized.
- Standing facilities (marginal lending facility—MLF; marginal deposit facility—MDF) should be freely accessible to banks and operated on an overnight basis; the MLF must be properly collateralized.

### Tender design and sequencing
- Given current NRBT limitations (monthly liquidity forecasts, data management gaps), the MLMO should initially be operated through fixed rate, full allotment tenders.
- As capacity improves and market participants understand the framework, consider moving to:
  - variable rates, fixed amount tenders;
  - fine-tuning OMOs.
- To stimulate interbank trading, adopt a wide IRC at the start.
- Engage banks to identify remaining obstacles to interbank trading (e.g., information availability, legal framework for collateralized lending).

### Liquidity monitoring and Treasury coordination
- Develop a weekly short-term liquidity monitoring and forecasting framework based on daily NRBT balance of accounts.
- NRBT should recommend to the Treasury:
  - stop placing time deposits with commercial banks;
  - engage in a study to assess the feasibility of setting up a Treasury Single Account (TSA).
- NRBT should engage the Treasury for regular (weekly) exchange of short-term cash flow forecasts.
- Operationalize an Emergency Liquidity Assistance (ELA) framework.

### Role and design of the SRD
- Reorganize the SRD to support the IRC framework: SRD to serve exclusively to adjust structural liquidity, while policy rate changes signal monetary policy stance.
- Technical adjustments to consider:
  - (i) consider some remuneration of the SRD;
  - (ii) consider allowing banks to meet the SRD on average over the reserve maintenance period (RMP) at an advanced stage of modernization;
  - (iii) link SRD non-compliance penalty to the policy rate.
- Two-stage sequencing recommended:
  - Stage 1: Introduce mid-rate IRC with fixed rate and full allotment tenders for OMOs.
  - Stage 2: Once robust weekly liquidity monitoring and forecasting capacity exists, move to variable rate, fixed allotment OMOs, introduce SRD averaging over the RMP, and fine-tuning OMOs.

### Collateral framework
- Current restriction of collateral eligibility to public sector securities is appropriate initially.
- Remove the current preference order among monetary policy collateral (preference order advisable only for discretionary operations such as ELA).
- Continuously monitor counterparties’ balance sheets.
- Develop broader collateral policy including:
  - eligibility criteria;
  - sound collateral valuation techniques (in absence of active liquid securities market, invest in theoretical pricing using implied yields from bond issuances);
  - risk mitigation measures and haircuts capturing market, liquidity, default, and maturity risks.
- Provide considerations for a future ELA collateral framework.

### Communications assessment and recommendations
- Assessed adequacy of the institutional framework for NRBT communications and applied Natural Language Processing (NLP) across lexical, syntactical, and semantical dimensions, plus comparative regional insights.
- Recommendations included enhancements to institutional framework and specific measures to improve communication effectiveness.

### Recommendations (Key items and implementation authorities)
- Recommendation 1 (NRBT): Consider a gradual modernization of monetary policy implementation by setting up an active liquidity management framework to allow the NRBT to steer short-term money market rates to its announced policy rate. ¶48
  - Priority: High
  - Timeframe: Near-term
- Recommendation 2 (NRBT): Start monetary policy modernization by Introducing a conventional mid-rate interest rate corridor (IRC) with fixed rate and full allotment tenders. ¶57
  - Priority: High
  - Timeframe: Near-term
- Recommendation 3 (NRBT): Move to mid-rate IRC with variable rates and fixed allotment tenders once a liquidity monitoring and forecasting framework is fully operational. ¶62
  - Priority: Medium
  - Timeframe: Medium-term
- Recommendation 4 (NRBT): Engage discussions with the banks to assess what specific actions the NRBT could take to support interbank market activities. ¶64
  - Priority: High
  - Timeframe: Medium-term
- Recommendation 5 (NRBT): Adopt a wide IRC at the start. ¶65
  - Priority: High
  - Timeframe: Near-term
- Recommendation 6 (NRBT): Strengthen liquidity monitoring and forecasting by developing a capacity to analyze and forecast the autonomous factors of liquidity (Afs) on a weekly basis. ¶68
  - Priority: High
  - Timeframe: Medium-term
- Recommendation 7 (NRBT): NRBT should recommend to the Treasury to stop placing time deposits in the commercial banks, and to engage in a study to assess the feasibility of setting a Treasury Single Account (TSA) framework. ¶73
  - Priority: High
  - Timeframe: Near-term
- Recommendation 8 (NRBT): Develop operational guidelines for the activation of an ELA framework. ¶74
  - Priority: High
  - Timeframe: Medium-term
- Recommendation 9 (NRBT): Engage in discussions with the Treasury for the exchange of cash flow forecasts. ¶75
  - Priority: High
  - Timeframe: Near-term
- Recommendation 10 (NRBT): Rely on the SRD to adjust the structural liquidity position:
  - Consider remunerating the SRD. ¶81
  - Consider introducing averaging over the reserve maintenance period and linking noncompliance penalty to the policy rate. ¶85
  - Priority: High
  - Timeframe: Near and Medium-term
- Recommendation 11 (NRBT): Remove the preference order for monetary policy collateral and undertake necessary actions to standardize the mobilization procedures for assets. ¶97
  - Priority: Medium
  - Timeframe: Medium-term
- Recommendation 12 (NRBT): Develop sound valuation methodologies to mark collateral to market. ¶98–¶99
  - Priority: High
  - Timeframe: Medium-term
- Recommendation 13 (NRBT): Coordinate with the Government of Tonga to issue bonds (consistent with the Government's financing priorities) at maturities that could help calibrate a term-structure sovereign yield curve. ¶100
  - Priority: Medium
  - Timeframe: Medium-term
- Recommendation 14 (NRBT): Develop and implement risk mitigation measures, including haircuts that capture key financial risks such as market, liquidity, and credit risk. ¶97
  - Priority: High
  - Timeframe: Medium-term
- Recommendation 15 (NRBT): Develop a formalized institutional policy for communications along with key complementary frameworks. ¶109–¶110
  - Priority: High
  - Timeframe: Medium-term
- Recommendation 16 (NRBT): Develop a Financial Stability Report. ¶111
  - Priority: High
  - Timeframe: Medium-term
- Recommendation 17 (NRBT): Clearly communicate, following international best practice, the implementation of the proposed monetary operational framework. ¶116
  - Priority: High
  - Timeframe: Near-term
- Recommendation 18 (NRBT): Expand outreach and adopt multitiered communication for monetary policy communication. ¶117–¶118
  - Priority: Medium
  - Timeframe: Medium-term
- Recommendation 19 (NRBT): Enhance the effectiveness of communication by:
  - (i) improving readability indices of monetary policy decisions by reducing the wording complexity;
  - (ii) producing more focused monetary policy statements in terms of topics; and
  - (iii) integrating more forward-looking elements in monetary policy statements. ¶135
  - Priority: Medium
  - Timeframe: Medium-term

- Timeframe definitions:
  - Near term: < 12 months;
  - Medium term: 12 to 24 months.

*Italic: Preface and Executive Summary, IMF Technical Assistance Report — Monetary and Capital Markets Department TA mission to NRBT (January 27–February 7, 2025).*

### Introduction

### Introduction

### Monetary policy objectives and legal amendments
- De jure and de facto objective of the NRBT: maintain internal and external price stability.
- Internal price stability defined as maintenance of inflation below a reference rate (currently set at five percent).
- External price stability defined as stability of the Pa’anga (TOP).
- The Monetary Policy Statement (MPS) notes aims to: ensure an adequate level of foreign reserves and low and stable inflation, while supporting macroeconomic stability and economic growth.
- The NRBT has not formally announced a hierarchy of monetary policy objectives; in practice it has relied on exchange rate stability as a nominal anchor because of the small domestic production base and high proportion of imports.
- Amendments to the NRBT Act are being considered to:
  - Establish price stability as the core priority with the primary focus on low and stable inflation.
  - Establish financial stability as the second objective.
  - Develop provisions for the NRBT to proactively facilitate and nurture private sector growth.
- The NRBT Act 2020 (Revised Edition) states that the principal objectives of the Reserve Bank are to maintain internal and external monetary stability, and to promote financial stability and a sound and efficient financial system. The NRBT should conduct its activities in a manner that supports macroeconomic stability and economic growth.

### Currency peg, FX regime, and currency basket
- The Tongan pa’anga is pegged to a basket comprising the US dollar, New Zealand dollar, Fijian dollar, and the Australian dollar.
- There is a ±five percent monthly adjustment limit; the US dollar is used as an intervention currency.
- The NRBT determines weights of the currency basket with the primary objective of stabilizing the basket; design allows some exchange rate flexibility via currencies other than the US dollar.
- The NRBT attempts to maintain foreign reserves above three to four months of imports cover.
- The NRBT employs capital flow management measures that have curtailed capital mobility and vulnerabilities to the peg.
- Largest component of capital flows: remittances (insensitive to interest rate differentials).
- Largest institutional investors (two pension funds) are required to invest domestically.
- Exchange controls in place:
  - All foreign currency payments from pa’anga accounts are subject to exchange control requirements.
  - Banks must provide underlying documentation when purchasing FX from the NRBT.
  - Banks and authorized FEDs delegated approval authority for payments up to ceilings; NRBT prior approval required above ceilings.
  - Export proceeds subject to repatriation and surrender requirements.
  - Banks’ nostro account balances subject to a daily ceiling; any excess must be sold to the NRBT within two business days.

### NRBT balance sheet evolution and liquidity dynamics
- FX reserves are the main asset on the NRBT balance sheet; significant FX reserve accumulation has driven balance sheet growth over two decades.
- Excess liquidity resulted from unsterilized FX purchases; excess liquidity has remained unremunerated by the NRBT since 2012.
- De facto zero-interest rate monetary policy stance maintained since 2012; no trading activities in the interbank market.
- NRBT could issue NRBT notes to mop up excess liquidity but has not activated this instrument for the last 10 years; instead relied on changes in the Statutory Reserve Deposit (SRD) ratio.
- FX assets are the main contributor to structural liquidity position and strongly influence the autonomous factors (Afs).

Key balance sheet figures (percent of GDP) from NRBT balance sheet evolution table:
- Total Assets: 78.94 (2023), 41.78 (2013), 18.79 (2003)
- Foreign Assets — Short/Long Term Investments: 69.77 (2023), 33.88 (2013), 7.02 (2003)
- Foreign Assets — Accrued Interest: 1.61 (2023), 0.74 (2013), - (2003)
- Foreign Assets — IMF: 0.89 (2023), 0.66 (2013), 1.16 (2003)
- Foreign Assets — (other lines in table present as shown in source)
- Domestic Assets — Advance to Banks: - (2023), - (2013), 0.86 (2003)
- Domestic Assets — Claim on Government of Tonga: - (2023), - (2013), 3.93 (2003)
- Currency in circulation: 9.67 (2023), 5.46 (2013), 3.83 (2003)
- Statutory Reserve Deposits: 10.11 (2023), 2.30 (2013), 4.75 (2003)
- Demand Deposits: 47.66 (2023), 26.52 (2013), 8.75 (2003)
- Non-Bank Deposits: 2.05 (2023), - (2013), - (2003)
- Payable to Government: 1.17 (2023), 0.39 (2013), - (2003)
- Other Liabilities: 5.59 (2023), 4.65 (2013), 0.28 (2003)
- Paid up Capital: 0.41 (2023), 0.71 (2013), 0.23 (2003)
- General Reserves: 1.86 (2023), 1.42 (2013), 0.29 (2003)
- Revaluation Reserve Account: 0.42 (2023), 0.34 (2013), 0.67 (2003)

### Monetary policy instruments, Interest Rate Corridor (IRC), and SRD
- NRBT policy rate historically: zero percent (since July 2016 as floor).
- November 2016 recommendations: implement a floor IRC with policy rate initially set at zero; recommendation to tighten policy if overheating emerges was not implemented.
- November 2016 Board approvals (implemented then):
  - NRBT Policy Rate: zero percent (July 2016) — applied to standing deposit facility (interest on commercial banks’ Exchange Settlement Accounts).
  - NRBT Note Rate: two percent below banks’ average one month deposit rate.
  - Repo Rate: 3.9 percent (two percent above interbank lending rate).
- As of April 19, 2024 Board resolutions (Board Paper No. 25/2024):
  - NRBT Policy Rate (Interest on the banks’ Exchange Settlement Accounts): zero percent—floor rate.
  - NRBT Notes Rate: two percent—mid-rate.
  - Repo Rate: fixed at four percent as the ceiling rate.
  - Repo conditions remain the same except for revised interest rate policy; IRC to be reviewed annually.
- Current IRC structure (as per Table 3):
  - Repo Operations: 4% interest rate; seven-day or longer maturity as requested by banks; collateral: NRBT Notes; Government of Tonga bills and bonds.
  - NRBT Notes: 2% interest rate; maturity N/A; issuance at NRBT discretion; undisclosed to the market; no issuance so far.
  - Deposits with NRBT: 0% interest rate; maturity N/A; at banks’ discretion; interest rate is the NRBT Policy Rate.
- SRD history and design:
  - Pre-2006 SRD was 15 percent.
  - Reduced to 12.5 percent effective March 3, 2006.
  - Reduced to 10 percent effective April 2, 2007.
  - Reduced to five percent effective August 4, 2009.
  - Increased to 10 percent effective July 2017.
  - Increased to 15 percent effective April 2023.
- Box 1: SRD design — banks required to maintain unremunerated deposits with NRBT as a percentage of local and foreign currency denominated deposits measured at month end; maintenance period is monthly with no averaging; all SRD are in local currency; given low financial dollarization, maintaining SRD in local currency for FX deposits is appropriate.

### Operational gaps and analytical capacity
- Absent components for an effective operational framework:
  - No clear operational target.
  - No interest rate that would incentivize savings in Tongan pa’anga; interbank market inactive.
  - Missing components: liquidity forecasts, standing facilities, open market instruments, and supporting arrangements for decision-making.
- NRBT has not developed an appropriate analytical and forecasting framework to guide policy rate decisions or instrument choice given monetary policy cost considerations.
- Short-term liquidity forecasting capacity incomplete:
  - Forecasts conducted monthly using NRBT monthly balance sheet.
  - Forecast involves NFA and currency in circulation; bank-provided information used to assess reserve demand.
  - Discrepancies between NRBT forecasts and actual outcomes tend to be large.

### Use of unconventional and non-market instruments; past TA implementation
- NRBT responses to shocks have relied on unconventional tools and SRD changes; examples:
  - 2016: introduced a minimum loan to deposit ratio of 80 percent (targets never achieved; not enforced).
  - 2017: increased SRD from five to 10 percent in response to elevated inflation as a signaling device.
  - 2023: offered deposit facilities for Government and retirement funds at the NRBT for targeted mopping up of excess liquidity; retirement fund deposits discontinued in 2024 to support banks’ lending and liquidity.
  - Supported development of financial instruments: credit guarantee facility and concessional credit facility (still to be implemented).
  - Reduced spread of exchange rates that authorized dealers can offer to customers to make FX cheaper for importers and better for exporters.
  - Utilized exchange rate adjustments (limited to five percent monthly) during high inflation to cushion impacts of high import costs.
  - Consistent reliance on comprehensive exchange controls limiting arbitrage activity.
- A previous IMF TA mission completed in 2017 reviewed framework and operations; implementation of TA recommendations only partially achieved due to natural disaster shocks and the COVID-19 outbreak. Policy responses to shocks relied largely on non-market-based measures.

### TA request, objectives, and report structure
- Tongan authorities requested assistance in implementing a modernized monetary policy framework to improve monetary transmission and incentivize interbank trading.
- Rationale: Implementing a modernized approach with instruments to absorb excess liquidity is crucial to restart the interbank market and allow NRBT to steer short-term interbank rates towards the policy rate.
- Structure of the report (subsequent sections as described):
  - Section I: overview of monetary policy implementation over the last ten years.
  - Section II: review of FX operation and determination of optimal FX basket weights.
  - Section III: design of an interest rate corridor (IRC) supported by an appropriate SRD instrument to absorb excess liquidity and align interbank rates with the NRBT policy rate.
  - Section IV: discussion of collateral policy and recommendations for estimating collateral haircuts.
  - Section V: examination of NRBT communication using natural language processing (NLP) and recommendations on a modernized communication policy framework.

*Source: tarea2025065-print-pdf - Introduction*

### 22. For its foreign exchange interventions (FXI) the NRBT relies on a foreign exchange

### 22. For its foreign exchange interventions (FXI) the NRBT relies on a foreign exchange 

### FX standing facility and exchange control procedures
- Licensed commercial banks can purchase FX from the NRBT to cover their bona fide transactions with customers via a foreign exchange standing facility.
- The NRBT sets the US dollar–pa’anga rate on a daily basis and buys and sells foreign exchange at these exchange rates to the commercial banks.
- Commercial banks are required to list all the underlying customers, and all foreign currency payments from pa’anga accounts are subject to the exchange control requirements; supporting documents are provided accordingly to verify whether the transactions are bona fide.

### FX interbank market and nostro limits
- Licensed commercial banks may trade among themselves in the FX interbank market.
- There are no limits on the spreads for interbank foreign exchange transactions among commercial banks.
- The NRBT does not intervene directly with market participants at their quoted rates.
- Commercial banks are allowed to maintain nostro accounts, but their total balances may not exceed T$1 million daily. Any excess must be sold to the NRBT within two business days.

### Historical exchange-rate dynamics and basket protection
- The pa’anga currency has depreciated relative to the USD over the long term since 2006 with a period of appreciation from 2008 to 2012.
- The long-term movement in the USD/TOP exchange rate has been somewhat countered by long-term trends in other exchange rates, most notably since 2020.
- AUD/TOP and NZD/TOP exchange rates have countered the impact of USD/TOP movements on the stability of the basket, providing the NRBT with some protection against the basket’s exposure to USD/TOP.

### Potential liberalization
- Once the NRBT has modernized its monetary operational framework along the lines recommended by the mission, consideration may be given to gradually liberalizing the FX regime.
- Overriding objective: gradually shift from reliance on FX controls (to limit residents’ ability to undertake arbitrage operations) towards a framework that relies more directly on market-based incentives.

### Currency Basket Design — current objective
- Official objective of the currency basket: maintain low and stable inflation in line with the primary NRBT mandate of price stability.
- In the absence of operational targets for inflation, exchange rate stability brought about by the currency basket serves as the main tool to maintain price stability.

### Import shares and trade invoicing (selected statistics)
- Tongan imports are predominantly sourced from 15 countries, accounting for around 92 percent of its total import.
- Example import share movements:
  - China increased from 1 percent in 1998 to 9 percent in 2022, and to 14.1 percent in 2024.
  - Import shares from Fiji have fluctuated between 1 and 10 percent since 1998.
  - Shares for the US, New Zealand and Australia have been more stable.
- 2024 data is as of March 2024.
- Table 5 (Trade Data by Country and Currency, 1998 / 2010 / 2024 / 2024 invoicing) — selected entries:
  - Australia: 11.16 (1998), 6.54 (2010), 7.97 (2024), 7.5 (AUD)
  - China: 0.32 (1998), 3.56 (2010), 7.05 (2024), 2.0 (CNY)
  - Fiji: 10.38 (1998), 5.58 (2010), 6.36 (2024), 5.5 (FJD)
  - New Zealand: 24.56 (1998), 24.17 (2010), 25.96 (2024), 25.0 (NZD)
  - United States: 22.27 (1998), 14.55 (2010), 18.76 (2024), 60.0 (USD)
- Information on the current weights applied by NRBT is not disclosed to the public.

### Basket volatility and currency relative volatility (Table 6 summary)
- The currency basket index has displayed low volatility over the review period.
- USD has been the most stable currency relative to the TOP over the past two decades; FJD has been the most volatile; AUD and NZD displayed moderate volatility.
- Selected descriptive statistics (exact values from Table 6):
  - Mean: TOP/USD = -0.003; TOP/NZD = 0.000; TOP/AUD = 0.003; TOP/CNY = -0.005; TOP/FJD = 0.011; Current Basket Index = 0.002
  - Median: TOP/USD = 0.000; TOP/NZD = -0.014; TOP/AUD = -0.015; TOP/CNY = 0.000; TOP/FJD = 0.000; Current Basket Index = -0.003
  - Max: TOP/USD = 2.789; TOP/NZD = 3.833; TOP/AUD = 5.157; TOP/CNY = 2.686; TOP/FJD = 24.955; Current Basket Index = 3.296
  - Range: TOP/USD = 5.714; TOP/NZD = 6.980; TOP/AUD = 9.021; TOP/CNY = 5.598; TOP/FJD = 27.850; Current Basket Index = 5.982
  - Mean Abs Dev.: TOP/USD = 0.361; TOP/NZD = 0.428; TOP/AUD = 0.417; TOP/CNY = 0.350; TOP/FJD = 0.451; Current Basket Index = 0.293
  - Variance (n): TOP/USD = 0.258; TOP/NZD = 0.371; TOP/AUD = 0.415; TOP/CNY = 0.244; TOP/FJD = 0.803; Current Basket Index = 0.196
  - Standard Deviation (n): TOP/USD = 0.508; TOP/NZD = 0.609; TOP/AUD = 0.644; TOP/CNY = 0.494; TOP/FJD = 0.896; Current Basket Index = 0.443
  - Skewness: TOP/USD = -0.348; TOP/NZD = 0.546; TOP/AUD = 1.255; TOP/CNY = -0.271; TOP/FJD = 14.853; Current Basket Index = 0.684

### Methodology for determining optimal weights
- Objective: choose weights to stabilize the basket and minimize fluctuations of import prices.
- FX composite model (developed by the Monetary Operations Division of MCM (MCMCO)) used to determine optimal weights.
- Pass-through reduced-form equation (as specified):
  - ∆Log(π_t) = ∑_{j=1}^N β_j ∆log(e_{j,t−p}) + ε_t
  - Where: π_t = import price index at time t; β_j = pass-through coefficient; e_{j,t} = bilateral exchange at time t; p = number of lag; N = number of trade partners; ∆ = difference operator; ε_t = error term.
- Import-price volatility expressed as variance:
  - σ_π^2 = F(w,.) = ∑(β_j w_j)^2 σ_j^2 + 2∑_{i≠j} β_i β_j w_i w_j cov(e_i,e_j)
  - Variance of e_{j,t} = σ_j^2; w_j is weight assigned to each exchange rate.
- Optimal weights obtained by solving a minimum variance portfolio optimization:
  - Minimize w′Zw subject to:
    - ∑_j w_j = 1
    - 0 ≤ w_j ≤ 1
    - F(w,.) ≤ σ̅^2
  - Z is the variance-covariance matrix of bilateral exchange rates; σ̅^2 is a desirable inflation variance (Annex IV provides details on choice of σ̅^2).

### Currencies considered and rationale
- Based on Tonga’s current currency framework, five currencies are considered: USD, AUD, NZD, FJD, and CNY.
- Estimated trade invoicing data indicates all five currencies have a non-negligible proportion in invoicing trade, with USD and NZD having the largest impacts.
- CNY included to assess (i) whether it helps stabilize the basket and (ii) the optimal weight of its inclusion.

### Results of the FX composite model
- The mission estimated the FX composite model and optimized weights using monthly data from 2006 to 2024.
- The currency basket index produced by the optimized weights demonstrated greater stability over the sample period relative to the index produced by the current currency basket weights.
- The mission discussed detailed results, including the optimal weights for all currencies, with the authorities.
- Tools delivered:
  - Technical presentation on model methodology.
  - Excel-friendly tool for NRBT staff to estimate the FX composite model (model programmed using Python).

### Monetary policy conduct — conceptual considerations and recommendations
- Key conceptual points:
  - Short-term interest rate and exchange rate are tied by an interest rate parity condition when there is some de facto capital mobility; they should be managed coherently to achieve final monetary policy objectives.
  - Inflation can be defined as the final objective using an exchange rate anchor, while the interest rate remains the monetary policy instrument.
  - Exchange rate anchors are more prone to policy inconsistencies, calling for fiscal and monetary discipline; they are hardly compatible with fiscal dominance and conventional monetary targeting.
- Liquidity management lessons and operational elements:
  - Liquidity management stabilizes money market rates, influences domestic financial conditions, and anchors market expectations; framework should curtail liquidity surpluses and limit over-injection of central bank liquidity.
  - Common features under peg regimes include: reserve requirements (RR) with ~one-month averaging, short-term open market operations (OMOs), interest rate corridors, and use of interest rates to offset FX flows.
  - Operational objectives achievable with RR, OMOs, and standing facilities (SFs) include: (i) offset short-term liquidity shocks and dampen interest rate volatility; (ii) lower precautionary demand for reserves, funding costs, and lending rates; (iii) build liquidity buffers and contain financial stability risks while releasing liquidity in adverse reversals.
- Recommendation 1:
  - Consider a gradual modernization of monetary policy implementation by setting up an active liquidity management framework to allow the NRBT to steer short-term money market rates to its announced policy rate.
- Suggested modernization approach (gradual):
  - Two complementary workstreams:
    - Analytical capacity to assess the appropriate stance of monetary policy (suitable level of short-term interest rates given the peg to the basket or currencies).
    - Operational capacity to determine and calibrate money market interventions to align short-term money market rates with the desired policy stance, including activation and calibration to absorb excess liquidity.
  - NRBT policy rate should be aligned with policy rates in countries included in the currency basket adjusted with a risk premium to:
    - Maintain covered interest rate parity.
    - Limit arbitrage opportunities.
    - Support attractiveness of local currency for residents.
    - Limit incentives to circumvent capital controls (e.g., repatriation of export proceeds).
    - Limit need to rely on capital controls which lose effectiveness over time.
  - Operationally, NRBT should develop a framework for market-based active liquidity management to support better control of excess reserves and support the peg, with the policy rate used to communicate the policy stance linked to a medium-term inflation objective.
  - NRBT liquidity management operations should aim at aligning market conditions (short-term interbank market interest rates) with the announced policy stance to allow predictable placement of surplus liquidity and short-term funding among commercial banks.

*Source: IMF Technical Assistance Report*

### 45. The current reliance on the SRD to that end does not offer the NRBT an adequate level of

### 45. The current reliance on the SRD to that end does not offer the NRBT an adequate level of

### Short-term liquidity management: findings
- The current reliance on the SRD has various limitations: its impact on lending rate and inflation appears modest at best, and the presence of sizeable excess reserves can significantly reduce SRD effectiveness as banks would simply shift the excess reserves into the mandatory reserve category.12
- Changes in the SRD ratio cannot be relied upon to adjust liquidity conditions to short-term changes in the autonomous factors of liquidity in a timely fashion.
- Recent experience: in 2023 the NRBT had to offer deposit facilities for the retirement funds for targeted mopping of excess liquidity, then had to cancel them in 2024 to support the banks’ lending activities and their liquidity needs.
- Immediate priority: enhance short-term liquidity management by relying on the IRC to align short-term money market conditions with the desired stance of monetary policy.
- Under the proposed framework the NRBT would offer liquidity facilities to banks only (to mop up or provide liquidity to offset undesirable changes induced by evolution of the autonomous factors) and avoid financial intermediation that competes with the banking sector.
- Critical capacity gap: the NRBT needs to develop short-term liquidity monitoring and forecasting based on analysis of the NRBT balance sheet.

12 See in particular the 2024 Article IV Consultation Annex VI “Monetary Policy Transmission Mechanism in Tonga”.

### Design of the Interest Rate Corridor (IRC): current situation
- Positive features:
  - IRC includes two standing facilities banks can access at their initiative, setting a fluctuation band for short-term money market conditions.
  - OMOs in the form of issuance of NRBT Notes correspond to active liquidity management needs in a context of structural excess liquidity.
  - Calibration of interest rates for instruments is anchored on what the NRBT Board considers an appropriate level for short-term money market conditions (the interest rate decided for NRBT Notes).
- Missing / problematic features:
  - NRBT has not started the issuance of its Notes, despite a large structural liquidity surplus and an inactive interbank market.
  - The “announced” policy rate (the rate applied to the deposit standing facility) sends a confusing signal as to where the NRBT would want the short-term money market to trade.
  - The structure and objective of the IRC (steer short-term money market rate to its desired level) has not been fully disclosed to market participants.
- Consequence: the current IRC does not allow the NRBT to be in “the driver’s seat” regarding the operational objective of monetary policy.

### Key recommendations on IRC and operations
- Recommendation 2 (NRBT): Start monetary policy modernization by the Introduction of a conventional mid-rate interest rate corridor (IRC) with fixed rate and full allotment tenders.
  - Implement a mid-rate IRC that would involve:
    - Announcing a policy rate which would serve to communicate the monetary policy stance and would be the NRBT target for short-term interbank market rates;13
    - A commitment to use OMOs (auctions of NRBT Notes) to steer interbank market rates close to the operational target;
    - Setting standing facilities (SFs) that banks can use at their initiative, with interest rates pre-set by the NRBT, allowing banks to borrow from the NRBT (refinance standing facility) or deposit funds with the NRBT (deposit standing facility).
  - Mid-rate IRC systems provide stronger incentives for interbank trading than floor IRC systems, though they are more demanding to operate.14
- Specific operational recommendation:
  - Introduce a seven-day main liquidity management operation (MLMO), conducted weekly and anchored on the NRBT policy rate (see Table 7).
    - Given current structural excess liquidity, this means issuing seven-day NRBT Notes at the NRBT policy rate to reduce excess liquidity.
    - When structural liquidity turns negative, conduct seven-day repo operations at the policy rate.
  - Operate the MLMO initially through tenders at a fixed rate (the NRBT policy rate) and with full allotment, given limited short-term liquidity forecasting capacity.
    - NRBT will accept all bids from banks; volume determined by banks.
    - Conduct a meeting of the NRBT liquidity management team before each weekly MLMO to practice assessing liquidity developments and prepare for future calibration to variable-rate, fixed-amount tenders (Box 2).
- Table 7 (Design of the Proposed Interest Rate Corridor) — key elements preserved from source:
  - Categories of Monetary Policy Operations: OMOs (NRBT discretion) and Standing Facilities (Banks’ discretion)
  - Types of Instruments / Maturity / Frequency / Procedures:
    - Main liquidity management operation-MLMO: Issuance of NRBT Notes / Repo operations 1/ / Seven-day / Weekly / Fixed-rate, full allotment or variable rate fixed amount tenders
    - Fine-tuning OMOs: Deposits with NRBT / Repo operations 1// / Overnight / As needed / Policy rate serves as maximum/minimum bid rate on absorbing/providing operations
    - Marginal lending facility-MLF: Overnight repo operations / Overnight / IRC Upper bound (policy rate + 100/200 pbs). Collateral: NRBT Notes, Government of Tonga bills and bonds
    - Marginal deposit facility-MDF: Overnight deposits / Overnight / IRC Lower bound (policy rate – 100/200 pbs)
    - 1/ Collateral similar as for MLF.

13 This policy rate would serve as the interest rate for conducting sterilization operations (where the forecast is one of excess liquidity) and for liquidity providing operations (where the liquidity forecast indicates a liquidity shortfall).
14 They require better liquidity forecasting frameworks, more frequent OMOs, and supporting measures such as RR with reserve averaging to properly steer interbank rates and contain volatility.

### Standing facilities (SFs)
- SFs (MDF and MLF) must limit interest rate volatility and ensure automatic access.
- Banks must be able to access SFs daily through overnight operations.
- Criterion for MLF access: availability of collateral provided by banks.
- NRBT must not limit the volume of MDF operations; restrictions could cause interbank rate to exceed MLF or fall below MDF, undermining alignment with the NRBT policy rate.
- Recommendation 3: Move to mid-rate IRC with variable rates and fixed allotment tenders once a liquidity monitoring and forecasting framework is fully operational.
  - Migration to fixed-quantity, variable-rate OMOs depends on:
    - Market participants’ understanding of NRBT liquidity management operations;
    - Improving NRBT liquidity forecasting capacity;
    - Coordination with the Ministry of Finance on reliable government cash-flow forecasts.

### Interbank market development: findings and recommendations
- A functioning interbank market, particularly the overnight market, is essential for market-based monetary policy implementation through an IRC.
- Banks indicate general willingness to participate in interbank activities once structural excess liquidity has been absorbed and incentives restored.
- Obstacles reported by banks:
  - Capacity to assess potential counterparties’ soundness due to limited financial information.
  - Uncertainty regarding legal framework for collateralized interbank lending operations.
  - Some foreign banks face group policies preventing interbank operations with counterparties not fully compliant with AMLCFT regulations.
- Recommendation 4: Engage discussions with the banks to assess what specific actions NRBT could take to support interbank market activities.
- Recommendation 5: Adopt a wide IRC at the start.
  - Rationale: a wide corridor increases opportunity cost for banks to use SFs, encouraging interbank transactions and distributing liquidity among banks; useful at IRC introduction and when markets are underdeveloped.
  - Note: currently most central banks whose currency are included in the basket for the Tonga currency have a width of the IRC of 50 bps (Reserve Bank of Australia; Reserve Bank of New Zealand; European Central Bank; US Fed; Bank of England). Therefore, when it is introduced in Tonga, the width of the corridor would be larger.

### Short-term liquidity monitoring and forecasting: framework and recommendations
- Framework: analyze projected changes in main items of central bank balance sheet (net foreign assets; net lending to banks/OMOs; other items net/net domestic assets) and liabilities (currency in circulation; bank reserves; net position of the government) — see Table 8.
- Demand for bank reserves = Required reserves + excess reserves.
- Supply of bank reserves = autonomous liquidity position (first four items beyond central bank short-run control) + policy position (central bank direct lending to banks and net lending through OMOs).
- Central bank needs adequate information on autonomous factors of liquidity supply to steer liquidity.
- Government net position with the central bank often accounts for the most significant changes in autonomous liquidity supply; government cooperation in providing timely information on government transactions that affect liquidity supply is crucial.
- Recommendation 6: Strengthen liquidity monitoring and forecasting by developing a capacity to analyze and forecast weekly the autonomous factors of liquidity based on the NRBT balance of accounts.
  - Follow-up technical assistance will be provided to NRBT to assist implementation.
- Recommendation 7: NRBT should recommend to the Treasury to stop placing time deposits in the commercial banks, and to engage a study to assess the feasibility of setting a Treasury Single Account-TSA framework.
  - Treasury currently maintains accounts at NRBT and commercial banks:
    - One time deposit in a commercial bank for an amount of 41 million.
    - Aggregated balances of Treasury current accounts in commercial banks were about 100 million.
    - Treasury accounts with the NRBT (including project accounts and a main operating account) represent about 175 million.
- Recommendation 8: Develop operational guidelines for the activation of an ELA framework.

*IMF Technical Assistance Report*

### 68. Liquidity provisions for financial stability purposes must be administered by an ELA

### 68. Liquidity provisions for financial stability purposes must be administered by an ELA instrument

### Emergency Liquidity Assistance (ELA) administration
- Finding: The NRBT does not have in place an active ELA framework.
- Legal/mandate context: Promoting financial stability and a sound and efficient financial system is institutionally mandated in Article 4 of the NRBT Act; NRBT is exclusively responsible for regulation and supervision of banks and other financial institutions and should participate in resolution processes for troubled banks and financial institutions.
- Recommendation (from mission): The NRBT should develop operational guidelines for the activation of an ELA framework.
- Recommendation 9: Engage in discussions with the Treasury for the exchange of cash flow forecasts.
- Operational note: Treasury currently does not provide short-term cash flow forecasts to the NRBT, though such information is available on a weekly basis and Treasury officials have expressed willingness to share it with the NRBT.

### Reserve requirements — Overview of NRBT reserve balances
- Finding: Commercial banks’ statutory reserves balances (SRD) have steadily increased over the last two decades, with spikes in 2017 and 2023 reflecting changes in the SRD ratio.
- Finding: Excess reserve balances (total Exchange Settlement Account, ESA) have significantly outpaced SRD balances over time and most notably since 2020.
- Policy action in 2023: NRBT increased the SRD ratio to 15 percent and attempted to mop up a portion of ESA liquidity.
- Key statistic: ESA balances remain highly elevated at 364 million TOP as of July 2024.
- Distributional finding: The share of SRD balances is concentrated—most SRD balances are held by Bank B; Bank B’s SRD size has marginally increased over time.
- ESA-specific findings:
  - ESA follows similar long-term trend as SRD but with more short-term volatility.
  - Most of total ESA balance is held by Bank B.
  - Bank D’s share of ESA is significantly lower than its share of SRD and has sharply declined to almost nil, coinciding with a sharp increase in Bank B’s ESA share.
  - Box 3 note: The fifth percentile of Bank D’s ESA balances is five million TOP, which may complicate further increases in the SRD ratio to absorb a portion of the current ESA.

### ESA statistical properties and percentiles (January 2014 to end-2024)
- System-wide ESA percentiles:
  - 95th percentile: 399.67 million TOP
  - Fifth percentile: 139.70 million TOP
- ESA percentiles for individual banks (as presented):
  - Bank A: 5th = 28 ; 95th = 98
  - Bank B: 5th = 47 ; 95th = 219
  - Bank C: 5th = 18 ; 95th = 38
  - Bank D: 5th = 5 ; 95th = 93

### Design of the Statutory Reserve Deposit (SRD) — Conceptual considerations
- Functional reorientation:
  - Finding: The SRD function should be reconsidered to support a proposed IRC framework; within an interest rate based monetary policy framework (and related IRC system) the SRD would serve exclusively to adjust the structural liquidity position while the policy rate signals monetary stance.
  - Communication requirement: Changes in the SRD ratio should not convey a message regarding the stance of monetary policy and this needs to be clearly communicated to the market.
- Structural liquidity role:
  - Mechanism: An increase in the SRD ratio reduces the structural liquidity position by locking away liquidity; when starting from a liquidity surplus, OMO liquidity-absorbing operations would have to be reduced to leave interest rates unchanged.
  - Implication: SRD may be used to adjust structural position but not short-term liquidity fluctuations.
  - Recommendation 10 (summary): Rely on the SRD to adjust the structural liquidity position—Consider remunerating the SRD; Consider introducing averaging over the reserve maintenance period and linking SRD noncompliance penalty to the policy rate.
- Averaging provision:
  - Rationale: Allowing banks to comply on average over the reserve maintenance period (RMP) facilitates intertemporal smoothing of autonomous factor shocks and reduces short-term interest rate volatility, though volatility may rise at the end of the maintenance period.
- Remuneration:
  - Principle: Ideally, SRD should be remunerated at the opportunity cost to avoid interfering with monetary policy stance and undermining financial deepening, unless SRD aims to tax specific funding sources.
  - Mechanics: Remuneration is typically linked to the policy rate; a remuneration below opportunity cost creates a wedge between funding and lending rates.
  - Analytical tool: Compute the breakeven deposit rate and the minimum LR (lending break-even rate) to assess the wedge and marginal funding costs.
- Penalties for non-compliance:
  - Principle: Penalties should be financial and higher than the central bank’s standing marginal facility rate to incentivize compliance.
  - Guidance: Penalty rates are often anchored to the policy rate or a short-term market rate; penalties generally apply an interest rate to the shortfall for the duration of the shortfall.
  - Caution: Penalties should not be excessive to avoid reserve hoarding, over-compliance, and market functioning distortion.
- Governance:
  - Finding: Organizational ownership of the SRD framework depends on SRD objectives; if primarily monetary policy, administration should rest with operational or implementation committee; if macroprudential, the financial stability committee would be involved.

### SRD design recommendations (mission)
- Rely on the SRD to adjust the structural liquidity position and entrust administration to the NRBT unit in charge of monetary policy implementation.
- Consider remunerating the SRD, ideally at a level close to the policy rate; given Tonga’s circumstances (large excess liquidity and weak financial position), in the short-term a remuneration at a lower interest rate, or even no remuneration, would be acceptable.
- Consider allowing banks to meet the SRD on average over the reserve maintenance period to lessen the impact of un-forecasted autonomous factor shocks on short-term interbank market interest rate volatility.
- Link SRD non-compliance penalty to NRBT’s Policy Rate; set the penalty above MLF to incentivize compliance but avoid setting it too high to prevent liquidity hoarding.

### Collateral policy — current framework and assessment
- Legal basis:
  - NRBTA provisions: Section 38 empowers NRBT to extend loans to banks for price stability and financial stability objectives; Section 38(e) mandates Board to establish terms and conditions for collateralized lending; Section 48A authorizes repurchase of government and government-guaranteed securities as part of monetary policy operations.
  - Limitation: NRBTA is less explicit regarding asset types NRBT may accept in financial stability operations and modalities for collateral mobilization.
- Current collateral framework for monetary policy (Repos):
  - Document: “Standard Policy” on repurchase agreements (Repos), revised June 2024.
  - Role: Repo facility functions as the standard lending facility and the Repo rate is set as a ceiling for short-term market rates (Board Paper No. 26/2024).
  - Eligibility (preference order): NRBT notes, Government of Tonga Bonds, Government Treasury Bills.
  - Valuation:
    - NRBT notes issued at a discount.
    - Government Treasury Bills priced at their settlement price (issuance price).
    - Coupon-paying NRBT Notes and Government Treasury Bonds valued at face value.
  - Risk mitigation: NRBT does not implement haircuts, concentration limits, or overcollateralization; counterparties may borrow up to the value determined by NRBT.
  - Statutory limit note: Section 48A(2) sets a maximum limit on the amount of government debt the NRBT can hold at any time at 30 percent of the average ordinary income of the Government over the previous three years.
- Gaps and assessment:
  - Finding: NRBT has not established a collateral policy for financial stability lending (ELA) at the time of the mission.
  - Recommendation: Develop distinct but coherent collateral frameworks for monetary operations and for ELA; monetary operations should have predefined, public eligibility and valuation; ELA may need broader collateral acceptance, central bank discretion, and non-public valuation/haircut details to contain moral hazard.
- Conceptual guidance for modernizing collateral policy:
  - Key components to define: (i) eligibility requirements; (ii) valuation processes; (iii) risk control measures (haircuts, concentration limits, overcollateralization).
  - Policy tradeoffs: Parameters should manage financial and operational risks within the central bank's risk appetite while minimizing price distortions across asset classes and promoting market development via standardized approaches.
  - Eligibility assessment: Current restriction to public sector securities is appropriate given ample liquidity and Repo inactivity; NRBT should monitor distribution of government bond holdings to ensure banks’ access to Repo facility for short-term refinancing needs.
  - Market structure observation: While all banks hold significant liquid assets (primarily deposits, including with NRBT), only one bank has government bonds exceeding five percent of total assets and some banks hold no government bonds, meaning some banks would lack practical access to the Repo facility if needed.

*IMF Technical Assistance Report — excerpt*

### 87. A fair collateral valuation is essential to safeguard the central bank’s balance sheet while

### tarea2025065-print-pdf - 87. A fair collateral valuation is essential to safeguard the central bank’s balance sheet while

### Collateral valuation principles and market footprint
- Collateralized lending requires that central banks accept assets that may ultimately appear on their balance sheets in the event of a counterparty default; therefore, collateral should be marked to market based on sound valuation principles.
- Lending operations described are designed to absorb residual liquidity risk without disrupting traditional market pricing of risk.
- In the absence of a secondary securities market, accurate price discovery is hindered; the NRBT should develop suitable theoretical models to price all eligible collateral.
- Valuation approach hierarchy:
  - Primary: market-based valuation—collect quotes from market participants and aggregate into a composite price after filtering out outliers, based on specific criteria.
  - Secondary: when reliable quotes are not available, resort to theoretical models.

### Risk mitigation measures for collateral
- Risk mitigation measures manage financial and operational risks and generally fall into three categories:
  - Haircuts are calibrated to mitigate liquidity, market, credit, and other risks.
  - Margining provisions allow for additional collateral to be requested if collateral value falls during the loan; other things equal, haircuts should be lower with margining provisions in place.
  - Concentration limits limit exposure to a certain class of assets and diversify the collateral pools.
- Additional measure:
  - Over-collateralization requires additional collateral where there is an increased risk of counterparty default, as in the case of ELA.

### Haircut calibration components and approach to risk equivalence
- Objective: Haircuts should protect the NRBT’s balance sheet while achieving risk equivalence across asset classes—calibrated so expected losses are the same for all collateral types given the central bank’s risk appetite.
- Key components to include in haircut calibration models:
  - Liquidity risk — captures time-to-liquidation (expected time at which collateral can be liquidated at a reasonable discount); depends on size of central bank position relative to market absorptive capacities.
  - Market risk — model price volatility over the time-to-liquidation period; for marketable securities this typically uses interest-rate volatility models, with yields derived from the security’s discount curve.
  - Credit risk — model the eventuality of default of the ultimate debtor during the time-to-liquidation period and related expected losses and/or expected shortfalls.
  - Additional risks — e.g., exchange rate risk for foreign-currency collateral, valuation risk for pricing-model uncertainty for illiquid collateral, residual maturity risk where collateral cannot be sold, mobilization risk (e.g., legal risk for assets held abroad such as FX deposits).

### Valuation methodology specifics and market data
- Current NRBT valuation practice concerns:
  - Zero-coupon securities are valued at their settlement price (their discounted issue price).
  - Coupon-paying securities are priced at their nominal values.
  - The current methodology for zero-coupon securities fails to account for time elapsed since issuance; valuation of coupon-paying securities does not consider coupon payment schedules.
- Issuance-implied yields can be leveraged to develop theoretical valuation models, though yields are only observable on days of government debt issuance.
- Government Treasury Bonds data and implied yields:
  - Tonga’s Government bonds have an average maturity at issuance of five years.
  - In 2024, maturities issued: two bonds with maturity of three years, one with maturity of four years, and two with maturity of five years.
  - Most recent implied yields to maturity for 2024 are 2.5 percent, 2.5 percent, and three percent for maturities of three, four, and five years, respectively.
- Pricing approach suggested:
  - Combine issuance-implied yields with the NRBT Note rate to interpolate yields for different maturities.
  - Calculate bond price as the sum of discounted cash flows between the pricing date and bond maturity (see Annex VIII for fixed-income valuation details).
  - Nelson-Siegel (NS) methodology recommended to calibrate term structure curves for discount factors.

### Recommendations on collateral framework (Recommendation 11–14)
- Recommendation 11: Remove the preference order for monetary policy collateral and standardize mobilization procedures for assets.
  - Rationale: Eliminate preference order between monetary policy collateral; international best practice recommends against preference orders for collateral supporting monetary policy operations.
  - Action: Undertake necessary actions—such as establishing a Central Securities Depository and creating master agreements—to standardize mobilization procedures for both asset classes.
  - Concern: Current preference for NRBT Notes could potentially create price distortions among the country’s risk-free assets beyond liquidity-risk implications that OMOs address.
- Recommendation 12: Develop sound valuation methodologies to mark collateral to market.
  - Rationale: Review current valuation methodology and adopt a consistent valuation methodology across asset classes.
  - Implementation: Use issuance implied yields and theoretical bond-pricing (sum of discounted cash flows) as outlined; address valuation discrepancies for zero-coupon and coupon-paying securities.
- Recommendation 13: Coordinate with the Government of Tonga to issue bonds to help calibrate a term-structure sovereign yield curve.
  - Rationale: Issue bonds with maturities that allow construction of an entire yield curve for theoretical bond pricing models.
  - Suggested tool: Nelson-Siegel methodology to generate discount factors for each coupon payment and compute model-based mark-to-market values as sum of discounted cash flows.
- Recommendation 14: Develop and implement risk mitigation measures, including granular haircuts that capture market, liquidity, and credit risk.
  - Approach: Adopt granular haircut calibration, assess each key risk individually, and aggregate into a final haircut; use international comparisons where data constraints limit structural models.
  - Example benchmark: ECB applies six percent haircut for BBB-rated government securities; this could serve as the minimum haircut for NRBT covering the three financial risks for government securities with maturities less than one year.
  - Adjustments to minimum haircut may be required upward to account for:
    - (i) market risk — if Tonga’s yields exhibit greater volatility than Eurozone peers;
    - (ii) liquidity risk — if Tonga’s public securities are less liquid;
    - (iii) credit risk — if likelihood of sovereign default in Tonga is higher than that of BBB-rated government securities within the Eurosystem.
  - Additional adjustment: Haircut may be increased by five percent based on maturity brackets (see Annex VIII for further details).

### Communication — current situation and institutional gaps
- NRBT Strategic Plan 2023-2028 includes enhancing communication: establish effective, open, and clear channels; encourage open communication and feedback from employees; and communicate and monitor measurable key strategic results areas (KSRAs).
- NRBT communication instruments (overview from Table 10):
  - Monetary Policy Decision/Statement — Frequency: Semiannual — Available since: 2007 — Available electronically since: 2014 — Description: Flagship semiannual publication indicating how the Bank intends to conduct monetary policy over coming six months.
  - Monetary Policy Decision/Statement Press Release — Frequency: Semiannual — Available since: 2007 — Available electronically since: 2014.
  - Annual Report — Frequency: Annual — Available since: Early 90s — Available electronically since: 2002 — Provides overview including balance sheet, income statement, and audited financial statement cash flows.
  - Quarterly Bulletin — Frequency: Quarterly — Available since: 2002 — Available electronically since: 2002 — Comprehensive overview of economic and financial conditions including statistical data.
  - Monthly Economic Update — Frequency: Monthly — Available since: 2014 — Available electronically since: 2014 — Overview of global economy and outlook for Pacific Island Countries and Tonga.
  - Financial Data - Monetary Policy Data — Frequency: Monthly — Available since: 2010 — Available electronically since: 2010.
  - Other publications and data repositories — audited financial statements and several economic releases including banking sector development, overseas exchange transactions, remittances, GDP nowcast, and inflation nowcast.
- Observed gaps and needs:
  - NRBT has intensified communication but further improvements warranted; stakeholders request more information on Tonga’s financial markets and the health of the financial system.
  - No financial stability report is yet produced (one is currently being developed).
  - No institutional policy governing communications; governance elements (objectives, messages, audiences, channels, spokespersons, products, impact assessment, crisis communication) are not documented.
  - De facto elements summarized:
    - Objective: Enhance public awareness of NRBT policies and actions; suggested to add objective of enhancing monetary policy effectiveness via stronger public understanding and transparency.
    - Spokespersons: Governor and Deputy Governor; areas of responsibility not delineated; no role established for Board Members, Executive Directors, Department Directors—should be addressed in institutional policy.
    - Messages: Internal review executed by Economics and Research Department; Governor approves articulation; communications personnel not involved in policy deliberations—proposed institutional policy should address this.
    - Audiences: Financial market participants, media, economists and finance experts, real sector corporates, government, academia, and the public; guideline development recommended for tailoring messages.
    - Channels/tools/formats: Website is primary intermediary-free channel; marginal presence on social media; no online media rebroadcasting.
    - Communications products: Calendar of policy decisions, press releases, press conferences, periodic reports exist; notable absences include financial stability report or equivalent.
    - Impact assessment: Not currently done; recommendation to establish analytical techniques (machine learning techniques were employed in the mission to assess NRBT communications effectiveness and will be discussed in subsequent sections).
  - Organizational capacity:
    - Governance of communications subsumed solely by the Governor; no centralized communications unit.
    - One communications officer employed in Risk and Compliance department, but does not review or manage communications; role focused on media liaison and press conference planning.
    - Internal reporting: No internal communication reports; communication planning and monitoring are not in place to support Strategic Plan objectives.
    - NRBT has internal guidelines for departmental roles and responsibilities.

### Communication recommendations (Recommendation 15 and supporting elements)
- Recommendation 15: Develop a formalized institutional policy for communications along with key complementary frameworks.
  - Key elements recommended for an institutional communications framework:
    - Principles for effective communications: clarity, timeliness, precision, consistency, predictability, transparency, and equal access to information.
    - Communication framework covering objectives, principles, spokespersons, messages, audiences, channels, tools, formats, and effectiveness assessment.
    - Internal organization of communications function: information flow, procedures, guidelines, processes, and structure and staffing.
    - Communications products: calendar of policy decisions, press releases, press conferences to explain decisions, periodic monetary policy and financial stability reports, minutes of policy discussions, speeches, social media publications, and interviews.
- Implementation guidance:
  - Develop an implementation strategy with well-thought-out milestones balancing implementation needs and capacity constraints.
  - Define priorities and proceed gradually with mission recommendations.
  - Glean insights from other small central banks facing capacity constraints in building communication frameworks.

*IMF Technical Assistance Report | excerpt from source content.*

### 104. To assist with the implementation of this institutional policy, the NRBT should

### 104. To assist with the implementation of this institutional policy, the NRBT should

### Key complementary frameworks for implementing the institutional policy
- Develop a tool for the regular assessment of the effectiveness of its communications.
  - Develop a media monitoring function to track the media for NRBT-related publications.
  - Maintain a database tracking numbers of NRBT communication activities and statistics of owned communication channels—the central bank’s website and pages on social media.
  - Supplement media and social media monitoring with periodic surveys to assess:
    - level of trust in the central bank,
    - awareness of its mandate,
    - awareness of monetary policy objective and inflation target,
    - confidence in the central bank’s ability to reach its objectives.
  - Enhance internal capacity through capacity-building training and cross-departmental rotation.

- Establish a proactive external communication policy, supported by an internal message management process.
  - Shift from reactive communications (“we answer the questions determined by stakeholders”) to proactive communications (“we determine the key information—messages—we want to deliver to stakeholders”).
  - Conduct message formation and determination in advance and on a regular basis (e.g., weekly), independent of upcoming public engagements.
  - Deliver messages proactively whenever the central bank sees the need to reach its policy and communication objectives.

- Develop an organizational structure to manage communications and message content.
  - Increase staffing solely responsible for communications and enhance the skills mix.
  - Create a Communications division.
  - Priority: add one staff member to the existing communications officer; the two positions would constitute the Communications division.
  - Responsibilities: develop and monitor adherence to the institutional policy, assist with governance of communications, message management and content creation.

- Develop and execute crisis communication simulation exercises.
  - Hold crisis simulation exercises to be better prepared for unexpected adverse developments, test crisis communications protocol, and fine-tune crisis communications procedures and plans.

*Recommendation 16: Develop a financial stability report.*

### Rationale for a financial stability report (Recommendation 16)
- Serves as an early warning system identifying potential vulnerabilities and risks within the country's financial system.
- Supports development of the interbank market by providing participants with important information.
- Explains proactive measures taken by authorities to maintain stability or prevent financial crises.
- Improves transparency and accountability for the central bank by identifying risks.
- Ensures market participants are well-informed about the overall health of the financial system to maintain confidence in the central bank.
- Highlights and addresses financial distress before escalation, thus maintaining economic stability and promoting sustainable growth.

### B. Monetary Policy Communication — Current situation (paragraphs 106–109)
- Central banks must clearly articulate:
  - strategy for achieving monetary policy objectives,
  - policy interest rates movements,
  - expected developments in economic conditions,
  - constraints and risks associated with policies.
- Stakeholders, including NRBT’S monetary policy counterparts, advised the mission they are unaware of the central bank’s current monetary policy instruments and the availability of deposit facilities.
- NRBT has a well-developed set of tools to communicate its monetary policy framework and decisions; the key instrument is the monetary policy statement.
  - There are no other instruments providing a periodic review of monetary policy.
  - Periodic reviews such as the quarterly bulletin and the monthly economic update are expositions of statistics and data.
  - All communication is homogeneously distributed; NRBT does not distribute tiered communication for varying audiences.
- NRBT communicates reliance on non-conventional monetary policy tools in the Annual Report (e.g., activation of non-conventional liquidity mopping operations, reliance on the minimum loans/deposit ratio to influence bank’s lending).
- Communications mention adverse impact of a persistently high level of excess liquidity on monetary policy transmission and on the interbank and repo market activities.
- NRBT communicated its 2017 decision to implement a floor system at an interest rate of zero, but further communication on the framework remained absent, reflecting incomplete implementation.
- Currently no strategy exists for communicating operationalization of existing or proposed frameworks and no plan has been developed to engage banking counterparties and members of the public to promote greater awareness of a modernized framework.

### Recommendations for monetary policy communication
- Recommendation 17: Clearly communicate, following international best practice, the implementation of the proposed monetary operational framework.
  - Communicate operationalization of the modernized monetary policy framework with clear and strategic market communication.
  - NRBT statements could include:
    - policy measures to absorb excess liquidity in the first stage,
    - subsequent measures to induce interbank lending and steer interbank rates towards the policy rate in the second stage.
  - Possible key messages:
    - The essential reason for the change in approach (to enable the NRBT to rely on market-based mechanisms to influence conditions and mitigate vulnerabilities to the peg) and how this connects to other economy-wide objectives (e.g., mitigating the pandemic, dealing with a potentially systemic crisis).
    - The process of monetary policy implementation in coming months/years, including the role to be played by counterparties and the design of the SRD.
    - The commitment of NRBT to modernizing monetary policy regardless of the financial cost, given the important benefits of an effective transmission mechanism to macroeconomic stability.

- Recommendation 18: Expand outreach and adopt multitiered communication for monetary policy communication.
  - Tailor communication to audience by appropriate technical depth and level of detail.
    - Market participants (domestic and foreign) and specialized journalists require specific information.
    - Academic institutions and/or think tanks can help explain operationalization consistent with best practice.
  - Calibrate level of engagement with each stakeholder and identify appropriate channels and platforms (traditional communications, social media, NRBT’s website).
  - Extend outreach on monetary policy issues through regular meetings with external stakeholders (experts, media, Ministry of Finance, and other officials).
  - Increase dialogue with financial market participants and expand outreach to other counterparties during initial stages of monetary policy implementation.
  - Gather feedback on policy, communication effectiveness, operational framework and expectations to enhance efficiency of implementation and strengthen interest rate transmission.

### C. The Effectiveness of NRBT Communications (analysis and findings)
- A quantitative analysis examined effectiveness along publications’ readability, clarity, and thematic breadth (paragraph 113).
  - Multifaceted framework: lexical, syntactical, and semantical analyses.
  - Lexical analysis: readability metrics determine wording complexity.
  - Syntactical analysis: sentence structure complexity using dependency measures.
  - Semantical analysis: examines topic, communication stance (backward or forward-looking), addressed audience, and policy sentiment.
  - A novel text-based classifier finetuned for central bank communications (based on multilingual Large Language Models similar to ChatGPT) was used for semantical analysis.
- Scope of the analysis:
  - Examined more than 20 years of published documents (annual reports, monetary policy statements, press releases).
  - Collected similar publications from Fiji, Papua New Guinea, Samoa, Solomon Islands, and Vanuatu for cross-comparison.

Lexical analysis — key findings
- Tonga has significantly streamlined traditional publications in the last few years.
  - Monetary policy decisions were among the largest regarding number of words compared to the peer group, but the length has decreased since 2019.
  - Monetary policy decisions are now similar in size to other peer group countries.
  - Length of Tonga’s annual reports has substantially decreased since 2022, while other countries have maintained or increased size.
- Readability of monetary policy decisions:
  - The lexical readability, measured by the Flesch-Kincaid Reading Ease score, of NRBT’s monetary policy decisions was trending upward and was among the peer group's three most readable decisions until 2021.
  - The lexical readability has substantially fallen since 2021.
  - Decrease in readability explained by increased use of more complex wordings (measured in average syllables per word) and lengthier sentences (measured in average words per sentence).
  - By 2024, NRBT’s monetary policy decisions reached a readability score of below 40, placing the readability classification of Tonga’s monetary policy decisions as “difficult” for the public.
- Readability of annual reports:
  - Tonga’s annual reports are well-placed compared to other peer countries.
  - Word complexity increased in the last few years (reducing readability), but sentence length shortened (increasing readability), yielding little net change from 2023 to 2024.

Syntactical analysis — key findings
- Tonga’s monetary policy decisions have simpler sentence structures than the average peer country, favoring clarity, but have become more convoluted since 2022.
  - Document clarity measured by average syntactic dependency depth (higher = more convoluted; lower = more direct).
- Tonga’s annual reports sentence structure is similar to Fiji’s and has become more direct (less complicated) since 2022.
- Simplifying sentence structure can improve public understanding and engagement, especially for audiences without specialized financial knowledge.

Semantical analysis — approach
- State-of-the-art methodology using Large Language Models (LLMs) finetuned for central bank communication.
- LLMs classify text at the sentence level for:
  - topics,
  - communication stance,
  - audiences,
  - policy sentiment.
- The LLM was used to classify each sentence across documents and the analysis aggregates sentence-level classifications using heuristics (e.g., by topic).

*Source: IMF Technical Assistance Report.*

### 121. For over 25 years, the topic distribution of Tonga's central bank communication has

### tarea2025065-print-pdf - 121. For over 25 years, the topic distribution of Tonga's central bank communication has

### Communication topics and concentration
- Monetary policy has been the most prevalent topic in Tonga’s central bank communication for over 25 years.
- Exceptions where crisis management was more discussed occurred during global adverse events, specifically the global financial crisis in 2008 and the COVID-19 outbreak and following inflation surge.
- Tonga’s monetary policy decisions discuss more topics (are more diversified) than the average country in the peer group (Fiji, Papua New Guinea, Samoa, Solomon Islands, Vanuatu), which may undermine policy messages and economic reasoning.
- Topic concentration is evaluated using the Herfindahl-Hirschman Index (HHI):
  - Higher HHI values indicate greater focus on fewer topics (higher concentration).
  - Lower HHI values indicate more diverse thematic coverage.
- Observations:
  - Tonga’s monetary policy decisions show a trend toward even more diversification since 2015.
  - Tonga’s annual reports have similar concentration levels to other sample countries.

### Forward-lookingness of monetary policy decisions
- Forward-looking score definition: share of sentences classified as forward-looking as a percentage of the total sentences, irrespective of topics.
- Tonga’s monetary policy communication pattern mainly comprises statements regarding exchange rates, inflation, and economic activity, with an increased emphasis on inflation in recent years.
- Until 2022, Tonga had one of the least forward-looking monetary policy decisions compared to peer countries.
- Rationale for improving forward-looking content:
  - Forward-looking statements are critical to anticipating external shocks and maintaining macroeconomic stability.
  - Tonga’s reliance on imported goods amplifies the impact of global economic trends, requiring proactive measures to address inflationary pressures and manage liquidity.
  - Incorporating forward-looking elements helps align monetary policy with expected changes in international reserves, trade dynamics, and remittance flows.

### Textual metrics used to assess communication
- Metrics analyzed: Net Policy Sentiment (NPS), Net Confidence Index (NCI), Straightforwardness Index (SI), Explanation Index (EI).
- Definitions and interpretations (as provided):
  - Net Confidence Index (NCI) = (C − R) / (C + R)
    - C and R are the quantity of confidence-building and risk-highlighting sentences.
    - Domain: [−1,1]
    - Higher values indicate a predominantly optimistic communication tone; lower values indicate a cautious tone.
  - Straightforwardness Index (SI) = (N + |H − D|) / (N + H + D)
    - N, H, D are the quantity of neutral/balanced, hawkish and dovish sentences.
    - Domain: [0,1]
    - Measures homogeneity of communication signals within the same monetary policy decision; coexistence of hawkish and dovish sentences suggests multiple contingency scenarios or inconsistency.
  - Explanation Index (EI) = (C + R + N) / (H + D)
    - C, R, N, H, D are the quantity of confidence-building, risk-highlighting, neutral/balanced, hawkish and dovish sentences.
    - Domain: ≥0
    - Measures the number of explanatory sentences per stance-related sentence (policy-stance statements).
    - Example interpretation given: If the EI = 2, it means that for every policy stance statement, there are, on average, two explanatory sentences.

### Empirical findings from the metrics
- Net Confidence Index:
  - Evaluated across three samples: (a) monetary policy decisions and monetary policy reports; (b) financial stability reports; (c) all traditional central bank communications.
  - Tonga’s central bank communication strongly leaned toward confidence-building statements after 2014, especially about monetary factors.
  - Significant rise in net confidence-building statements in monetary policy communications from 2014 to 2017 aligns with the decline in lending interest rates to the private sector.
  - In recent years, communication on monetary factors remains confidence-building, albeit to a lesser degree than observed between 2014 and 2017.
- Straightforwardness Index:
  - Tonga’s monetary policy decisions generally have an unidirectional communication signal, but the straightforwardness index has notably declined after 2022.
  - Decline reflects a growing prevalence of mixed signals within the same monetary policy decisions, suggesting increasing discussion of multiple contingency scenarios.
  - The decline in straightforwardness during adverse economic conditions mirrors similar observations in some peer countries.
- Explanation Index:
  - Tonga’s monetary policy decisions devote five sentences to explain each stance-based sentence.
  - Each stance-based sentence in Tonga’s monetary policy decisions is supported by five explanatory sentences.
  - Other countries in the sample exhibit a similar ratio, except for Fiji, which typically includes more explanations for each policy action.
  - Note: The explanation index considers only the main body of the monetary policy decision; other documents may complement explanations.

### Recommendations to enhance communication (Recommendation 19)
- Enhance the effectiveness of communication by:
  - (i) improving readability indices of monetary policy decisions by reducing wording complexity;
  - (ii) producing more focused monetary policy statements by better managing the variety of topics discussed;
  - (iii) integrating more forward-looking elements in monetary policy statements.
- Additional implementation guidance derived from NLP analysis:
  - Improve readability of all communication:
    - Prioritize clear and concise language, focusing on essential messages.
    - Use shorter and more direct sentence structures to enhance accessibility and engagement for non-specialized audiences.
  - Produce more focused monetary policy statements:
    - Reduce topic diversification to avoid diluting central policy messages and priorities.
    - Ensure main objectives of monetary policy remain prominent to prevent misunderstandings about core intentions.
  - Integrate more forward-looking elements:
    - Better prepare stakeholders for potential economic changes (such as inflationary pressures or fluctuations in international reserves).
    - Strengthen policy credibility and stakeholder confidence by increasing forward-looking content across monetary policy decisions and related publications.

### Implementing the proposed monetary operational framework
- Implementation principles:
  - The implementation needs to be gradual, properly articulated, and well explained.
  - NRBT should engage market participants to explain rationale, expected effects, and broad timeline once the framework is internally defined.
- Follow-up virtual technical assistance:
  - Proposed virtual TA period: from 2025 Q2 to 2026 Q2.
  - TA objectives:
    - Setting up a short-term liquidity monitoring and forecasting framework and estimation of demand for reserves.
    - Developing the collateral framework for liquidity-providing monetary operations.
- Two-stage implementation approach:
  - Stage 1 (priority actions):
    - Transition to the new IRC framework based on regular conduct of MLMO: weekly issuance of 7-day NRBT Notes at the policy rate, through fixed rate, full allotment tenders.
    - Consider increasing SRD ratio.
    - Consider some SRD remuneration.
    - Link SRD noncompliance penalty to policy rate.
    - Develop capacity to assess level of the policy rate.
    - Start developing a short-term liquidity monitoring and forecasting framework based on the autonomous factors (AFs).
    - Rely on current collateral framework with the addition of haircuts.
    - Develop communication strategy to outline MP implementation and modernization.
    - Implementation of recommended changes to the SDR could occur at this stage, except moving to SDR averaging over the RMP.
  - Stage 2 (subsequent actions):
    - Weekly issuance of NRBT Notes at variable rate, fixed allotment tenders.
    - Introduce averaging provisions over the RMP.
    - Remunerate SRD at opportunity cost.
    - Short-term liquidity forecasting framework fully operational.
    - Implement a modernized collateral framework.
    - Communication strategy fully operational.
    - Introduce fine tuning open market operations (OMOs).
- Proposed MCM Technical Assistance program:
  - Virtual TA from 2025 Q2 to 2026 Q2 to set up a short-term liquidity monitoring and forecasting framework and estimation of demand for reserves.
  - Virtual TA from 2025 Q2 to 2026 Q2 for collateral framework for liquidity-providing monetary operations (operational arrangements for collateral valuation, haircut, and margin requirements).

### Conclusion and operational policy recommendations
- If implemented:
  - By the end of Stage 1, NRBT would have transitioned to the new IRC framework based on regular MLMO through fixed rate, full allotment tenders.
  - Developing proper liquidity forecasting capacity anchored on standard monitoring of autonomous factors would enable a market-based framework to influence money market conditions.
  - The move to variable rate, fixed amount MLMOs, combined with RMP averaging and fine-tuning OMOs, would complete the transition to a modernized framework with effective monetary operations and transmission.
- Collateral framework recommendations:
  - Remove the preference order for monetary policy collateral and standardize mobilization procedures for assets.
  - Develop sound valuation methodologies.
  - Assist with the calibration of a term-structure sovereign yield curve.
  - Implement risk mitigation measures to protect the central bank’s capital from adverse market developments.
  - These recommendations aim to bridge gaps and institutionalize best-practice methods underpinning liquidity-providing monetary operations.

*Source: IMF Staff Calculations and Mission’s elaborations as presented in the IMF Technical Assistance Report.*

### 135. Recommendations were provided to enable the transition to a modernized,

### 135. Recommendations were provided to enable the transition to a modernized, comprehensive, and effective communications framework.

### Communications framework and monetary policy messaging
- Provide detailed guidelines for developing an appropriately designed and well-articulated institutional policy to govern communications and develop complementary frameworks.
- Monetary policy communication guidelines to:
  - Effectively communicate the implementation of the proposed monetary operational framework.
  - Produce multitiered communication and expand outreach.
- Recommendations to modify NRBT message content to improve effectiveness by:
  - Improving readability.
  - Being more focused.
  - Integrating more forward-looking elements.
  - Reducing mixed signals.

### Follow-up missions and technical assistance (TA)
- Virtual TA scheduled from 2025 Q2 to 2026 Q2 to assist with:
  - Setting up a short-term liquidity monitoring and forecasting framework, as well as estimation of demand for reserves.
  - Developing the collateral framework for liquidity-providing monetary operations.
- NRBT may consider requesting future IMF missions on:
  - Activation of an ELA framework.
  - Financial stability communication.
  - A Central Bank Transparency (CBT) review.

### Implementation of past IMF recommendations (Annex I)
- Emphasis on benefits of a more market-based monetary policy framework.
- Key past recommendations and progress:
  - Rely on the NRBT policy rate to communicate and implement monetary policy.
    - Comment: Adjusting the SDR ratio is hampered by sticky lending rates and sizeable excess reserves; reliance on the policy rate recommended.
  - Establish an interest rate corridor (IRC).
    - Comment: Set interest rate floor on excess reserves consistent with macroeconomic conditions to influence banks’ lending rates.
  - Review of the indicative inflation reference rate.
    - Progress: The NRBT adjusted the indicative inflation reference rate to five percent per annum, down from the prior 6–8 percent.
    - Recommendation: Implement a regular review process for the reference rate.
  - Enhanced and effective communication strategy.
    - Recommendation: Commit to transparent communication, stakeholder engagement, and public education.
- September 2016 MCM TA mission recommended reducing the indicative inflation reference range of six–eight percent; authorities reduced it to five percent per annum.

### Table A.1 — Selected recommendations and progress (summary)
- Continue consistency between the indicative inflation reference rate and historical/medium-term outlook.
  - Last adjustment in 2016 when the indicative inflation rate was lowered to 5 percent from 6–8 percent.
- Clarify recapitalization process and enhance accountability.
  - Progress: No progress.
- Continue to take advantage of exchange rate flexibility.
  - Progress: NRBT has used ER flexibility to enhance monetary policy effectiveness.
- Announce Board policy decisions shortly after meetings.
  - Progress: NRBT releases a Press Release of policy decisions by the Board the same day it is approved; decisions communicated via press releases, monthly economic publications and Monetary Policy Statements.
- Introduce a floor system with interest on excess reserves as monetary policy rate and announce current monetary policy rate of zero percent.
  - Progress: Floor system introduced in November 2016. Revised IRC system announced in April 2024 (see Annex II).
- Tighten policy by raising interest on excess reserves if overheating emerges.
  - Progress/comments: The NRBT has not moved its policy rate in over a decade; decision influenced by excess liquidity and GDP recovery considerations.
- Strengthen liquidity forecasting and management by analyzing autonomous factors.
  - Progress: No progress. Exercises conducted monthly exclusively with data provided by commercial banks.
- Consider issuing NRBT notes to encourage interbank transactions and establish interbank interest rates.
  - Progress: No progress. NRBT has not issued notes; at times offered deposit facilities for targeted mopping of excess liquidity.
- Introduce SRD averaging.
  - Progress: No progress.

### Annex II — NRBT Board Paper No. 26/2024 (April 2024): Interest Rate Corridor
- Current interest rate policies:
  - Policy Rate: Zero percent (July 2016).
  - NRBT Note Rate: Two percent below banks’ average one month deposit rate (since March 2010).
  - Repo Rate: 3.9 percent (since October 2010) - two percent above the interbank lending rate.
- Eligible securities for repo, in order of preference:
  - NRBT Notes.
  - Tonga Government Bonds.
  - Tonga Government Treasury Bills.
  - Note: As of Friday, 12 April, 2024, total banks’ holdings of eligible securities was $37.12 million (all Government Bond holdings).
- Repo facility terms and conditions:
  - Interest Rate: Current annual interest rate on Repo was based on the overnight interbank rate plus 200 basis points; last Board approved repo rate was 3.9 percent in October 2010.
  - Term: Repo is for a period of seven days or any other longer period as the commercial bank may request.
  - Limit: Minimum amount for Repos is T$50,000. The amount advanced should not exceed the market value of the securities pledged.
- Observations:
  - Repo facility and interbank lending market inactive for over a decade due to excess liquidity; no indicative interbank lending rate.
  - NRBT policy rate should influence interbank lending rate, not vice versa.
- Corridor rationale and medium-term suggestions:
  - Corridor: Standing lending facility (repo) as ceiling and standing deposit facility as floor to encompass interbank rates and reduce volatility.
  - Consider issuance of NRBT notes in the medium term to absorb excess liquidity and help establish interbank interest rates; notes issued via OMOs could be used as securities for repo or interbank dealings.
- Draft resolutions for endorsement:
  - Policy Rate (Interest on the banks’ Exchange Settlement Accounts): zero percent — Floor rate.
  - NRBT Notes Rate: two percent — mid-rate.
  - Repo Rate: Fixed at four percent as the ceiling rate.
  - Conditions for the Repo Facility to remain the same except for the revised interest rate policy.
  - Corridor to be reviewed annually as NRBT moves towards a more interest rate-based framework.

### Annex III — Tonga’s Exchange Controls and Foreign Exchange Interbank Market
- Legal and administrative framework:
  - Foreign Exchange Control Act 2018 establishes comprehensive exchange controls.
  - All foreign currency payments from pa’anga accounts are subject to exchange control requirements with supporting documents verifying bona fide nature.
  - Commercial banks must provide underlying documentation when purchasing foreign exchange from the NRBT.
- Delegation and approval authorities:
  - NRBT delegated approval authority to licensed commercial banks and authorized foreign exchange dealers (FEDs) for current and capital payments up to a given ceiling; all other payments require NRBT approval.
  - Commercial banks and authorized FEDs delegated authority to approve capital payments not exceeding T$100,000 an application, subject to unaudited accounts and company resolution.
  - NRBT approval required for outward capital payments exceeding T$100,000 accompanied by audited accounts and company resolution.
- Current account transactions:
  - Exporters must repatriate to a pa’anga account with a commercial bank licensed in Tonga within six months of exportation at least 60 percent of foreign exchange earnings, unless NRBT approves extension or retention.
- Provisions specific to financial institutions:
  - Commercial banks’ nostro account balances may not exceed T$1 million daily; any excess must be sold to the NRBT within two business days.
  - FX-denominated loans to customers must be financed from overseas banks or customer foreign exchange receipts; banks may not buy FX from the NRBT to finance such loans.
  - Statutory reserves deposit amount is 10 percent of deposits and similar liabilities of banks in both local and foreign currency; no separate requirements by currency.
  - Commercial banks’ net FX positions subject to prudential limits: overall net positions 25 percent of eligible capital and single currency net positions 12.5 percent of eligible capital.
- NRBT FX interventions and interbank market features:
  - NRBT relies on a foreign exchange standing facility allowing the four licensed commercial banks to purchase foreign exchange from the NRBT on demand.
  - NRBT sets the US dollar–pa’anga rate daily and buys and sells at these rates to commercial banks.
  - Licensed commercial banks may trade among themselves in the FX interbank market; NRBT does not intervene directly with market participants at quoted rates.
  - Commercial banks’ total nostro balances may not exceed T$1 million daily; excess must be sold to NRBT within two business days.
  - No limits on spreads for interbank foreign currency transactions among commercial banks.
  - NRBT licensed four commercial banks and sixteen FEDs.
  - Banks’ spreads on telegraphic transfers limited to: 220 basis points (US dollars), 300 basis points (Australian dollars), and 350 basis points (New Zealand dollars).
  - FED licensing types: Ten Type A, five Type B, one Type C; Type A authorized for money conversions, foreign currency payments, and inward/outward remittances; Type B authorized for inward remittances; Type C for money conversion only.

### Annex IV — FX Composite Model and Currency Basket Weight Determination
- Weights determined to minimize exchange rate volatility subject to inflation passthrough volatility constraint:
  - Constraint: F(w, .) ≤ σ̅^2 where σ̅^2 set by the central bank.
  - Baseline σ̅^2 set to the minimum of a rolling window 60-month volatility of the log of imported CPI over the sample period.
- Sensitivity analysis:
  - A scale factor multiplies the baseline σ̅^2: scale factor = 1 reproduces original result; lower (higher) value corresponds to tighter (looser) constraints.
  - Above a certain scale factor the inflation constraint is too high and does not affect minimization (global minimum).
  - Below a certain scale factor the constraint forces the optimization to find a sub-optimal solution with a higher minimized value.
- Results and outputs:
  - Figure B.1 (described) shows minimized function value vs. scale factor with dashed vertical line at scale factor = 1 and dashed horizontal line at global minimum.
  - Table B.1 presents the volatility of inflation constraint.
  - Table B.2 displays FX Composite Model Results: Inflation Passthrough Coefficients.
  - Table B.3 displays FX Composite Model Results: Total Inflation Passthrough.

*IMF Technical Assistance Report | 63–73*

### Annex V. The Use of Interest Rate Corridors in

### Annex V. The Use of Interest Rate Corridors in Monetary Policy Implementation

### Introduction
- Migration from reliance on direct (administrative tools) to indirect (market-based) monetary instruments has increased central banks’ focus on short-term money market rates and led to growing reliance on interest rate corridors (IRCs) to stabilize short-term money market rates.
- Banks’ discretionary access to central bank liquidity at pre-set interest rates provides an automatic mechanism to limit interbank market rate volatility:
  - lending standing facility = ceiling; deposit standing facility = floor (for maturities equivalent or close to those applying to the standing facilities).
- By the 1990s central banks used a variety of arrangements; later reliance on a lending and deposit standing facility to set an IRC within which short-term interbank market rates could fluctuate became prevalent.

### Primary Purpose of an IRC
- Primary monetary policy purpose: set boundaries for short-term market interest rate volatility.
- Mechanism:
  - When market rates rise above the lending facility rate, banks borrow from the central bank instead of the market, preventing market rates from moving above the standing facility rate.
  - When market rates fall below the deposit facility rate, banks deposit with the central bank, stabilizing market rates at the central bank’s deposit rate.
- Provided the IRC width is sufficiently wide, interbank market trading, development, and deepening can continue within the corridor.
- Role across regimes:
  - Inflation-targeting (IT) regimes: IRC facilitates steering short-term money market rates to the level desirable to achieve the inflation target.
  - Monetary-targeting frameworks: focusing short-term liquidity management on reserve money rather than banks’ reserve balances and short-term interest rates can produce unwarranted short-term interest rate volatility that muddles policy signals.
  - Pegged exchange rate regimes: IRC can help steer short-term interest rates to a level aligned with the relevant foreign interest rate adjusted for the risk premium.
- Containing day-to-day volatility of short-term interest rates is essential for anchoring the yield curve and enhancing monetary policy transmission.

### Mapping of the IRC with Other Monetary Policy Instruments
- Adequate pairing of the IRC with other instruments is critical for effectiveness.
- In developed money markets, IRC operates in support of OMOs to limit short-term money market interest rate volatility.
- Key factors affecting the function of standing facilities:
  - frequency and flexibility of OMOs,
  - willingness to accept some market rate volatility,
  - degree of reliance on standing facilities (greater willingness to accept volatility → lower reliance on standing facilities and/or wider corridor, and vice versa).
- Frequency of OMOs depends on:
  - degree of development of the interbank market;
  - width of the corridor;
  - design of reserve requirements (RR) and capacity to absorb large/unexpected liquidity shocks;
  - accuracy of short-term liquidity forecasting framework;
  - possibility to conduct fine-tuning (overnight OMOs).
- Typical practices:
  - Daily OMOs: used when interbank market is shallow; liquidity forecasting accuracy is limited; RR cannot absorb large shocks; market volatility is high and central bank prioritizes stabilizing interbank rates.
  - Less frequent OMOs (e.g., weekly): used when interbank market allocates liquidity smoothly; weekly OMOs incentivize interbank trading and provide better market signals. Inclusion of fine-tuning OMOs (overnight) allows response to large or end-of-maintenance-period shocks.
- Reserve requirement averaging:
  - Averaging provisions reduce banks’ recourse to standing facilities and stabilize demand/provision of interbank funds.
  - Absence of RR or averaging, combined with desire to limit short-term interest rate volatility, likely results in frequent use of the IRC unless fine-tuning OMOs absorb short-term shocks.

### Options for Operationalizing an IRC
- All corridor systems combine a short-term (overnight) lending standing facility with a deposit standing facility; OMOs may be used to influence market rates within the corridor.
- Suitability of introducing a formal policy rate must be assessed; it can strengthen policy signaling and guide interbank rates if conditions are met.
- Three configurations:
  - Corridor with no official policy rate:
    - Fits countries relying on reserve money as operational target transitioning toward interest-rate based framework.
    - Lack of point policy rate may make it harder to stabilize interbank rates within the corridor but provides flexibility to keep day-to-day operations consistent with reserve money targets.
    - Persistent drifts under strict reserve money targeting should trigger a shift of the corridor in the same direction; under flexible monetary targeting reassess longer-term reserve money path.
  - Floor system (standing deposit facility rate serves as target for interbank rates and policy rate):
    - Reduces need for fine-tuning operations; attractive for countries with weak liquidity forecasting capacity and/or structural liquidity surpluses.
    - Provides fewer incentives for banks to engage in interbank trading.
  - Mid-rate corridor system (policy rate is an announced target):
    - Policy rate used to price OMOs and align interbank rate; commitment to use OMOs to steer interbank rates close to the target.
    - More demanding to operate than floor systems: requires better liquidity forecasting, more frequent OMOs, and supporting measures such as reserve requirements with reserve averaging to properly steer rates and contain volatility.
    - Provides stronger incentives for interbank trading.

### Maturity of the Standing Facilities
- IRCs are typically operated at the shortest money-market maturity (overnight), where the central bank monopolizes liquidity supply and where most interbank trading occurs as banks adjust reserve positions to planned levels amid unexpected shocks.
- On the last day of the reserve requirement maintenance period, banks have an incentive to trade overnight to avoid liquidity effects of longer-term operations that carry over.

### The Optimal Width of the IRC
- Tradeoffs:
  - Too wide a corridor may hamper market development, policy signal transmission, and can lead to reserve hoarding and higher day-to-day interest rate volatility.
  - Wide corridors discourage reserve intermediation via the central bank and can reduce overnight interbank trading, complicating liquidity management and discouraging market participation.
- Transitional considerations:
  - Wide corridor can be useful initially to de-incentivize banks from relying on the central bank for short-term liquidity management.
  - As interbank markets mature and banks’ treasury functions improve, a narrow IRC can better anchor short-term rates, support development of longer-term securities markets, and improve transmission along the yield curve.
  - In underdeveloped financial markets these considerations are less critical.
- High interest rate levels and other volatility absorption mechanisms may justify a wider corridor.
- Reserve averaging and longer maintenance periods reduce daily volatility and help markets absorb shocks.
- Numerical ranges and views:
  - Across central banks the width of the corridor can go from 50 bps to above 200 bps.
  - Some central bank experts suggest that 200 bps are optimal.
  - The tradeoff is between interest rate volatility and market development.

### Implications of an IRC for Central Banks’ Costs
- Managing liquidity to stabilize short-term rates and align with the policy rate can be costly when structural liquidity surpluses are large, but failure to manage could be costlier:
  - Structural surpluses must be drained through OMOs under conventional flexible-rate fixed-quantity and fixed-rate full-allotment systems, or be remunerated at the deposit rate under a floor system, generating interest costs.
  - Many central banks in transition may rely on un-remunerated reserve requirements to reduce structural surpluses; failing to do so would produce below-target market rates and a looser-than-intended policy stance.
  - Periodically draining excess liquidity to save costs can increase liquidity risk, interest rate volatility, risk premiums, and weaken policy transmission.
- Cost comparisons across frameworks:
  - Providing an interest-earning deposit facility may be less costly than assumed because:
    - Precautionary demand for overnight balances can be very low and the use of the deposit facility can be close to zero in a well-managed mid-corridor system.
    - Theory suggests precautionary demand should be zero when the target is set at the middle of the corridor and the distribution of payment shocks is symmetric with zero mean; then targeting a zero-overnight balance minimizes banks’ liquidity costs.
    - In practice, some small excess reserves are likely because banks may not find profitable to lend very small amounts due to fixed trading costs and credit risks.
  - Paying interest on excess reserves can be partly or fully offset by reduced OMO costs under a floor system:
    - With a large structural surplus, alternative to remunerating excess reserves at the policy rate would be to mop most up using longer-maturity OMOs; interest on these instruments should be slightly higher than on the deposit facility if a positive term premium exists.
    - Likewise, in structural deficits, funds in the deposit facility would need to be injected via central bank term-lending or OMOs typically of longer maturity and higher interest.
    - Compared to a well-managed mid-corridor system with low excess reserves, moving to a floor system could in fact result in increased central bank profit.

*IMF Technical Assistance Report — Annex V. The Use of Interest Rate Corridors in Monetary Policy Implementation*

### References

### References

### Citations and Working Papers
- Bindseil, Ulrich and Juliusz Jabłecki. The Optimal Width Of The Central Bank Standing Facilities Corridor And Banks’ Day-To-Day Liquidity Management. ECB Working Paper Series No 1350, 2011.
- Gray, Simon, Nick Talbot. Monetary Operations. London: Bank of England, Handbooks in Central Banking no. 24, 2006.
- IMF. Evolving Monetary Policy Frameworks in Low-Income and Other Developing Countries. Washington D.C.: International Monetary Fund, Background Paper No. 15/172, 2015a.
- —. Evolving Monetary Policy Frameworks in Low-Income and Other Developing Countries—Background Paper—Country Experiences. Washington D.C.: International Monetary Fund, Background Paper No. 15/173, 2015b.
- Kahn, George A. Monetary Policy under a Corridor Operating Framework. Kansas City: federal reserve bank kansas city, Economic Review 2010 Q4, 2010.
- Laurens, Bernard. The Use of Standing Facilities in Indirect Monetary Management: Country Practices and Operational Features. Washington D.C.: International Monetary Fund, MAE Operational Paper OP/97/4, 1997.
- Laurens, Bernard J. Kelly Eckhold, Darryl King, Nils Maehle, Abdul Naseer and Alain Durré. The Journey to Inflation Targeting: Easier Said than Done. The Case for Transitional Arrangements along the Road. Washington D.C.: International Monetary Fund, Working Paper No. 15/136, 2015.
- Mæhle, Nils. Monetary Policy Implementation: Operational Issues for Countries with Evolving Monetary Policy Frameworks. Washington D.C.: International Monetary Fund, Working Paper WP/20/26.
- Vavra, David, Barry Topf, Zsolt Kondrat. Framework for Developing Money Markets in Frontier and Emerging Market Economies. EBRD, forthcoming.

### Footnotes and Empirical Observations
- In New Zealand, the spread between the policy rate and the 30-day bank bill averaged 17bps in 2010-14 when they operated a two-tier floor system.
- There can also be a persistent small positive spread between the central bank deposit rate and the overnight interbank rate under a floor system. In the case of Norway, interbank rates were on average around 20-30 bps above the deposit rate in 2010-2011 when they operated a conventional floor system.
- See Goodfriend (2002) for an early discussion of this in the case of the US, and Keister, Martin, and McAndrews (2015) for a fuller theoretical discussion.

*IMF Technical Assistance Report | References*

### Annex VI. Key Features of a Liquidity Monitoring and Forecasting Framework

### Objectives of Liquidity Forecasting
- Main purpose: produce short-term liquidity forecasts to generate information allowing the central bank to smooth changes in liquidity conditions.
- Purpose of smoothing:
  - Create stable liquidity conditions and steer the central bank’s operating target.
  - Help market participants distinguish between changes in monetary policy stance and temporary “noises.”
  - Clear signaling reduces market participants’ uncertainties about the central bank’s policy intentions, contributing to efficient monetary policy transmission.
  - Stable liquidity conditions facilitate banks’ liquidity management by moderating volatility and uncertainty, reducing liquidity management costs and limiting settlement risks.
- Liquidity forecasting centralizes all relevant information that determines the future stance of liquidity without central bank activities; central banks then decide on how much liquidity to add or withdraw.
- Liquidity forecasts are necessary whether the operating target is an interest rate or a quantity (monetary base or bank reserves); the forecasting process is the same but the use of the information differs.

### Liquidity Forecasting and Monetary Policy Instruments
- Central banks using direct monetary policy instruments have no immediate need for liquidity forecasts because they exert direct control over liquidity by controlling interest rates, directing credit, or setting bank-to-bank credit ceilings.
- Transitioning from direct to market-based indirect instruments requires a thorough liquidity forecasting framework, including institutional and organizational setup.
- Certain indirect instruments can reduce liquidity fluctuations:
  - Reserve requirements with averaging provisions and standing facilities can limit interest rate volatility.
  - Reserve averaging acts as a buffer, allowing banks to over- and under-fulfill reserve requirements during the maintenance period; interest rate elasticity of demand for reserves tends to be higher with reserve averaging, reducing fluctuations in money market rates from liquidity supply shocks.
- Interest rate corridor (IRC) context:
  - Standing facilities set an upper and lower boundary for interbank money market rates with the same maturity.
  - A narrow IRC limits market volatility but disadvantages include:
    - (i) liquidity injected or absorbed at the initiative of banks rather than the central bank;
    - (ii) interbank activity discouraged, negatively affecting market deepening;
    - (iii) reduced ability of the central bank to extract information on liquidity conditions and banks’ expectations from the money market.
  - Most interest rate targeting central banks conduct main operations at their initiative and aim at an IRC wide enough to encourage interbank transactions.
  - Liquidity forecasting consolidates expected liquidity conditions and helps avoid unwarranted excessive volatility.

### Liquidity Forecasting and the Payment System
- Payment system features affect the demand and effective supply of bank reserves.
- If the payment system is undeveloped and inefficient:
  - Banking system typically holds large, often highly volatile amounts of excess reserves.
  - There may be large and unstable reserve floats.
  - Both deficiencies complicate liquidity forecasting, weaken central bank control of liquidity supply, and hamper monetary management.
- Quality of liquidity forecasts improves with payment system efficiency.

### Determinants of Liquidity
- Liquidity forecasting analyzes projected changes in main items of the central bank’s balance sheet.
- Main components:
  - Autonomous supply of liquidity (short-run beyond central bank control): (i) net foreign assets; (ii) net position of the government with the central bank; (iii) currency in circulation; (iv) other items net.
  - Liquidity demand: demand for required and excess reserves.
- Forecasting requires identifying determinants and selecting appropriate techniques.
- Net position of the government with the central bank often accounts for the most significant changes in autonomous liquidity supply; close government-central bank cooperation on timely cash flow projections (revenues, expenditures, funding instruments) is crucial.
- In countries with heavily managed exchange rates and large/volatile capital flows, net foreign assets can drive significant liquidity supply changes; in the very short run, variations in net foreign assets are normally known with certainty due to typical two-day settlement lags.
- Currency projections can be highly uncertain under certain circumstances, e.g., after a period of high inflation.

### Liquidity Monitoring and Forecasting Framework (Organization and Process)
- A clear and efficient organization is essential for accurate, timely projections.
- Forecasts should be a rolling process incorporating every new piece of information promptly.
- Recommended organization: assign forecasting responsibility to a dedicated unit in the monetary operations department to ensure close linkage between forecasting and liquidity management. The unit should:
  - (i) communicate with information sources and ensure timely data receipt;
  - (ii) supervise consistency of forecasted components;
  - (iii) produce an overall liquidity projection regularly (daily) updated;
  - (iv) assess forecasting errors.
- Forecasting horizons and intervals differ by institutional setup:
  - With reserve requirements: horizon should comprise at least the current maintenance period.
  - Without reserve requirements: horizon should comprise at least the time between two discretionary monetary interventions.
  - Optimal frequency: daily forecasts; if data unavailable, start with weekly intervals.

### Conceptual Background
- Demand and supply of liquidity derived from central bank balance sheet; simplified stylized balance sheet:
  - Assets: Net foreign assets; Net lending to banks/OMOs 1/; Other items net/Net domestic assets.
  - Liabilities: Currency in circulation; Bank reserves (required and excess); Net position of the government 2/.
  - Footnotes:
    - 1/ In countries where there is a structural excess liquidity, the net “policy” position with banks can be on the liability side, for instance when the central bank absorbs liquidity by issuing its own securities.
    - 2/ In countries where the central bank is allowed to provide direct credit to the government, the net position of the government may appear on the assets side.
- Demand for bank reserves = Required reserves + excess reserves.
- Key determinants of reserve demand: institutional relationship between central bank and banking system; degree of money market development; variability and timeliness of payments; banks’ expectations about central bank liquidity management; characteristics of reserve requirements; design and efficiency of payment system; access to central bank credit.
- Institutional factors more relevant for long-run level of liquidity demand and must be considered in case of structural changes.
- Liability side: currency and bank reserves (uses of base money) mirrored by four sources on asset side; supply of bank reserves = autonomous liquidity position + policy position.
- Autonomous items are factors beyond short-run central bank control; policy position comprises central bank direct lending to banks and net lending through OMOs, under immediate control.
- Comparing forecasted demand with projected autonomous supply yields estimated excess supply or demand if the central bank does not intervene; based on this, the central bank adjusts its net policy position.
- Net policy position accounts for past policy operations that come into effect later (e.g., redemption of repo/reverse repo, maturing refinance credit).
- How the central bank adjusts depends on operating target (quantity target or interest rate target) and time horizon of target:
  - If targeting overnight interest rate:
    - Central bank uses projected daily excess supply/demand to define daily injections/withdrawals.
    - Relationship is less strict if steering rate within a corridor or with reserve averaging; central bank may only react to cumulated disequilibria over the maintenance period.
    - Without reserve requirements or reserve averaging, central bank may need day-to-day liquidity management.
  - If targeting a quantitative variable (reserve money or bank reserves):
    - Central bank adjusts supply to confirm with target level; remaining excess demand or supply is left for interbank money market rate to adjust.
    - If demand for bank reserves is interest inelastic (e.g., end of reserve period), central bank may need to accommodate demand to avoid financial system disturbances even if this leads to short-term deviations from the quantitative operating target.
- Footnote on NFA categorization:
  - NFA are categorized as autonomous regardless of exchange rate regime, assuming under flexible exchange rates the central bank does not intervene in the very short run and NFA are constant; under fixed exchange rates NFA changes are outside control because central bank committed to intervene to hold the exchange rate stable.

*IMF Technical Assistance Report | Annex VI — Key Features of a Liquidity Monitoring and Forecasting Framework*

### Annex VII. Balance Sheet Composition

### NRBT Balance Sheet Observations
- NRBT asset and liability distribution is uneven with the share dominated by Bank A and Bank B.
- The vast majority of NRBT assets are bank deposits and loans and advances.
- On the liability side, composition is predominantly demand deposits.

### Figures (Descriptive)
- Figure E.1: NRBT Asset Distribution (categories shown in chart: Cash (Notes & Coins); Bank Deposits; Government Bonds; Loan and Advances; Accrued Interest; Shares and Equities; Non Financial Assets; Other Assets) with distribution across Bank A, Bank B, Bank C, Bank D.
- Figure E.2: NRBT Liabilities (categories shown in chart: Demand Deposits; Savings Deposits; Time Deposits; Accrued Interest Payable; Payable Interest (Loans and Advances); Other Liabilities) with distribution across Bank A, Bank B, Bank C, Bank D.

*IMF Technical Assistance Report | Annex VII — Balance Sheet Composition*

### Annex VIII. Fixed-income Security Valuation and

### Annex VIII. Fixed-income Security Valuation and Haircut Methodology

### A. Fixed-income Security Valuation — key points
- Theoretical valuation is used to price collateral when no reliable market prices are available.  
- For fixed income securities, central banks rely on yield curves to compute a price for a hypothetical zero-coupon used for discounting cashflow payments.  
- The approach is based on arbitrage-free theory: the discount factor for a specific asset type is inferred from observed market prices of assets with the same characteristics.  
- Calibrating a yield curve disciplines parameters of an analytical model by means of observed bond market prices.  
- Formal bond pricing notation in the source:
  - If zBond(t) is zero rate at term t, C the coupon rate and t1,...,tN the coupon dates, the bond theoretical price is given by:
    - PB = sum_{i=1 to N} Prisky P(0,ti) C + Prisky(0,tn), where Prisky(0,t) = e^{−t * zBond(t)}.  
  - The zero rate zBond(t) is derived from the sovereign curve of the Government of Namibia, which is calibrated as below.

### A Model for Yield Curve Calibration — method and objective
- Observed market prices vector: m⃗⃗⃗ = (PM(B1) … PM(Bn))′.  
- Theoretical prices vector given a yield curve z: t⃗ = (P(B1; z) … P(Bn; z))′.  
- Vector of absolute pricing errors (∈A⃗⃗⃗⃗⃗) defined by: ∈A⃗⃗⃗⃗⃗ = t⃗ − m⃗⃗⃗.  
  - Note: “absolute” is used to mean the actual value of the difference, not the absolute value.  
- Vector of relative pricing errors (∈R⃗⃗⃗⃗⃗):
  - ∈R⃗⃗⃗⃗⃗ = ((P(B1; z) − PM(B1))/PM(B1) … (P(Bn; z) − PM(Bn))/PM(Bn))′.  
- Vector of squared relative pricing errors (∈Sq⃗⃗⃗⃗⃗⃗⃗) = (∈R,1^2 … ∈R,n^2)′.  
- Calibration objective: minimize the cost function F(z) measuring the squares of sum of squared relative pricing errors:
  - F(z) = ‖ ∈Sq⃗⃗⃗⃗⃗⃗⃗ ^T ‖.  
  - Find ẑ so that ẑ = arg min_z F(z).  
- Nelson-Siegel types of yield curves are used to assume a parametric form for z.

### Nelson-Siegel family — functional forms and parameter interpretations
- Original Nelson-Siegel (NS) parametrical function:
  - NS(τ; β0, β1, β2, λ0) = β0 + β1 * (1 − e^{−τ/λ0}) / (τ/λ0) + β2 * ((1 − e^{−τ/λ0})/(τ/λ0) − e^{−τ/λ0}), where τ is a given time interval and (β0, β1, β2, λ0) are parameters.  
- Monotone (constrained) form obtained by applying β2 = −β1, yielding:
  - NS(τ; β0, β1, −β1, λ0) = NS_monotone(τ; β0, β1, λ0) = β0 + β1 * e^{−τ/λ0}.  
- Six-parameter Nelson-Siegel-Svensson model:
  - NS(τ; β0, β1, β2, β3, λ0, λ1) = β0 + β1 * (1 − e^{−τ/λ0})/(τ/λ0) + β2 * ((1 − e^{−τ/λ0})/(τ/λ0) − e^{−τ/λ0}) + β3 * ((1 − e^{−τ/λ1})/(τ/λ1) − e^{−τ/λ1}).  
- Table F.1 — interpretation of Nelson-Siegel parameters (source: Mission’s elaborations):
  - β0: independent of time to maturity; often interpreted as the long-run yield level.  
  - β1: weighted by a function of time to maturity; influences the short end of the curve (unity at λ = 0, decays as λ grows).  
  - β2: weighted by a function of λ; zero for λ = 0, increases then decreases as λ grows; adds a hump to the curve.  
  - λ (lambda): affects weight functions for β1 and β2; determines the position of the hump.

### B. Haircut Methodology — key principles
- Haircuts are a key risk mitigation measure calibrated to achieve “risk equivalence”, so the central bank faces the same risk across all asset classes.  
- Haircuts must account for liquidity, market and credit risks with add-ons for specific risks: residual maturity, exchange-rate risk, valuation risk (model uncertainty), and mobilization risk (if assets are difficult to gain title to and liquidate).  
- The outlined approach in Box F.1 uses Value-at-Risk (VaR) methodology; noted that VaR is likely not workable in Tonga because of paucity of pricing. An alternative requires relative assessment of standalone risk components.  
- Final haircut calibration should benefit from additional market risk intelligence and be approved by the Board of the central bank and reviewed periodically.

### Box F.1 — A Primer on Haircut Estimation (method and numeric example)
- Concept:
  - In case of counterparty default, collateral must be sold over a liquidation time T2L > 0.  
  - Cash equivalent after haircut: (1 − h) · P0, where P0 is collateral value before risk mitigation.  
  - Haircut h should depend on price volatility and expected T2L; may be set to cover 99 percent of price changes within liquidation time. Additional add-on may address valuation uncertainty.  
- Two-asset illustrative example to achieve risk equivalence:
  - Asset A:
    - Traded regularly with transparent prices. Orderly liquidation immediate (one-week horizon).  
    - One-week price change distributed N(0, σ_{M,A}^2).  
  - Asset B:
    - Not traded regularly; theoretical value calculated daily. Orderly liquidation assumed 4 weeks (T2L = 4 weeks).  
    - Four-week market price change distributed N(0, σ_{M,B}^2).  
    - Valuation uncertainty N(0, σ_{V,B}^2).  
    - Spread and migration risk during liquidation N(0, σ_{S,B}^2).  
    - Assume independence across risk factors; total uncertainty variance for asset i (i = A,B) denoted σ_{T,i}^2.  
  - If risk tolerance defined as “preventing with 99 percent probability that the asset value of the liquidation falls short of the last valuation post-haircut”, haircuts set to:
    - h_i = σ_{T,i} Φ^{-1}(0.01), where Φ(.) is the cumulative standard normal distribution.  
  - Numerical example:
    - σ_{M,A}^2 = 1%  
    - σ_{M,B}^2 = 4%  
    - σ_{V,B}^2 = 2%  
    - σ_{S,B}^2 = 2%  
    - Resulting haircut array establishing risk equivalence: H = { h1; h2 } = { 2.33%; 6.60% }.  
- Methodological notes and references cited in source (no external material reproduced here other than as named in the text).

### Operationalizing haircuts for environments with sparse pricing
- The approach in Box F.1 (VaR-based) may not be workable where pricing is scarce; use a relative assessment of standalone risk components instead.  
- Table F.2 (Tentative Standalone Risk Assessment, source: Mission’s elaborations) illustrates operationalization via relative risk assessment among asset classes across risk dimensions. Implementation notes from the source:
  - At aggregated balance-sheet level, NRBT Notes considered most liquid; liquidity risk set to zero.  
  - Liquidity risk increases across asset classes from government securities, Government Loans, Loans and Advances (corporate and households), Shares and Equities, to Non-financial Assets.  
  - Same logic applies for market and credit risk, reflecting asset volatility and debtor-default risk respectively.  
  - Applicable add-ons include residual maturity, exchange rate, theoretical valuation, and mobilization risks.  
  - Example: for government securities, maturity add-ons can increase by 2.5 percent with the maturity brackets; mobilization risk for government securities is zero.

*Source: Mission’s elaborations.*

---


_Source: https://www.imf.org/-/media/files/publications/tar/2025/english/tarea2025065-print-pdf.pdf_
