Monetary Policy Impact

1. Global Landscape — Monetary Policy Design Choices

This guidance synthesises the tier-4 design layer (holding limits, interest-bearing decisions, programmability) and the disintermediation / financial-stability framing across all 20 jurisdictions tracked in this domain: six in the Americas (

United States,

Bahamas,

Brazil,

Canada, Eastern Caribbean,

Jamaica), seven in Asia-Pacific (

Australia,

China,

India,

Japan,

Singapore,

South Korea,

Thailand), four in Europe (

EU,

Russia,

Sweden,

UK), and three in the Middle East & Africa (

Ghana,

Nigeria,

South Africa).

Status distinction (of 20). Precise positioning matters for monetary-policy analysis, because disintermediation risk only becomes empirical once a CBDC is in public hands:

• Launched / live (6): Bahamas (Sand Dollar), Nigeria (eNaira), Jamaica (JAM-DEX), Eastern Caribbean (DCash — pilot concluded Jan 2024, DCash 2.0 successor suspended Feb 2026), China (e-CNY, pilot-scale and operational at scale), India (e₹, retail + wholesale pilots).
• Active / restarted pilots & design (7): Brazil (Drex — Besu platform decommissioned, rebuilding), Russia (digital ruble — narrow live government/budget use since 1 Jan 2026, full launch mandated 1 Sep 2026), South Korea (Project Hangang deposit-token model, Phase 2 from Mar 2026), Australia (Project Acacia wholesale), Singapore (live SGD wholesale CBDC pilot), Thailand (retail pilot concluded, wholesale active), Ghana (eCedi pilot declared "completed" 22 May 2026).
• Research / monitoring (6): Canada (retail scaled down Sep 2024), Japan (decision pending 2026), Sweden (research, contingent on the digital euro), EU (advanced pre-issuance, legislation pending), South Africa (research/PoC, "no compelling immediate need"), plus the wholesale-leaning Asia-Pacific programmes above where retail is shelved.
• Anti-CBDC (1): United States — only major economy to formally bar CBDC development (Executive Order 14178, in force). A statutory moratorium passed both chambers in differing, unreconciled forms and is NOT enacted; only EO 14178 is binding. New Fed Chair Kevin Warsh (sworn in 22 May 2026) reaffirmed opposition to a retail CBDC.

1.1 Holding Limit Approaches

The calibration of individual holding limits is the single most consequential monetary policy design decision in retail CBDC architecture. Every jurisdiction with a launched or advanced-design retail CBDC has adopted or proposed holding limits (or a structurally equivalent cap) as the primary safeguard against bank disintermediation, but the approaches vary dramatically in magnitude and mechanism. Wholesale-only programmes (Australia, Singapore wholesale, South Africa, Thailand wholesale) apply no retail holding limit by construction.

European Union — EUR 3,000 per person (proposed). The European Commission's impact assessment for the digital euro modelled a holding limit of EUR 3,000 per individual. The ECB tested scenarios at EUR 500, 1,000, 2,000, and 3,000. At the EUR 3,000 threshold, the ECB estimated a maximum deposit outflow of EUR 699 billion under a flight-to-safety scenario — approximately 8.2% of retail sight deposits. Under business-as-usual conditions, just over EUR 100 billion would leave bank balance sheets — less than 2% of retail sight deposits. The Council position (December 2025) empowers the ECB to set the limit within a Council-agreed ceiling, reviewed at least every two years; businesses and public authorities have a zero limit (waterfall to bank account).

• Source: ECB — Technical data on financial stability impact (October 2025)
• Source: ECB Occasional Paper No. 326 — Know your (holding) limits
• Source: Bruegel — On the digital euro holding limits

United Kingdom — GBP 10,000–20,000 per individual (proposed). The Bank of England and HM Treasury proposed this range in their February 2023 consultation (unchanged in the 2025 design phase). Financial Stability Paper No. 53 (November 2025) modelled deposit outflow scenarios under severe stress:

The Treasury Select Committee recommended limits closer to the

EU's EUR 3,000, arguing the proposed range was too high. The BoE has described limits as "transitional safeguards" that could be loosened over time.

• Source: BoE Financial Stability Paper No. 53 (November 2025)
• Source: UK Finance — Understanding the Bank of England's latest digital pound analysis

China — Tiered wallet system. The e-CNY employs tiered wallet caps calibrated to KYC strength rather than a single universal limit:

This tiered approach enables differentiated risk management: the most anonymous wallets have the lowest caps, while full-KYC wallets permit substantially higher holdings. Critically, the 1 January 2026 reclassification of e-CNY held in commercial-bank wallets as bank deposit liabilities means the balance caps no longer serve a disintermediation function in the traditional sense — the funds remain on bank balance sheets regardless.

• Source: PBOC White Paper — Progress of Research & Development of E-CNY in China (July 2021)
• Source: DL News — 10 Chinese banks to pay interest on digital yuan (2026)

India — INR 1,00,000 per wallet (~US$1,200). The RBI set a flat holding cap with additional transaction constraints: INR 10,000 per transaction, INR 50,000 daily cumulative, and a maximum of 250 token denominations. These limits are among the most restrictive globally, reflecting the RBI's cautious approach and the early-stage nature of the pilot.

• Source: Axis Bank Digital Rupee product page; Kotak Mahindra Bank CBDC FAQ

Nigeria — Tiered wallet balance caps (eNaira). From the 2021 eNaira Regulatory Guidelines (no 2025–26 amendment found): Tier 0 N120,000 ($75); Tier 1 N300,000 ($187); Tier 2 N500,000 ($312); Tier 3 N5,000,000 ($3,125). Daily transaction limits range N20,000–N2,000,000. These caps are tied to KYC tier rather than disintermediation modelling.

• Source: IMF Working Paper WP/23/104 — Nigeria's eNaira, One Year After (May 2023)

Bahamas — Tiered wallet caps (Sand Dollar). Tier 1: B$500 balance / B$1,500 per month; Tier 2: B$5,000 balance / B$10,000 per month (B$100k/yr); Tier 3: B$8,000 (or 1/20th of annual revenue) / B$20,000 per month. Maximum B$8,000 personal holding. Non-interest-bearing.

Eastern Caribbean — DCash tiered caps (pilot). Registered Wallet: EC$3,000–EC$20,000 (set by the distributing FI); Value-Based Wallet: EC$1,000–EC$2,700 per month. Merchant limits NOT VERIFIED. (DCash 2.0 successor suspended Feb 2026; these were the pilot parameters.)

Russia — RUB 300,000 monthly top-up limit (digital ruble). The Central Bank of Russia applies a RUB 300,000 monthly replenishment (digitisation) cap on individual wallets — incoming platform transfers can exceed it, and there are no caps on wallet-to-wallet transfers; business/legal-entity wallets are uncapped. The CBR explicitly frames the top-up limit as a bank-disintermediation safeguard, supplemented by non-remuneration and its commitment to "fully offset" liquidity outflows.

• Source: Carnegie Endowment — Can the Digital Ruble Shield Russia From Western Sanctions? (October 2024)

South Korea — Pilot caps (Project Hangang deposit tokens). Phase 1 imposed a 1 million won (~$689) balance cap at any one time and 5 million won cumulative. Production digital-won holding limits are not publicly disclosed; because the model is wholesale-CBDC-backed commercial-bank deposit tokens, interest policy sits with the issuing banks.

Provider-set or no published retail limit.

• Jamaica (JAM-DEX): The BOJ imposes no minimum/maximum wallet limits — providers set their own (e.g., JN Pay basic J$50,000; up to J$100,000/day with additional KYC).
• Brazil (Drex): No specific Drex holding limit published; rDrex are tokenized bank deposits subject to the FGC deposit guarantee (NOT VERIFIED as a distinct CBDC cap).
• Ghana (eCedi): BoG has not published holding limits or wallet tiers — NOT VERIFIED.
• Thailand (retail pilot): Not publicly disclosed (NOT VERIFIED); the Deposit Protection Act benchmark of 1 million baht per depositor per institution is the relevant tokenized-deposit reference.
• Japan: Not publicly disclosed; BOJ is testing "restricting holdings"/"restricting transactions" with differentiated caps for corporate vs P2P; holding limits deemed essential by the Japanese Bankers Association.
• Sweden: Potential holding caps to limit deposit migration are under consideration but not decided.
• Canada, United States, Australia, Singapore (retail), South Africa: Not applicable — no retail CBDC issued or pursued (Canada's research, SWP 2024-4, finds holding limits effectively prevent deposit crowding-out).

1.2 Interest-Bearing vs Non-Remuneration Decisions

The question of whether a CBDC should bear interest is among the most debated in monetary economics. As of June 2026, the global consensus strongly favours non-remuneration for retail CBDCs, with one notable exception — China.

Non-remunerated / no interest (the overwhelming majority):

Interest-bearing CBDC — China's 2026 break with orthodoxy (the sole exception):

Effective 1 January 2026,

China became the first major economy to offer interest on CBDC holdings. Commercial banks pay an initial annual rate of approximately 0.05% on verified e-CNY wallet balances (Categories 1–3), with the first quarterly payout occurring on 20 March 2026.

At 0.05%, the yield is deliberately negligible and is designed to collapse the distinction between e-CNY wallets and conventional demand deposits rather than to attract savings. The concurrent reclassification of e-CNY held in commercial-bank wallets as bank deposit liabilities under the reserve requirement framework means the interest-bearing feature functions as a deposit-integration mechanism, not a conventional monetary policy tool.

PIIE analysts noted this reclassification means "much of the e-CNY is no longer a CBDC" under standard definitions. CNBC reported the shift to interest-bearing status "shows previous efforts to promote digital yuan didn't really work."

• Source: BeInCrypto — China Breaks CBDC Orthodoxy: Digital Yuan to Pay Interest Starting 2026
• Source: PIIE — China gives up on state-backed digital cash (2026)
• Source: CNBC — Interest-bearing e-CNY shows previous efforts didn't really work (January 2026)

IMF framework on remuneration design. IMF Fintech Note 2023/010 (Das, Mancini-Griffoli et al.) established that:

• Non-interest-bearing CBDC with quantity caps has limited monetary policy impact

• Interest-bearing CBDC could strengthen the transmission of policy rate changes

• Design choices on remuneration, holding limits, and access determine the magnitude of monetary policy effects

• Tiered remuneration (lower rate for store-of-value holdings) can mitigate disintermediation while preserving monetary policy stance

• Source: IMF Fintech Note 2023/010 — Implications of CBDCs for Monetary Policy Transmission

1.3 Programmability and Conditional Payments

Programmability — the ability to attach spending conditions, expiry, or purpose-binding to CBDC — is a recurring monetary-policy lever and political flashpoint. The matrix shows three distinct postures across the 20:

Programmability built and used:

• India (e₹): Purpose-bound money expanded 30 Aug 2024 (fuel, groceries, education, healthcare, dining, travel); user-level programmability (HDFC, 29 Aug 2024); welfare schemes (Gujarat G-SAFAL, AP DEEPAM 2.0, Odisha Subhadra, Puducherry PMGKAY food-currency pilot). The most operationally advanced programmable retail CBDC.
• Russia (digital ruble): Smart-contract "coloured funds" — 17,000+ smart contracts in the pilot; programmable spending restrictions for targeted budget funds (Tatarstan, Chuvash).
• China (e-CNY): Conditional payments, time-limited vouchers, purpose-binding, automated triggers — but the PBOC controls permitted programmability; no DeFi, no public blockchain explorer.
• South Korea (Hangang): Purpose-Bound Money model — children's pocket money, EV charging subsidies (world-first national-treasury digital-currency operation), childcare, culture/regional vouchers.
• Singapore (PBM): Purpose-Bound Money is the flagship innovation, distinguishing programmable payments, programmable money, and purpose-bound money; wraps CBDC, tokenized deposits, or stablecoins.
• Thailand: Programmable Payment Sandbox (SCB 10X "Purpose Bound Money", KBANK; second-round True Money/Bitkub/OM Platform from Apr 2026).
• Australia (Acacia/eAUD): Atomic settlement, embedded compliance, conditional payments, escrow tested.
• Bahamas (Sand Dollar): Issuer can set expiration dates or assign specific fund-usage rules.
• Sweden: Phase 3 confirmed feasibility of programmable payments and smart contracts.

Programmability constrained or excluded by policy:

• EU (digital euro): "Programmable money" explicitly excluded ("would never be programmable money"); conditional payments supported only as PSP value-added services off the core ledger.
• UK (digital pound): No government/central-bank programmability (to be guaranteed in primary legislation); private-sector programmability with user consent only — a 31,476-signature petition opposed programmable CBDC.

Not enabled / not verified: Nigeria (smart-contract capability in architecture but unenabled after launch); Brazil (originally the key differentiator, now narrowed post-pivot to collateral/lien reconciliation); Jamaica, Ghana (NOT VERIFIED); Japan (under research); South Africa (wholesale/sandbox exploration only); Canada, US (not applicable).

1.4 Waterfall and Reverse Waterfall Mechanisms

The

EU's digital euro design introduced a critical innovation for managing holding limits without disrupting user experience: automated waterfall/reverse waterfall mechanisms.

Waterfall mechanism: When a payee receives a digital euro payment that would push their balance above the holding limit, the excess automatically transfers ("falls") to the user's linked commercial bank account. This prevents involuntary breaches and ensures the digital euro flows back to the banking system in real time. Businesses and public authorities (zero limit) rely on waterfall for every receipt.

Reverse waterfall mechanism: When a payer lacks sufficient digital euros for a payment, the shortfall is automatically topped up from their linked bank account. This ensures usability even when wallet balances are low.

The ECB's preparation phase (concluded October 2025; Governing Council moved the project to its next phase 29–30 October 2025) confirmed these mechanisms work at scale while preserving privacy and scalability. The design has been closely studied by the

UK and other jurisdictions as a potential template.

• Source: ECB — Preparation phase closing report (October 2025)

No other jurisdiction has yet implemented or formally proposed waterfall mechanisms at the same level of specificity. China's tiered wallet system and India's fixed caps do not include automatic overflow/underflow features — exceeding a cap simply blocks the transaction or requires manual wallet upgrading. Russia's RUB 300,000 cap applies only to top-ups, leaving incoming platform transfers uncapped.

1.5 M0 vs Deposit-Money Classification

The classification of CBDC within the monetary aggregates has direct implications for central bank balance sheets, reserve requirements, and the seigniorage model.

M0 (cash-equivalent) classification — the standard approach:

Reclassification to deposit money — China's 2026 shift:

China's e-CNY was originally designed as M0 (cash equivalent). The PBOC Action Plan (published 29 Dec 2025, effective 1 Jan 2026) reclassified e-CNY held in commercial-bank wallets as bank deposit liabilities, integrating them into the reserve requirement framework. This effectively shifts the e-CNY from M0 toward M1 (demand deposits), blurring the boundary between CBDC and conventional bank deposits.

Implications:

• e-CNY no longer reduces bank deposit bases — it appears on bank balance sheets as a liability

• The PBOC incorporates e-CNY wallet balances into reserve calculations

• Deposit insurance applies to e-CNY holdings in commercial bank wallets

• The PBOC's balance sheet shrinks relative to a pure M0 model

• Source: State Council of the PRC — China to enhance digital yuan management (December 2025)

• Source: China Daily — China to usher in digital deposit money era for e-CNY (December 2025)

Tokenized-deposit / deposit-token models (Brazil, South Korea):

Brazil's Drex uses a tokenized-deposit model: retail Drex (rDrex) represents tokenized commercial bank deposits rather than direct central bank liabilities, while wholesale Drex (wDrex) is a BCB liability for interbank settlement. The BCB explicitly designed this so the central-bank balance sheet does not change and bank intermediation is preserved; tokenized deposits benefit from the FGC deposit guarantee.

South Korea's Project Hangang takes a structurally similar route — a wholesale-CBDC settlement layer underpinning won-pegged commercial-bank deposit tokens (1:1 convertible), which the BOK itself describes as "an intermediate stage between a CBDC and stablecoins" rather than a full retail CBDC.

Singapore's retail Purpose-Bound Money likewise runs on commercial-bank deposits, not a direct MAS liability.

• Source: Wharton Initiative on Financial Policy — Lessons from Brazil

2. Disintermediation Risk Analysis

2.1 The Core Problem: What Happens to Bank Deposits When CBDC Launches?

Bank disintermediation — the migration of retail deposits from commercial banks to central bank money — is the preeminent financial stability concern with retail CBDCs. The risk operates through two channels:

Steady-state channel (gradual substitution): Over time, households and businesses shift a portion of their demand deposits into CBDC wallets. This reduces banks' cheap deposit funding base, forcing them to replace it with more expensive wholesale funding or central bank facilities. The cost of credit rises.

Crisis channel (deposit flight / digital bank runs): During financial stress, depositors rapidly convert bank deposits to CBDC — the safest possible asset (central bank money). Unlike physical cash withdrawal, CBDC conversion can happen instantly and at scale, potentially accelerating bank runs beyond anything experienced in the pre-digital era.

A central design insight running through the matrix is that the tokenized-deposit / deposit-token architectures (Brazil's rDrex, South Korea's Hangang deposit tokens, Singapore's retail PBM, and post-2026 China) sidestep the steady-state channel by construction: the money never leaves the banking system, so there is no deposit base to drain. The trade-off is reduced monetary-policy leverage and classification ambiguity.

2.2 ECB Modeling: EUR 699 Billion Maximum Outflow

The ECB's October 2025 technical assessment provides the most detailed publicly available modeling of CBDC disintermediation risk. Using supervisory data from over 2,000 banks and newly constructed models:

Business-as-usual scenario (EUR 3,000 limit):

• Total outflows: ~EUR 100 billion (just over)
• Percentage of retail sight deposits: less than 2%
• Average bank ROE impact: ~30 basis points
• Assessment: manageable; well within regulatory liquidity assumptions

Flight-to-safety scenario (EUR 3,000 limit):

• Total outflows: EUR 699 billion
• Percentage of retail sight deposits: ~8.2%
• Percentage of total banking sector assets: ~2.2%
• Small institutions disproportionately affected
• Assessment: severe but contained by holding limit design

ECB conclusion: "The impact of the digital euro would not harm financial stability within the euro area, even under a highly unlikely and extremely conservative crisis scenario."

The analysis confirmed that the combination of holding limits and non-remuneration provides effective protection. Importantly, the EUR 699 billion figure assumes every eligible person in the eurozone simultaneously fills their wallet to the limit — an extreme scenario.

• Source: ECB — Technical data on financial stability impact (October 2025)
• Source: Reuters/Investing.com — Digital euro could drain up to 700 billion euros of deposits in bank run

2.3 Bank of England Modeling: GBP 250 Billion Without Limits

The BoE's Financial Stability Paper No. 53 (November 2025) modelled the UK digital pound under severe illustrative stress scenarios:

Without holding limits:

• Additional central bank lending required: ~GBP 250 billion
• Firms falling below 100% Liquidity Coverage Ratio: 21%
• Outflow rate assumptions: 75% for individual uninsured deposits; ~10% for insured deposits
• Total potential banking sector outflows: ~GBP 500 billion

With holding limits:

• GBP 20,000 limit: ~GBP 112 billion central bank lending; lower LCR breach rate
• GBP 10,000 limit: ~GBP 58 billion
• GBP 5,000 limit: ~GBP 21 billion

Steady-state impact: A sustained shift of deposits to the digital pound could increase interest rates on bank loans by 0.8 percentage points or more.

The BoE emphasized that holding limits are transitional safeguards that could be raised or removed as the financial system adapts. (Note: the BoE's May–June 2026 signals about easing "overly conservative" holding limits concerned systemic stablecoins, not the digital pound's own limits.)

• Source: BoE Financial Stability Paper No. 53 (November 2025)

2.4 Other Modelled Assessments (Canada, Sweden, Russia)

Three further jurisdictions provide modelled disintermediation analysis without a live retail CBDC:

• Canada (SWP 2024-4): Bank of Canada research finds holding limits "effectively prevent deposit crowding-out"; adoption-friction work (SWP 2024-52) shows even a $1.50/month friction drastically cuts uptake, implying low steady-state migration absent subsidies. Retail work was scaled down Sep 2024 with "not currently a compelling case."
• Sweden: The Riksbank estimates e-Krona demand of ~1–3% of Swedish GDP would be manageable; non-remuneration and potential holding caps are the contemplated safeguards.
• Russia: The CBR frames the RUB 300,000 top-up cap plus non-remuneration as its disintermediation defence and has committed to "fully offset any liquidity outflow." Pilot scale remains tiny (~2,500 wallets, June 2025, self-reported); media projections put up to ~RUB 50bn of annual bank revenue at risk and the CBR projects up to 5% of cashless payments within 5–7 years (self-reported projection).

2.5 Evidence from Launched CBDCs

Six jurisdictions have launched retail CBDCs (one now suspended) and provide empirical rather than purely modelled evidence on disintermediation effects:

Bahamas (Sand Dollar — launched 20 October 2020):

• Circulation: B$2,482,045 as of early 2025, representing ~0.39% of physical cash supply

• Adoption: ~138,000 personal wallets (Dec 2024, +16% YoY); e-money penetration remains low

• Disintermediation evidence: Academic research (MDPI, 2025) found the Sand Dollar is "associated with reductions in loan issuance, deposit levels, and deposit account ownership"

• Mitigation: Non-interest-bearing design with tiered balance caps (max B$8,000 personal)

• Assessment: Small but measurable negative impact on commercial-banking deposits in a very small economy. The CBOB acknowledged the structural risk

• Source: MDPI — CBDCs and Liquidity Risks: Evidence from the SandDollar's Impact on Deposits and Loans in the Bahamas (2025)

• Source: Intereconomics — The Evolution of SandDollar (2023)

Nigeria (eNaira — launched 25 October 2021):

• Adoption: 98.5% of wallets inactive after one year (IMF WP/23/104); 0.37% of currency in circulation / N18.31bn as of Feb 2025; only ~2.2 million transactions by July 2024 (all self-reported, flagged)

• Disintermediation evidence: Negligible at current adoption levels. The IMF noted that "eNaira wallets may be perceived, or even effectively function, as a deposit at the central bank, which may reduce demand for deposits in commercial banks"

• Context: Nigeria's banking sector was under separate stress from the naira redesign (2023) and FX crises, making it difficult to isolate eNaira-specific effects

• Assessment: No measurable disintermediation, primarily because adoption was negligibly low. The CBN has since pivoted toward a wholesale eNaira v2.0 and is studying a private stablecoin (cNGN)

• Source: IMF Working Paper WP/23/104 — Nigeria's eNaira, One Year After (May 2023)

Jamaica (JAM-DEX — launched 2022):

• Adoption: ~282,274 registered wallets (end-2024, +7% YoY); ~0.1% of currency in circulation; J$260.05m issued (end-2025)

• Transaction growth: value +550% YoY and volume +267% YoY in 2025 — but driven primarily by existing users transacting more, not new sign-ups (self-reported, BOJ)

• Disintermediation evidence: No specific data published on bank deposit impact; growth remains too small relative to the banking system to produce measurable effects. JAM-DEX is non-interest-bearing, explicitly to reduce the incentive to shift deposits from banks

• New development: JN Bank launched JN Pay (second public wallet) July 2025; Governor Byles (term ends Aug 2026) called slow adoption "a disappointment" and is pushing DTIs to retrofit POS machines

• Source: Silicon Caribe — Bank of Jamaica Pushes Banks to Accelerate Jam-Dex Adoption (March 2026)

• Source: Kansas City Fed — Observations from the Retail CBDCs of the Caribbean

Eastern Caribbean (DCash — pilot Mar 2021 to Jan 2024; DCash 2.0 suspended Feb 2026):

• Adoption at conclusion: 4,000+ accounts, 21 FIs + 10 government agencies, EC$2.45M (~US$906,551) in circulation (~0.16% of currency, Mar 2023) — self-reported
• Disintermediation evidence: None measurable; pilot too small. The ECCB suspended DCash 2.0 development (13 Feb 2026) in favour of a conventional Fast Payment System + CARICOM (CAPS) integration — a deliberate regression from CBDC

China (e-CNY — pilot from April 2020, operational at scale):

• Adoption: cumulative CNY 16.7 trillion (~US$2.37 trillion) and 3.48 billion transactions as of Nov 2025; only ~0.16% of M0 (June 2023); average transaction CNY 428, down from CNY 1,895 in 2023 (self-reported PBOC aggregates)
• Disintermediation evidence: No published disintermediation data; the Jan 2026 reclassification to deposit money means e-CNY in commercial-bank wallets no longer drains the deposit base — it sits on bank balance sheets. PIIE argues the reclassification effectively eliminates the siphon risk (and that e-CNY may no longer qualify as a CBDC)

India (e₹ — retail pilot from Dec 2022):

• Adoption: ~7–8 million retail users (Dec 2025); ~19 banks; e₹-R in circulation ₹771.7 crore (~US$90M) as of 31 March 2026, down ~24% YoY from ₹1,016.5 crore — the first reported circulation contraction (self-reported, RBI AR 2025-26, flagged). e₹-W collapsed from ₹10 crore (Mar 2023) to ₹0.08 crore (Mar 2024)
• Disintermediation evidence: NBER WP 32457 found "limited disintermediation effects" during the pilot; non-interest-bearing design and tight caps (₹1,00,000 wallet, ₹50,000 daily) constrain store-of-value behaviour

Key takeaway from launched CBDCs: Actual disintermediation has been minimal across all launched retail CBDCs, primarily because adoption itself has been minimal — and in India's case is now declining year-on-year. The Bahamas is the only jurisdiction showing measurable negative impact on deposits, and this occurs in a micro-economy where small shifts have outsized effects. The evidence does not validate the extreme disintermediation scenarios modelled by the ECB and BoE — but it also does not invalidate them, because no launched CBDC has achieved the scale at which those risks materialise. Eastern Caribbean's suspension and Nigeria's wholesale pivot underline that the binding constraint to date has been demand, not financial stability.

2.6 IMF/BIS Research on Financial Stability Impact

IMF WP/24/226 — "Central Bank Digital Currencies and Financial Stability: Balance Sheet Analysis and Policy Choices" (2024): Provides a formal balance-sheet framework showing how CBDC issuance affects commercial banks' funding, the central bank's balance sheet, and overall financial stability. Concludes that holding limits, non-remuneration, and interoperability with deposits are the three primary tools for mitigating disintermediation.

• Source: IMF WP/24/226

IMF Fintech Note 2025/008 — "Evaluating the Implications of CBDC for Financial Stability" (2025): Updated framework covering deposit substitution risk, bank run dynamics, and design safeguards. Provides country-calibration methodologies for stress testing.

• Source: IMF Fintech Note 2025/008

BIS Working Paper No. 1280 — "CBDC and banks: disintermediating fast and slow" (2024): Analyses both gradual deposit migration and rapid crisis-driven outflows. Finds that "suitably-designed CBDC — in particular, with holding limits — may induce a desirable welfare effect precisely because it does disintermediate banks, reducing the fragility of a system prone to runs."

• Source: BIS Working Paper No. 1280

BIS Foundational Principles (2020): The first of the three foundational principles for CBDC design states: "A central bank should not compromise monetary or financial stability by issuing a CBDC." This principle underpins all subsequent design work by BIS member central banks.

• Source: BIS — Central bank digital currencies: foundational principles and core features (October 2020)

Bundesbank research (2024): The Deutsche Bundesbank found that the digital euro would strengthen financial stability "within certain limits," provided holding limits and non-remuneration are maintained.

• Source: Deutsche Bundesbank — Will the digital euro strengthen financial stability? (2024)

3. Comparative Table (t4 + Disintermediation Across 20 Jurisdictions)

4. What Works and What Doesn't

4.1 What Works

Holding limits (EU/UK/India/Russia model). The most effective tool for managing disintermediation risk in a direct-liability CBDC. The ECB's modelling demonstrates that even under extreme flight-to-safety scenarios, a EUR 3,000 limit constrains outflows to approximately 8.2% of retail sight deposits — a severe but survivable shock. The BoE's analysis shows that lower limits (GBP 5,000) reduce required central bank lending to just GBP 21 billion. India's tight ₹1,00,000 wallet cap and Russia's RUB 300,000 top-up cap show the same instinct applied in live programmes. The empirical evidence from the Bahamas (where holding limits exist but the economy is small) and the theoretical work from BIS WP 1280 both support the conclusion that well-calibrated limits are the primary safeguard.

Conditions for effectiveness:

• Limits must be set low enough to constrain store-of-value behaviour
• Limits must be high enough to support routine payment needs (a EUR 3,000 limit covers approximately 95% of individual monthly payment needs in the eurozone)
• Automatic overflow mechanisms (waterfall) prevent user friction

Non-remuneration (universal consensus minus China). Zero interest on CBDC holdings removes the primary incentive for depositors to substitute bank deposits for CBDC. When combined with holding limits, non-remuneration effectively confines CBDC to its intended role as a payment instrument. Every launched retail CBDC (Bahamas, Nigeria, Jamaica, India, Eastern Caribbean) and every advanced-design programme (EU, UK, Russia, Sweden-likely, Japan-recommended) uses or favours non-remuneration; only China pays interest, and that is a deposit-integration device rather than a savings inducement. The IMF Fintech Note 2023/010 confirmed that "non-interest-bearing CBDC with quantity caps would have limited monetary policy impact."

Gradual rollout (all active jurisdictions). Every jurisdiction with an active CBDC programme has adopted a phased approach:

• China: pilot cities expanded from 4 (2020) to 26–29+ across 17 provinces, then 22 authorized operators (April 2026) en route to full-scale deployment
• India: 17 banks expanded to 19; ~6 million users growing to ~7–8 million over 2+ years (though circulation contracted in FY25-26)
• EU: investigation phase (2021–2023), preparation phase (2023–2025), PSP pilot (12 months from H2 2027), potential first issuance 2029
• UK: consultation (2023), design phase (2024–2026 with a decision later in 2026), potential build/legislation late 2020s
• Russia: phased mandatory adoption (systemically important banks by Sep 2026, all universal-licence banks by Sep 2027, remaining by Sep 2028)

Gradual rollout allows central banks to observe real adoption patterns, measure actual deposit migration, and calibrate safeguards based on evidence rather than models alone.

Two-tier distribution (universal). All 20 programmes that contemplate any distribution use a two-tier model where the central bank issues CBDC to commercial banks/PSPs, which then distribute to the public. This preserves banks' intermediation role, maintains the commercial relationship between citizens and their banks, and avoids the extreme scenario where the central bank competes directly with the banking system for retail deposits.

Tokenized-deposit / deposit-token model (Brazil, South Korea, Singapore retail, China post-2026). Designs where retail tokens represent commercial bank deposits rather than central bank liabilities eliminate the steady-state disintermediation channel by construction. Funds never leave the banking system; they are simply represented in a new digital form. Brazil's BCB explicitly chose this so Drex "didn't change the balance sheet of the central bank"; South Korea's BOK calls its deposit-token design "an intermediate stage between a CBDC and stablecoins." The trade-off is reduced monetary-policy leverage and classification ambiguity.

"Not launching" as a defensible monetary-stability choice (South Africa, Canada). South Africa's SARB (Nov 2025 Position Paper) concluded a retail CBDC is technically feasible but has "no compelling immediate need," prioritising PayShap and NPS reform; Canada scaled retail work down for the same reason. Both treat declining to issue as the responsible disintermediation-management decision when existing rails already serve the public.

4.2 What Doesn't Work

No limits without analysis. No jurisdiction has launched or proposed a direct-liability retail CBDC without holding limits and survived scrutiny. The BoE's modelling shows that a limitless digital pound could require GBP 250 billion in central bank lending and push 21% of financial firms below regulatory liquidity thresholds. The ECB's EUR 699 billion figure applies even with a EUR 3,000 cap; without any cap, the potential outflow is orders of magnitude larger. Any central bank contemplating a direct-liability retail CBDC without holding limits faces unacceptable financial stability risk.

Forced / mandated adoption without demand (Russia, Nigeria cautionary cases). Russia's mandatory adoption timeline — systemically important banks by September 2026, large retailers (>RUB 120M revenue) required to accept — combined with government payment channels, risks creating adoption without demand. Over 50% of Russians are unwilling to use the digital ruble; ~50% view it as a state-control tool (media-sourced, indicative). Nigeria's eNaira suffered from forced circulation during the 2023 naira redesign crisis, damaging public trust and contributing to the 98.5%-inactive outcome. Mandates create compliance without genuine utility and risk embedding negative perceptions.

Insufficient merchant infrastructure (Jamaica). Jamaica's JAM-DEX demonstrates that without merchant acceptance infrastructure, even a technically functional CBDC has no use case. Governor Byles pushed banks to retrofit POS machines and, in May 2026, called slow adoption "a disappointment." Growth has been driven by existing users, not new adoption. Merchant infrastructure is a prerequisite, not a consequence, of CBDC success.

Launching ahead of a value proposition (Eastern Caribbean, Nigeria). The Eastern Caribbean's DCash and Nigeria's eNaira both reached live status before demonstrating durable demand; the ECCB has now suspended the DCash 2.0 successor (Feb 2026) in favour of a conventional Fast Payment System, and Nigeria has pivoted to a wholesale eNaira and is studying a private stablecoin. Live issuance is not a one-way ratchet — a CBDC can be wound down when the demand case fails.

4.3 Mixed Results: Interest-Bearing CBDC as Monetary Tool

China's January 2026 decision to make e-CNY interest-bearing breaks with the global consensus and introduces a potential new monetary policy transmission channel. The theoretical case for interest-bearing CBDC is established in IMF Fintech Note 2023/010:

Potential benefits:

• Strengthens transmission of policy rate changes directly to retail holdings
• Creates a new floor for deposit rates (banks must at least match the CBDC rate to retain deposits)
• Enables future use of negative rates in extreme deflationary scenarios
• Provides a real-time monetary policy lever reaching all wallet holders simultaneously

Risks and concerns:

• Directly competes with bank deposits if the CBDC rate exceeds deposit rates
• Complicates monetary policy communication (multiple rates to manage)
• Creates procyclical incentives: during stress, higher CBDC rates could accelerate deposit flight
• China's experience is too early to evaluate (first payouts: 20 March 2026)

China's specific context: The 0.05% rate is deliberately negligible relative to deposit rates. The primary purpose appears to be completing the reclassification of e-CNY as deposit money rather than deploying a new monetary policy tool. Whether this represents a replicable model or a pragmatic response to adoption challenges remains to be seen — no other jurisdiction in the matrix pays or plans to pay interest.

5. Recommended Framework

5.1 Holding Limit Calibration Methodology

Central banks should follow a structured calibration process, drawing on the ECB and BoE methodologies:

Step 1: Define the holding limit objective function. The limit must balance two competing goals:

• Usability floor: The limit must be high enough to support routine payment needs (monthly grocery bills, utility payments, in-store purchases). For reference: median monthly household expenditure is approximately EUR 2,700 in the eurozone and GBP 2,500 in the UK
• Stability ceiling: The limit must be low enough to prevent CBDC from becoming a store of value that drains bank deposits. The ceiling is determined by stress-test modelling

Step 2: Model deposit outflow scenarios. Central banks should model at minimum three scenarios using supervisory banking data:

1. Business-as-usual: What fraction of the population will adopt CBDC, and at what average holding? Historical precedent from China, India, Nigeria, and Jamaica suggests initial (and, in India, declining) adoption is far below theoretical maximum
2. Moderate stress: A localised banking crisis where 20–30% of uninsured depositors convert to CBDC
3. Severe stress / flight-to-safety: A systemic crisis where all eligible individuals fill wallets to the maximum. This is the ECB's EUR 699 billion scenario

Step 3: Assess banking sector capacity. For each scenario, evaluate:

• How much central bank lending would banks need to replace lost deposits?
• How many firms would breach regulatory liquidity thresholds (LCR, NSFR)?
• What is the impact on lending rates and credit supply?
• Is sufficient eligible collateral available for central bank lending operations?

Step 4: Select the initial limit. Country precedents suggest the initial limit should err conservative:

Step 5: Build in adjustment mechanisms. The Council of the EU's position (December 2025) provides a template: the ECB sets the limit within an overall ceiling agreed by the Council, reviewed at least every two years. Downward revisions via delegated acts; upward revisions requiring legislative procedure. This asymmetry (easier to tighten, harder to loosen) is a prudent governance design.

5.2 Interest Rate Policy Options

Central banks must choose among four remuneration strategies:

Option 1: Zero remuneration (recommended default).

• Used by: EU (proposed), UK (proposed), India, Russia, Japan (recommended), and all launched retail CBDCs
• Advantages: simplest to implement; avoids competition with deposits; maintains clear CBDC-as-payment-instrument positioning
• Disadvantages: no monetary policy transmission via CBDC channel; no ability to use negative rates

Option 2: Positive fixed rate.

• Used by: China (0.05% from January 2026) — the only jurisdiction in the matrix
• Advantages: supports reclassification to deposit money; small incentive for wallet maintenance
• Disadvantages: requires ongoing rate management; creates (small) competition with deposits; complicates monetary communication
• When appropriate: when CBDC is deliberately being integrated into the deposit money classification (as in China's case)

Option 3: Dynamic/tiered rate.

• Used by: no jurisdiction currently
• The IMF's Fintech Note 2023/010 describes this option: offering a lower remuneration for store-of-value holdings (below bank deposit rates) while potentially a higher rate for active transaction balances
• Advantages: fine-grained monetary policy transmission; can discourage hoarding while encouraging payment use
• Disadvantages: complex to implement and communicate; requires sophisticated rate-setting infrastructure
• When appropriate: advanced-economy contexts where CBDC is positioned as a new monetary policy channel

Option 4: Negative rate capability.

• Used by: no jurisdiction currently; theoretically explored by Sweden and the ECB
• Advantages: provides an effective lower bound tool in deflationary scenarios; resolves the "zero lower bound" problem
• Disadvantages: politically toxic; likely to destroy public trust and CBDC adoption; no empirical precedent
• Assessment: should be retained as a theoretical option in CBDC system architecture but not deployed without extraordinary economic justification and public mandate

5.3 Bank Transition Management Strategies

Central banks introducing retail CBDCs should implement structured transition support for the commercial banking sector:

Liquidity backstop facilities. Central banks should announce standing lending facilities explicitly designed to replace deposit outflows caused by CBDC adoption. The CBR has committed to "fully offset any liquidity outflow." The BoE's Financial Stability Paper 53 models the required lending capacity. Central banks should pre-position this capacity before CBDC launch.

Phased mandatory adoption timelines. Russia's phased approach (systemically important banks by September 2026; all universal-licence banks by September 2027; remaining institutions by September 2028) provides a template for graduated implementation — but Russia's >50% public unwillingness is a cautionary signal that mandates must be paired with genuine utility. Banks should have at minimum 12–18 months notice before mandatory integration.

Compensation models for intermediaries. The EU is developing fee structures for digital euro PSPs. The debate centres on inter-PSP fees (Spain proposed 0.2%; EuroCommerce proposed 0.05% with a 2-cent cap). Central banks should ensure the compensation model creates viable business incentives for intermediary participation without imposing excessive costs on merchants.

Deposit rate flexibility. As noted in BIS Working Paper 1280, CBDC introduction creates incentive for banks to offer more competitive deposit rates to retain funding. Central banks should monitor deposit rate dynamics and be prepared to adjust monetary policy operations if CBDC adoption creates unexpected upward pressure on funding costs.

5.4 Stress Testing Requirements

Before any retail CBDC launch, central banks should conduct and publish results of:

1. Deposit outflow stress tests across multiple holding-limit scenarios (the ECB's EUR 500–3,000 range and the BoE's GBP 5,000–unlimited range provide templates)
2. Bank-by-bank impact assessments identifying institutions most vulnerable to deposit migration (small and medium banks with high retail deposit dependency)
3. Collateral adequacy tests verifying that sufficient eligible collateral exists for central bank lending operations at modelled outflow levels
4. Payment system capacity tests ensuring the CBDC platform can handle crisis-level transaction volumes (mass conversion from deposits to CBDC)
5. Interoperability stress tests validating that waterfall/reverse waterfall mechanisms function under peak load

The IMF's Fintech Note 2025/008 provides an updated stress-testing framework, and IMF WP/24/226 provides the formal balance-sheet methodology.

6. Decision Matrix

The following matrix synthesises the key design decisions and their monetary policy implications. Central banks should use this as a starting framework, calibrating to domestic conditions.

6.1 Holding Limit Decision

6.2 Remuneration Decision

6.3 Classification Decision

6.4 Transition Safeguards Decision

Wikilinks

Jurisdiction profiles (20):

United States,

Bahamas,

Brazil,

Canada, Eastern Caribbean,

Jamaica,

Australia,

China,

India,

Japan,

Singapore,

South Korea,

Thailand,

EU,

Russia,

Sweden,

UK,

Ghana,

Nigeria,

South Africa

International frameworks: BIS, IMF

Cross-references: Legal Framework & Authorization, Technology Design Choices, Privacy & Data Protection, Financial Inclusion, Cross-Border Interoperability, Economic Outcomes, Stablecoin Coexistence

Verification Notes

This is a refresh synthesis (June 2026). The canonical source of truth for every per-jurisdiction fact is .claude/cbdc-frameworks-fact-matrix.json. No web search was performed; facts mirror the matrix, and adoption/transaction/projection figures are flagged as self-reported with their source. "X of 20" aggregates are recomputed from the matrix.

Facts Refreshed This Cycle (against the matrix)

1. United States — Anti-CBDC. Only EO 14178 is binding; the statutory CBDC moratorium passed both chambers in differing, unreconciled forms (Senate 21st Century ROAD to Housing Act 89-10, 12 Mar 2026; House amended 396-13, 20 May 2026) and is NOT enacted. GENIUS Act = S.1582 / Public Law 119-27. Kevin Warsh sworn in as Fed Chair 22 May 2026, reaffirming opposition to a retail CBDC.
2. China — Interest-bearing from 1 Jan 2026 (0.05%, first payout 20 Mar 2026); reclassified toward deposit money; 22 authorized operators after +12 banks on 2 Apr 2026.
3. India — e₹-R circulation ₹771.7 crore at 31 Mar 2026, ~24% YoY decline (first reported contraction; self-reported, RBI AR 2025-26).
4. Russia — RUB 300,000 monthly top-up cap (replenishment only); narrow live government/budget use since 1 Jan 2026; full launch mandated 1 Sep 2026.
5. Eastern Caribbean — DCash 2.0 successor suspended by ECCB Monetary Council 13 Feb 2026 in favour of Fast Payment System + CAPS.
6. Brazil — Drex Hyperledger Besu platform decommissioned 4 Nov 2025; MP 1303 lapsed 8 Oct 2025; rebuilding from scratch.
7. Canada — Stablecoin Act (Bill C-15) Royal Assent 26 Mar 2026, not yet in force, expressly excludes CBDCs.
8. Australia — Project Acacia final report 18 May 2026; A$4.4m wCBDC issued across 8 use cases.
9. EU / Sweden — ECB preparation phase concluded Oct 2025; Governing Council moved to next phase 29–30 Oct 2025; ECON Committee vote scheduled 23 Jun 2026. Sweden's path now contingent on the digital euro.
10. Ghana / South Africa — eCedi pilot declared "completed" 22 May 2026 (retail unlaunched); SARB Position Paper 27 Nov 2025 ("no compelling immediate need"), wholesale via Project Khokha 2x.

Verified (in-matrix; underlying figures previously web-checked)

• ECB EUR 699 billion outflow (EUR 3,000 limit, flight-to-safety) — ECB technical data, Oct 2025.
• BoE GBP 250 billion lending requirement (no-limit scenario) — Financial Stability Paper No. 53, Nov 2025.
• China 0.05% interest rate — the exact 0.05% figure derives from Bank of China International Securities analysis reported in secondary sources; the State Council release confirms interest is paid but does not specify the rate.
• Bahamas Sand Dollar deposit/loan impact — MDPI academic paper (2025); circulation ~0.39% of cash.
• Nigeria eNaira 98.5% inactive wallets — IMF WP/23/104 (self-reported, flagged).
• Jamaica JAM-DEX +550% transaction value / +267% volume in 2025 (self-reported, BOJ) — from a very low base; no deposit-impact data.

Items Not Independently Verified (per matrix)

• Exact wallet balance caps for China Categories 2 and 3 — NOT VERIFIED to primary PBOC source.
• Brazil Drex retail holding limit and interest treatment — NOT VERIFIED (rDrex = tokenized deposits under FGC).
• Eastern Caribbean DCash interest, programmability, merchant limits — NOT VERIFIED.
• Jamaica programmability; Ghana holding limit / interest / programmability; Thailand retail holding limit; Japan/Sweden holding limits and interest decisions — NOT VERIFIED / not yet decided.
• China e-CNY actual disintermediation data post-2026 reclassification — NO DATA FOUND (too early to measure).
• Adoption, transaction-volume, and multi-year projection figures (China, India, Nigeria, Jamaica, Bahamas, Eastern Caribbean, Russia, Australia) are self-reported by the issuing central bank or its trackers and are flagged accordingly.

---

Last updated: 2026-06-04