## ftnea2025008

## Source details

**Canonical URL:** [ftnea2025008](https://www.imf.org/-/media/files/publications/ftn063/2025/english/ftnea2025008.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/ftn063/2025/english/ftnea2025008.pdf.md)
- [Structured JSON version](/-/media/files/publications/ftn063/2025/english/ftnea2025008.pdf.json)

---

### Key findings and scope
- CBDC is a new liquid, safe, and widely accessible payment instrument and store of value that could substitute for other forms of money, including bank deposits.
- A foundational principle: CBDC should be designed to coexist with existing forms of money.
- Agents may choose to decrease holdings of bank deposits and safe/money-like assets in favor of CBDC, introducing competitive pressures that can affect financial stability.
- The Note focuses narrowly on the effects of CBDC on domestic financial stability and does not provide an integral assessment of whether to issue CBDC.
- Contribution:
  - Provide a framework to think through financial stability implications across economies and draw insight from quantitative studies.
  - Offer guidance for practitioners on tools, data, and models to evaluate financial stability risks and mitigation strategies.

### Six transmission channels through which CBDC affects financial stability
- Liability channel
  - CBDC can induce deposit outflows and change banks’ cost and composition of funding.
  - Four bank-response scenarios:
    - Increase reliance on short-term wholesale funding (less stable, costlier, increases interconnectedness).
    - Substitute deposit funding with long-term debt (reduces rollover risk, raises interest costs).
    - Rebalance liabilities toward more foreign funding (raises currency-mismatch risks).
    - Increase reliance on central bank borrowing (raises funding costs, sovereign–bank nexus, central bank balance sheet risks; may cause collateral shortages).
  - Banks may respond by raising retail deposit rates, shifting sight deposits to term deposits, reducing fees, or offering complimentary services—affecting profitability and operational costs.
- Asset channel
  - Competitive pressures may shrink bank asset holdings, deplete excess reserves, force asset sales, change asset composition, and potentially increase risk-taking.
  - Banks may raise lending rates to protect margins, which can reduce lending volumes.
  - Nonbank issuers (e.g., MMFs) may sell safe assets or move into riskier assets to retain yields.
- Fee income channel
  - Reduced usage of bank checking accounts and transactions can lower fee income (account maintenance, interchange, overdrafts); cross-selling revenues may decline.
- Run-risk channel
  - CBDC can facilitate runs by offering a safe, convenient alternative to deposits.
  - Compositional effects matter: smaller deposit bases and more long-term funding can mitigate run severity; greater reliance on short-term wholesale funding heightens vulnerability.
  - CBDC availability may help retain funds domestically in crises if depositors trust the central bank and inflation expectations are managed.
- Information channel
  - Deposit outflows reduce banks’ borrower information (transactional signals), increasing information asymmetries and default risk.
  - CBDC transaction data can help monetary authorities with real-time monitoring and early intervention, but use must balance privacy and operational risks.
- Resilience (payment system) channel
  - CBDC can lower concentration risks, improve interoperability, provide backup payment rails, and enhance liquidity management if infrastructure is resilient; poor design risks excessive dominance and operational fragility.

### Overarching considerations and adoption determinants
- Channels are interconnected and can reinforce one another.
- Liability, asset, and run-risk channels are likely most significant given deposits’ central funding role.
- Magnitude depends on CBDC adoption level, which may be gradual or limited.
- Adoption determinants:
  - Trust in the banking system (lower CBDC adoption when trust is strong).
  - Maturity of payment systems.
  - Strength of network effects.
  - Trust in the central bank.
- Design choices (e.g., unremunerated CBDC) affect attractiveness; switching costs can slow adoption and reduce immediate competitive pressures.
- Stablecoins may coexist or compete with CBDC and can generate similar pressures but have different vulnerabilities.

### Quantitative evidence and key numeric findings
- General: existing studies focus on liability and asset channels; results are model- and country-dependent.
- Studies typically find limited financial stability risks under mild adoption; banking profitability may decline but systemic risk is often contained in many country contexts.
- Scale of CBDC adoption and remuneration are crucial determinants of impacts.
- Selected study results (preserve exact values quoted in source):
  - Garcia and others (2020): "0.02–1 pp decrease in RoE"; scenarios: (5–33 percent of total deposits).
  - BIS (2021): "0.05–0.3 pp decrease in RoE"; scenarios: (5–25 percent of deposits).
  - Bouis and others (2024): "3–8 pp decrease in RoE"; "0–17 pp decrease in NSFR"; "1–83 pp decrease in LCR"; scenarios: (6–17 percent of deposits).
  - Gross and Letizia (2023): "<0.05 pp decrease in RoA in the Euro Area"; "0.11–0.15 pp decrease in RoA in Bahrain"; "0.1–0.6 percentage point decrease in Net Interest Margin in the United Arab Emirates".
  - Chen and Phelan (2025): "30–40 bps increase in deposit rates"; "5–6 pp increase in probability of banks in distress state"; "3–4 pp increase of probability of banks in crisis state"; scenario: (9–13 percent of total money supply, M2).
  - Chang and others (2023): "40–50 bps increase in deposit rates"; "58–75 percent decrease in profits" (model estimate, US).
  - Bidder, Jackson, and Rottner (2024): "CBDC increases the probability of bank runs by 1.2 pp"; estimated for the Euro area.
- Magnitudes in baseline scenarios (CBDC demand around 10 percent of demand deposits): decreases in bank RoE on the order of around 0.3 to 1 percentage point in some studies; larger declines (3–8 pp) in studies with higher assumed costs and fee-income drops.
- Notable extreme-assumption result: Chang and others (2023) forecast a 70 percent drop in profitability in their baseline given nearly 30 percent deposit losses and restrictive assumptions.

### Run-risk headline statistic (from reviewed literature)
- Introduction of CBDC increases the probability of bank runs by 1.2 percentage points, to 2.5 percent.
  - Interpreted as an increase from roughly every 70 years (without CBDC) to roughly every 40 years (with CBDC), per the authors.

### Practical analysis steps, data needs, and model categories
- Recommended practical steps:
  - Start with balance sheet analyses to trace potential responses across banks, central bank, government, and private sector.
  - Use financial and macrofinancial models for richer dynamics and behavioral responses.
  - Root analysis in granular data and update routinely post-launch.
- Key data categories to collect (preserve wording examples):
  - Balance sheet composition of banks: loans, bonds, other security holdings, reserve holdings; liabilities: retail deposits, wholesale deposits, other bond finance; regulatory capital metrics such as CET1/RWA and total capital/RWA; encumbered versus unencumbered assets; concentration metrics; reserve holdings (required vs. excess) and reserve borrowing across banks. "The data needs to have a time dimension, to examine trends."
  - Profitability of banks: interest income, interest expense, fee income, bottom-line net income before tax (all by bank and over time); interest income over interest-bearing assets; interest expense over liabilities; detailed fee and commission income/expenses.
  - Central bank balance sheet and income: reserve lending, holdings of government and private bonds, reserve holdings of banks, reserve requirements, interest income/expense, net income/seigniorage, policy rates/ corridor components.
  - Other metrics: competition measures (Skew and Herfindahl indices, share of largest banks), cash-in-total money ratios, extent of other currency use, reserve coverage.
- Model categories and roles:
  - Category #1 — Balance sheet analyses: visualization, first step; cannot analyze information and payment resilience channels.
  - Category #2 — Financial models (utility, consumer choice, SFC): analyze deposit/lending responses, income/expense effects, sometimes endogenous CBDC demand.
  - Category #3 — Macro models (NK, DSGE, OLG, NM): capture real effects (real GDP, employment) and most channels except information and payments resilience. Require advanced macro and coding expertise.
- Modeling notes preserved:
  - "The aggregate CBDC adoption scenarios are either directly assumed in an ad hoc manner or calculated by adding up holdings for individual(s) which are based on assumptions regarding CBDC holding limits per household/individual."
  - "The level of utility (value) from holding different forms of money... can be informed by certain benchmarks, for example by equating CBDC base utility... to that of physical currency, or that of deposits."

### Mitigation options, design levers, and policy sequencing
- High-level mitigation options:
  - Price measures: low or zero CBDC remuneration rates, fees on transactions or conversions; remuneration can be set to "zero" or "at a sufficiently low level" or, "in theory, remuneration could also be negative."
  - Quantity measures: hard limits on holdings, transactions, and convertibility; waterfall and reverse-waterfall mechanisms to link CBDC wallets to bank accounts.
  - Access parameters: retail-only versus wholesale access alters substitution effects and channel magnitudes.
  - Two-tier distribution with intermediaries: intermediaries can offset fee and information losses by providing wallets and services.
- Calibration and sequencing:
  - Tradeoff: limiting adoption can reduce financial stability risks but may undermine CBDC objectives like financial inclusion.
  - Differential measures by user type (individuals vs. merchants) and tiered remuneration (e.g., higher limits remunerated at lower or negative rates).
  - Time-varying or contingent application: measures can be revised over time or applied in stress periods; central banks should be ready to provide liquidity support where appropriate.
- Cross-policy interaction:
  - Macroprudential policies and crisis management remain relevant complements.
  - CBDC can redirect outflows from stablecoins and improve competition and operational resilience in payments.

### Calibration, operational experience, and examples
- Calibration challenges:
  - Modeling relies on simplifying assumptions; reactions change over time.
  - Need for forward-looking assessment and ongoing data collection post-launch (holdings, flows, anonymized demographics).
  - Large data requirements if limits vary by user type.
- Operational precedents and examples:
  - Central Bank of the Bahamas increased holding limits for individuals from BSD 5,000 to BSD 8,000 after the pilot phase.
  - ECB used scenario-based assessment of deposit outflows for limits calibration (ECB 2024).
- Final practical advice:
  - Balance sheet analyses are straightforward and accessible to many central banks.
  - Category #2 financial models require advanced modeling expertise; Category #3 macro models require deep macroeconomic modeling skills.
  - Detailed fee and commission data enhance transaction-flow competition analysis.

### Stock-take of central banks’ approaches (Box 1 highlights)
- Central banks largely favor nonremunerated CBDC; more than half of central banks in the BIS survey do not intend to remunerate CBDC (Illes, Kosse, and Wierts 2025).
- Most central banks in advanced stages consider holding limits; less than 30 percent of surveyed central banks in advanced economies are considering transaction limits.
- Two calibration approaches for holding limits:
  - Top-down: country-level desired adoption divided by eligible population (example: EUR 3,000–4,000 per individual proposed relative to EUR 1–1.5 trillion banknotes).
  - Bottom-up: individual-level limits aggregated (example: GBP 10,000–20,000 would allow 75–95 percent of income earners to use Digital Pound for salaries).
- Proposed holding limits in selected projects (values preserved exactly as in table):
  - Sand Dollar (The Bahamas) — Live — B$ 500 (1%) — B$ 8,000 (18%) — B$ 8,000–B$ 1,000,000 (18–2227%)
  - JAM-DEX (Jamaica) — Live — None2 — None2 — None2
  - e-Naira (Nigeria) — Live — NGN 120,000 (3%) — NGN 5,000,000 (141%) — No limit
  - e-Cedi (Ghana) — Live (plans) — Yes (TBD) — Yes (TBD) — Yes (TBD)
  - Digital Euro (Eurosystem) — Live (plans) — TBD3 — TBD3 — 0
  - Digital Pound (United Kingdom) — Live (plans) — GBP 10,000–20,000 (22–43%)4 — — “Significantly higher than individuals”
  - D-Cash (ECCU) — Pilot — None — None — -
  - e-CNY (China) — Pilot — RMB 10,000 (6%) — None — -
  - Digital Rupee (India) — Pilot — INR 100,000 (16%) — -
- Policy considerations: design features are a first line of defense, but macroprudential tools, crisis management, and supervision/regulation gaps must be addressed comprehensively.

### Model development needs and illustrative model applications
- Model development suggestions:
  - Add “macroeconomic shells” to nominal structural models to assess real economy pass-through.
  - Enhance liquidity focus to analyze banking liquidity dynamics and collateral scarcity.
  - Combine frameworks that endogenize bank interest rates and incorporate NBFIs and bank money creation.
- Example model applications and findings:
  - SFC model (Bouza and others 2024) for Georgia: excess reserves deplete or require borrowing; RoA falls but solvency preserved; central bank net income follows a hump shape.
  - New Keynesian model (Abad, Nuño, and Thomas 2025) for the euro area: CBDC take-up of 14 percent of GDP reduces bank deposits by 11 percent of GDP and lending drops by 0.6 percent relative to pre-CBDC stock; system regimes (floor, corridor, ceiling) vary by CBDC share with thresholds: below 4 percent of GDP (floor), beyond 4 percent (corridor starts), beyond 10 percent (ceiling state where banks structurally borrow from central bank).

*Fintech Note: Evaluating the Implications of CBDC for Financial Stability — INTERNATIONAL MONETARY FUND*

### Introduction and Summary of Findings ...................................................................................

### Introduction and Summary of Findings

### Key findings
- CBDC is a new liquid, safe, and widely accessible payment instrument and store of value that could substitute for other forms of money, including bank deposits.
- A foundational principle of CBDC is that it should be designed to coexist with existing forms of money.
- Agents may choose to decrease their holdings of bank deposits as well as safe and money-like assets in favor of CBDC, introducing competitive pressures in the financial system and potentially affecting financial stability.
- Financial stability should be seen holistically, including banks’ balance-sheet metrics (profitability, liquidity, solvency) and broader elements (risk-taking behavior, interconnectedness, resilience).
- The Note identifies six interrelated channels through which CBDC could affect financial stability: liability channel, asset channel, fee income channel, run-risk channel, information channel, and resilience channel.
- The relative strength and economic significance of these channels are theoretically ambiguous and depend on factors such as the level of CBDC adoption, country characteristics, and design features.
- Quantitative studies to date find CBDC would not pose significant financial stability risks under scenarios of mild adoption, though country characteristics need to be factored in. Findings suggest banking system profitability may decrease, but the overall impact on financial stability is likely to be limited—especially in countries with low competition, low reliance on deposit funding, and sufficient access to alternative sources of funding.
- Analysis requires data on balance sheet composition of banks and the central bank, profitability, metrics of competition, and other financial indicators; richer dynamics require financial and macrofinancial models.
- Countries can contain financial stability risks with appropriate CBDC design and policies (for example, quantity limits on CBDC holdings or fees applied to holdings above a certain limit), though restricting adoption can undermine other CBDC objectives. Traditional safeguards—prudential and crisis management policies—remain useful.

### Scope and purpose
- The Note focuses narrowly on the effects of CBDC on domestic financial stability and does not provide an integral assessment of the implications of CBDC nor opine on the appropriateness of issuing CBDC.
- The contribution is twofold:
  - Provide a framework to think through financial stability implications of CBDC across different types of economies and draw insight from studies that have quantified some effects.
  - Offer guidance for practitioners on tools and models to evaluate financial stability risks and discuss strategies to mitigate them.

### Stylized environment before CBDC
- Financial system agents hold cash, bank deposits, private forms of digital money (such as e-money), and safe and money-like assets (short-term government securities, commercial paper, certificates of deposit, well-regulated stablecoins fully backed with safe and liquid assets).
- Holding decisions are based on product attributes (risk, return, cost, convenience) and consumption and saving preferences (Li 2023).

### Transmission channels (overview)
- Six main channels through which CBDC may affect financial stability (Figure 1):
  - Liability channel
  - Asset channel
  - Fee income channel
  - Run-risk channel
  - Information channel
  - Resilience channel
- Channels are closely interconnected and operate simultaneously.
- Design features and policies aimed at mitigating financial stability risks are discussed separately in the Note.

### Liability channel (detailed mechanisms and bank responses)
- The liability channel concerns CBDC-induced competitive pressures on the liability side of financial institutions, notably composition changes and cost of funding.
- Introduction of CBDC could lead to outflows of bank deposits into CBDC; four main bank-response scenarios are highlighted:
  - Increase reliance on short-term wholesale funding:
    - Wholesale funding is inherently less stable, exposes banks to rollover risks, and is generally costlier than retail deposit funding (Huang and Ratnovski 2011; Gertler, Kiyotaki, and Prestipino 2016).
    - Rebalancing toward wholesale funding may increase interconnectedness between banks and NBFIs (Bouis and others 2024). A shock to MMFs could negatively affect wholesale funding availability.
    - On the positive side, wholesale funding could instill greater market discipline because wholesale providers are not protected by explicit deposit insurance schemes (Mancini-Griffoli and others 2018; Das and others 2023).
  - Substitute deposit funding with long-term debt:
    - Long-term debt can reduce rollover risks and improve liquidity but typically carries higher interest rates and may reduce bank profitability.
  - Rebalance liabilities toward more foreign funding:
    - May lead to higher currency mismatches and funding volatility. Foreign lenders tend to exhibit home bias when funding conditions deteriorate (Giannetti and Laeven 2012).
  - Increase reliance on central bank borrowing:
    - Could raise funding costs and heighten the sovereign–bank nexus. Central bank borrowing rates are typically set above market rates to discourage excessive use.
    - Significant expansion in central bank borrowing would grow central bank balance sheets and expose them to financial risks.
    - Outside of a bank run situation, central banks may not consider it a viable policy option nor have statutory authority to do so (Mancini-Griffoli and others 2018; Bouis and others 2024).
    - An expansion of the central bank balance sheet may lead to more risk exposure to the private sector, either indirectly (if banks engage in more collateralized reserve borrowing) or directly (if the central bank purchases assets/collateral). Prudent risk management by the central bank is important.
    - Greater reliance on central bank liquidity may lead to collateral shortages as a higher share of assets used as collateral would be pledged with the central bank, particularly in emerging markets with limited high-quality liquid assets, potentially straining repo markets and amplifying fragility (Greppmair, Paludkiewicz, and Steffen 2025).
- Banks may respond to CBDC competition by raising retail deposit rates to retain funding. In less competitive banking sectors, higher profit margins give banks more leeway to increase deposit rates.
- Banks might shift sight deposits toward term deposits, improving maturity alignment and liquidity, or retain deposits via reduced fees or complimentary services, which may raise operational costs.
- Nonbank issuers of safe and money-like assets may face substitution away into CBDC, reducing demand, lowering prices, and raising yields—though substitution might be muted if CBDCs are not remunerated.

### Asset channel (introductory note)
- The asset channel is closely tied to the liability channel and operates through the asset side of financial institutions’ balance sheets. (Further detail in subsequent sections of the Note.)

### Quantitative evidence and modeling guidance (summary)
- Existing quantitative studies are primarily theoretical and calibrated to specific country circumstances, focusing on banks’ balance sheets.
- Studies generally indicate limited financial stability risks under mild CBDC adoption, with potential declines in banking profitability but limited systemic risk in certain country contexts.
- Country-specific characteristics—competition in banking, reliance on deposit funding, access to alternative funding—are critical determinants of outcomes.
- Practical analysis steps recommended:
  - Start by exploring how balance sheets of financial institutions, the central bank, government, and private sector could respond to CBDC.
  - Use financial and macrofinancial models to capture richer sector dynamics and behavior.
  - Root analysis in data such as balance-sheet composition, profitability, banking competition metrics, and broader financial-system indicators.

### Mitigation options (high-level)
- Design and policy tools to contain risks include quantity limits on CBDC holdings and fees on holdings above a limit.
- Design choices that restrict adoption can undermine other CBDC objectives.
- Traditional safeguards—prudential regulation and crisis management—remain relevant.

*Fintech Note: Evaluating the Implications of CBDC for Financial Stability — INTERNATIONAL MONETARY FUND*

### introduction of CBDC may affect the size and composition of financial institutions’ assets and increase

### introduction of CBDC may affect the size and composition of financial institutions’ assets and increase

### Asset and liability adjustments; lending rates
- Competitive pressures from CBDC may induce banks to decrease their asset holdings and the size of their balance sheets.
- Banks facing deposit outflows may:
  - Shrink reserve assets held in excess of regulatory requirements to service outflows, which could suffice to cover retail deposit declines.
  - Sell other assets to service deposit outflows.
  - Experience reduced interest income if excess reserves are remunerated and decline.
- Competitive pressures may induce banks to change asset composition and take excessive risk, particularly banks deemed too big to fail.
- Depending on banking sector competition, banks may increase lending rates to mitigate declining interest margins, which:
  - Supports interest margins but may result in fewer loans.
  - Benefits banks with more market power, which can pass on higher funding costs to borrowers while limiting reductions in lending volumes.
- Nonbank issuers of safe and money-like assets (e.g., MMFs) may:
  - Sell asset holdings or change asset composition in response to competition.
  - Shift to riskier assets to offer higher returns and mitigate fund migration into CBDC.

### Fee income channel
- CBDC-induced competitive pressures could reduce fee income for banks and nonbank PSPs by:
  - Reducing usage of bank checking accounts and associated account-related fees (account maintenance, withdrawals, penalties for non-sufficient funds, overdrafts, late payments).
  - Lowering transaction-related fees (interchange, assessment, and end-user fees).
- Deposit outflows could reduce commercial banks’ revenues from cross-selling other services if customers have fewer interactions or leave the bank.
- Households and firms would likely still need some type of bank account for services CBDC would not offer.

### Run-risk channel
- CBDC may facilitate bank runs by providing a safe and convenient store of value, facilitating flight to safety away from bank deposits.
- The effect may be muted where liquid and safe alternatives and low switching costs already exist; deposit insurance, supervision, and central bank liquidity reduce run incentives.
- CBDC availability in a banking crisis may help retain funds domestically and limit currency crisis risks, conditional on depositors’ expectations about inflation and trust in the central bank.
- Compositional balance-sheet changes can affect run severity:
  - Smaller deposit bases from conversion into CBDC imply less severe impacts from retail deposit runs.
  - Increased share of long-term funding (term deposits, bond issuance) and central bank borrowing may reduce runs.
  - Greater dependence on short-term wholesale funding increases exposure to liquidity risk, especially when short-term funding finances long-term illiquid assets.
- CBDC may facilitate runs on issuers of safe and money-like assets, potentially triggering wholesale funding withdrawals from banks (example: MMF runs reducing demand for bank-issued commercial paper).
- Withdrawal of wholesale funding and reduced interbank transactions could undermine key interest rate benchmarks (SOFR, €STR, Federal Funds rate) if benchmarks rely on active liquid secured interbank market transactions.

### Information channel
- Bank deposit outflows can result in the loss of borrower information (transaction data, consumption/saving patterns, cash flows, payment history, income, fund balances), increasing information asymmetries and default risks.
- CBDC transaction data could be valuable to monetary authorities to support financial stability by:
  - Allowing real-time transaction monitoring to improve monetary policy and financial stability actions.
  - Enabling agile monitoring of flows into CBDC in stress times to identify weak banks and intervene early, potentially reducing depositors’ incentives to run.
  - Strengthening regulatory compliance and law enforcement, reducing illicit activity (dependent on CBDC infrastructure choices, e.g., smart contracts).
- Policymakers must balance data use and privacy protection; aggregated or anonymized data could be used without privacy concerns, while personal data use poses risks of data leakage, abuse, cyberattacks, operational instability, and undermined trust.

### Payment system resilience channel
- CBDC can lower concentration and fragmentation risks amid declining cash use by providing public infrastructure and an appropriate regulatory framework that promotes interoperability among PSPs.
- Integration of PSPs into a CBDC ecosystem can provide access to a safe and liquid digital asset, improving liquidity management and lowering settlement costs.
- Reduced market concentration can improve operational resilience by making single-firm failures less systemically relevant.
- A well-functioning CBDC system could serve as a back-up for the traditional payment system, but requires a resilient and efficient infrastructure; large operational outages could have financial stability repercussions.
- Poorly designed CBDC risks displacing existing competitors and becoming overly dominant.

### Overarching considerations
- Channels are interconnected and can reinforce one another (e.g., lower fees lead to higher perceived risk and higher investor return requirements).
- Bank liability, asset, and run-risk channels are likely to be the most significant overall, given deposits’ role as a major funding source in many systems.
- Magnitude of channels depends on CBDC adoption, which may be gradual or limited, particularly in non-crisis times.
- Adoption determinants include:
  - Trust in the banking system (lower incentives to adopt CBDC when trust is stronger).
  - Maturity of payment systems (lower incentives when mature).
  - Strength of network effects (higher network effects reduce incentives to adopt CBDC).
  - Trust in the central bank (higher trust could increase CBDC adoption).
- CBDC design features affect attractiveness and accessibility; most central banks are considering unremunerated CBDC, which curtails usefulness for savings.
- Switching costs can result in sticky behaviors, slowing retail adoption unless clear benefits are observed.
- Even if widely adopted, CBDC may not imply large competitive pressures if it substitutes cash or only partially substitutes private digital money.
- CBDC could increase deposit holdings in less competitive banking systems by reducing banks’ market power and prompting banks to offer higher deposit rates.
- Reduced bank net income may not threaten financial stability unless profits fall significantly and capital is strained for a significant segment of the banking system; lower profit margins increase vulnerability to future shocks.
- Short- to medium-term channels reflect initial reactions and transition phases; medium- to long-term outcomes may differ (e.g., market consolidation, business model evolution).
- Stablecoins may coexist with CBDC and could generate similar competitive pressures; unlike CBDC, stablecoins are vulnerable to value volatility and run risks. CBDC and stablecoins could compete or complement each other.

### Insights from quantitative studies
- No quantitative studies analyze all channels simultaneously; studies focus mainly on liability and asset channels and, to a lesser extent, run-risk channel.
- Modeling frameworks vary in granularity and complexity; heterogeneity in methods requires careful consideration when comparing results.
- Measures of financial stability used in studies include bank RoE, RoA, net stable funding ratio, liquidity coverage ratio, deposit rates, aggregate deposit balances, lending rates, and model-specific run-risk measures.
- Liability channel findings:
  - Impact on deposits depends on banking sector competition: fully competitive sectors likely see conversion of bank deposits to CBDC and higher deposit rates; less competitive sectors may see deposit increases.
  - Chang and others (2023) estimate deposits could grow by 0.1 percentage points (as a share of total households’ wealth) in a scenario with non-fully competitive banks and costly CBDC adoption.
- Studies mostly agree that CBDC introduction would decrease banking sector profitability through the liability channel, regardless of deposit volume changes.
  - Profitability loss cannot be fully recouped by increasing lending rates due to lending market competition.
  - Even if deposits increase, profitability can decrease due to higher deposit rates.
- The scale of CBDC adoption is the most important determinant of impact magnitudes; remuneration of CBDC increases take-up in models that estimate demand.
- Effects on profitability through liability and asset channels are not sizable in baseline scenarios where CBDC demand is around 10 percent of demand deposits:
  - Garcia and others (2020) and BIS (2021) point to decreases in bank RoE of around 0.3 and 1 percentage points, respectively, in their highest CBDC demand scenarios (RoE starting points cited: around 15 percent in Garcia and others (2020) and 7.5 percent in BIS (2021)).
  - Bouis and others (2024) find a larger decline in RoE between 3 and 8 percentage points, due to assumed higher costs of alternative refinancing and large fee-income drops.
  - Gross and Letizia (2023) point to a decrease in RoA in the euro area banking system of less than 0.05 percentage points, with an application to Bahrain suggesting a drop of about 0.11–0.15 percentage point.
- Some studies report much larger effects under extreme assumptions:
  - Chang and others (2023) forecast a large 70 percent drop in profitability in the baseline scenario after a drop of nearly 30 percent in deposits (based on assumptions that deposits are the only funding source and investment remunerated at the policy rate).
- In most cases, the impact on financial stability seems manageable:
  - Adalid and others (2022) estimate euro area banks would need central bank refinancing only if CBDC demand were very high (18 percent of deposits).
  - Chen and Phelan (2025) find CBDC increases the probability of banks being in a crisis state by 3 percentage points (to around 6 percent).
  - Berg and others (2024) find announcements hinting at a higher probability of CBDC in the euro area have no impact on banks’ stock prices.
- Literature points to slight increases in run risks, with most analysis theoretical; Bidder, Jackson, and Rottner (2024) is noted as an empirical exception.

*International Monetary Fund*

### introduction of CBDC increases the probability of bank runs by 1.2 percentage points, to 2.5 percent.

### introduction of CBDC increases the probability of bank runs by 1.2 percentage points, to 2.5 percent

### Main finding on run probability
- Introduction of a CBDC increases the probability of bank runs by 1.2 percentage points, to 2.5 percent.
- The authors characterize this change as an increase from an event roughly every 70 years on average without CBDC to roughly every 40 years on average with CBDC.

### Empirical evidence and calibration
- Empirical studies cited:
  - Carapella and others (2024)
  - Infante and others (2024)
- These studies find that past crisis episodes have produced large reallocations across NBFIs (for example, outflows from prime MMFs into government MMFs) but have shown limited reallocation toward central bank liabilities in the context of the Federal Reserve’s Overnight Reverse Repurchase Facility.
- The literature indicates that with some optimally calibrated holding limits for CBDC, the increase in the probability of bank runs can be negligible.

### Five key factors affecting the impact of CBDC on banks
- (1) A more competitive banking sector would see higher outflows of deposits than found in the earlier models.
- (2) Banks’ strategic levers to compete with CBDC (such as banks bundling services with deposit offerings) might decrease the impact of CBDC on deposit outflows.
- (3) Higher reliance on deposit funding is likely to increase the impact of CBDC on bank profits.
- (4) The availability of efficient alternative funding sources for banks would soften the impact of CBDC on banks’ balance sheets.
- (5) Lower competition in lending markets would allow banks to raise lending rates more easily to support margins.

### Limitations and additional channels
- Results from the literature are relatively benign overall, but inclusion of more transmission channels could lead to stronger effects on financial stability.
- The fee income channel is largely absent from the models surveyed.
- Some channels could reinforce each other (see the “Conceptual Framework” section).
- The willingness of the central bank to provide more reserves to banks would soften the impact of introducing CBDC on banks.
- Countries with deep and efficient local capital markets, foreign investors, and favorable risk profiles could see banks tapping private funding sources at potentially low costs; conversely, the cost of alternative sources of financing could be high for banks when local capital markets are undeveloped.

*Source: ftnea2025008 - introduction of CBDC increases the probability of bank runs by 1.2 percentage points, to 2.5 percent.*

### References Outcome Variables

### ftnea2025008 - References Outcome Variables

### Quantitative findings from reviewed studies (liability, asset, run-risk, and fee income channels)
- Liability channel: reported impacts and assumptions from selected studies
  - Adalid and others (2022): "Banks need refinancing by the central bank only in high demand scenario"; Scenario-driven, quantity changes with scenario (0.5–18 percent of deposits); Euro area.
  - BIS (2021): "0.05–0.3 pp decrease in RoE"; Scenario-driven, quantity changes with scenario (5–25 percent of deposits); Mix of G7 economies, Sweden, and Switzerland.
  - Bouis and others (2024): "3–8 pp decrease in RoE"; "0–17 pp decrease in NSFR"; "1–83 pp decrease in LCR"; Scenario-driven, quantity changes with scenarios (6–17 percent of deposits); Mix of euro area, Japan, and the US.
  - Chen and Phelan (2025): "30–40 bps increase in deposit rates"; "5–6 pp increase in probability of banks in distress state"; "3–4 pp increase of probability of banks in crisis state"; Scenario-driven, quantity changes with scenario (9–13 percent of total money supply, M2); US.
  - Chang and others (2023): "40–50 bps increase in deposit rates"; "58–75 percent decrease in profits"; Estimated from the model; US.
  - Garcia and others (2020): "0.02–1 pp decrease in RoE"; Scenario-driven, quantity changes with scenario (5–33 percent of total deposits); Canada.
  - Gross and Letizia (2023) and its applications: "<0.05 pp decrease in RoA in the Euro Area"; "0.11–0.15 pp decrease in RoA in Bahrain"; "0.1–0.6 percentage point decrease in Net Interest Margin in the United Arab Emirates"; Estimated from the model; Euro area, the US, Bahrain, Tunisia, Georgia, Qatar, Dominican Republic, United Arab Emirates, and various others.
- Asset channel: selected findings
  - BIS (2021): "2–20 pp increase in lending rates to maintain profitability unchanged"; Scenario-driven, quantity changes with scenario (5–25 percent of deposits); Mix of G7 economies, Sweden, and Switzerland.
  - Bouis and others (2024): "4–8 percent decrease in loans provided to the private sector"; Scenario-driven, quantity demanded changes with scenarios; Mix of euro area, Japan, and the US.
- Run-risk channel
  - Bidder, Jackson, and Rottner (2024): "CBDC increases the probability of bank runs by 1.2 pp"; Estimated from the model; Euro area.
- Fee income channel
  - Bouis and others (2024): "Fee income decreases by 5 percent in proportion of the drop in deposits"; Scenario-driven, quantity demanded changes with scenarios; Mix of Euro area, Japan, and the US.
- Definitions and notes preserved exactly as in the source
  - Note: "bps = basis points; CBDC = central bank digital currency; G7 = Group of Seven countries; LCR = liquidity coverage ratio; NSFR = net stable funding ratio; pp = percentage points; RoA = return on assets; RoE = return on equity."

### Practical considerations and recommended data collection
- Purpose: "gain a detailed, upfront understanding of the financial system’s characteristics, its components, and the broader macrofinancial environment."
- Examples of guiding diagnostic questions to inform financial stability assessment:
  - "What is the level of competition in the banking system as per metrics such as deposit-policy or deposit-lending rate spreads and concentration measures? Has competition been trending in some direction? How does it compare to peer countries?"
  - "Have cash-in-total money ratios been declining? Are other currencies used to a notable extent?"
  - "How have measures of profitability for the banking system evolved over time (including return on assets and net interest margins)?"
- Recommended data categories and specific items (Table 2)
  - Balance sheet composition of banks
    - Asset side: loans, bonds, other security holdings, reserve holdings
    - Liability side: retail deposits, wholesale deposits, other bond finance
    - Capitalization: "regulatory capital metrics such as CET1/RWA and total capital/RWA"
    - Distinguish financial assets by encumbered versus unencumbered; for the unencumbered, distinguish eligible versus not eligible for central bank borrowing
    - Concentration metrics (total assets and deposits)
    - Reserve holdings (required vs. excess) and reserve borrowing across banks
    - "The data needs to have a time dimension, to examine trends"
  - Profitability of banks
    - Selected P&L components: interest income, interest expense, fee income, bottom-line net income before tax; all by bank, at the banking system level, with a time dimension
    - Interest income over interest-bearing assets, loan interest rates
    - Interest expense over liabilities, deposit rates
    - Detailed fee and commission income/expenses for transaction-flow competition analysis
    - Comparable data for peer countries
  - Central bank balance sheet
    - Asset side: reserve lending to banks, holdings of government and private sector bonds
    - Liability side: reserve holdings of banks, possibly bonds issued by the central bank
    - Judge net reserve (excess reserve) position of the banking system and the central bank
    - Reserve requirements
  - Central bank income/expense and net income (seigniorage)
    - Interest income sources (bond holdings, reserve lending)
    - Interest expenses (issued bonds, banks’ reserve holdings)
    - Cost for maintaining cash
    - Net income, seigniorage redistribution to the sovereign
    - Policy rates: "all corridor components"
  - Other metrics: competition and high-level monetary indicators
    - Competition: Skew and Herfindahl indices, share of the largest three or five banks in total banking system assets
    - High-level monetary indicators: cash-in-total money ratios, extent of use of other currencies, reserve coverage
- Data availability note: "Most of the data indicated in Table 2 should be available through conventional statistics and data repositories maintained by central banks and supervisory institutions."

### Analytical frameworks and model categories
- Three model categories (Table 3)
  - Category #1: Balance sheet analyses
    - Selected references: "Juks (2018, 2020); Bindseil (2020); Adalid and others (2022); Malloy and others (2022); Bouis and others (2024)"
    - Role: "visualization, in either tabular or graphical form, of these interconnected balance sheets"; recommended as a first step; useful to assess net reserve position and balance sheet effects; cannot analyze information and payment system resilience channels.
  - Category #2: Financial models (utility, consumer choice, SFC)
    - Selected references: "Chang and others (2023); Gross and Letizia (2023); IMF (2023a); Li (2023); Meller and Soons (2023); Whited, Wu, and Xiao (2023), Bouza and others (2024); León, Moreno, and Soramäki (2024); Li, Usher, and Zhu (2024)"
    - Role: include banks optimizing deposit remuneration and utility-maximizing consumers; can estimate deposit and lending rate changes, income/expense effects, solvency impacts, and sometimes CBDC demand.
  - Category #3: Macro models (NK, DSGE, OLG, NM)
    - Selected references: "Andolfatto (2021); Barrdear and Kumhof (2021); Jiang and Zhu (2021); Banet and Lebeau (2022); Ferrari Minesso, Mehl, and Stracca (2022); Assenmacher, Bitter, and Ristiniemi (2023); Abad, Nuño, and Thomas (2025)"
    - Role: capture real effects (real GDP, employment) and can analyze most channels except information and payments resilience.
- Use cases and limitations
  - Balance sheet analyses: straightforward, suitable for initial country application; rely on simplifying ad hoc assumptions.
  - Category #2 financial models: useful for compositional effects, deposit/lending rate responses, seigniorage impacts; require advanced modeling expertise and coding.
  - Category #3 macro models: require advanced coding and macroeconomic modeling expertise; need quarterly macroeconomic data.
- Data input considerations by model type
  - Category #2 example inputs: deposit-policy rate spreads, shares of currency-in-circulation, measures of velocity, central bank policy parameters (central bank lending and deposit facility rates, reserve requirements).
  - Category #3: general macroeconomic data, usually at quarterly frequency.
- Additional methodological notes (preserved wording)
  - "The aggregate CBDC adoption scenarios are either directly assumed in an ad hoc manner or calculated by adding up holdings for individual(s) which are based on assumptions regarding CBDC holding limits per household/individual."
  - "The level of utility (value) from holding different forms of money... can be informed by certain benchmarks, for example by equating CBDC base utility... to that of physical currency, or that of deposits."

### Design options and policy measures to mitigate downside risks
- High-level guidance
  - "Design options determine the extent to which CBDC is adopted and, in turn, the magnitude of deposit outflows and the strength of the liability, asset and fee income channels."
  - "There is no one-size-fits-all approach to address financial stability risks. Instead, CBDC design and policy choices should be tailored to unique country circumstances."
- Menu of design levers
  - Price measures
    - "Sufficiently low (or zero) CBDC remuneration rates, fees on CBDC transactions, or fees on conversions to other forms of money."
    - Remuneration can be set to "zero" or "at a sufficiently low level" to favor payment use over store-of-value.
    - Note: "In theory, remuneration could also be negative."
  - Quantity measures
    - "Hard limits on holdings, transactions, and convertibility into CBDC."
    - Useful during initial deployment or stress times; may produce operational hurdles (failed transactions if limits exceeded).
    - Waterfall (and reverse waterfall) mechanisms proposed to link CBDC wallets to commercial bank accounts to automatically transfer CBDC amounts above established limits.
  - Access parameters
    - Choice between retail-only or also wholesale access affects substitution for safe assets and magnitude of channels (asset, liability, run-risk).
  - Two-tier distribution with intermediaries
    - "Involving financial intermediaries in the distribution of CBDC can help offset their potential loss of fees and information."
    - With two-tier distribution, intermediaries could charge fees on CBDC services, offer wallets, programmable payments, and use CBDC data for credit evaluation (information channel).
- Calibration and policy sequencing
  - Tradeoff: "design features that limit or slow adoption may at the same time reduce the benefits that a wider utilization of CBDC could bring, such as financial inclusion."
  - Differential measures: "Design features can vary by type of account holder (such as individuals versus merchants, or small versus large merchants)" and can "differ by CBDC holdings or transaction levels."
  - Tiered remuneration example: "CBDC wallets with higher holding limits are remunerated at lower rates, potentially negative."
  - Time-varying and contingent application: price and quantity measures "can be revised over time" and "can be applied during a transition period" or "applied contingently, such as when signs of financial market stress emerge."
  - In stress scenarios: "central banks should stand ready to provide emerging liquidity support"; contingent convertibility limits could be envisaged as a last resort, particularly "in countries with no deposit guarantees and weak crisis resolution frameworks."
- Cross-policy interactions
  - CBDC could redirect outflows away from stablecoins, potentially improving financial stability if stablecoins are poorly regulated or denominated in foreign currencies.
  - CBDC adoption can improve competition in concentrated payment markets and operational resilience.

### Calibration challenges and operational considerations
- Key calibration challenges
  - Modeling relies on "simplifying assumptions"; bank and consumer reactions "depend on multiple factors, which are likely to change over time."
  - Necessity of forward-looking assessments and ongoing data collection post-launch (holdings, flows between bank deposits and CBDC, demographic characteristics of depositors/CBDC holders in anonymized form).
  - Large amounts of data required to understand user preferences and price sensitivities, especially if limits vary by user type or wallet.
  - Calibration may require fine-tuning over time and balancing stakeholder predictability needs.
- Operational examples and precedents
  - "Central Bank of the Bahamas... increased holding limits for individuals from BSD 5,000 to BSD 8,000 after the pilot phase."
  - ECB use: "ECB has developed a scenario-based assessment of deposit outflows for calibrating limits (ECB 2024)."
- Final practical advice
  - "Balance sheet analyses... are usually straightforward to implement and will not require advanced analytical capacity within a central bank."
  - "Financial models in Category #2... require some more advanced modeling expertise."
  - "Macroeconomic models (Category #3) require advanced coding skills and a deep understanding of macroeconomic modeling."
  - To augment transaction-flow competition analysis, obtain detailed data on banks’ and nonbank PSPs’ fee and commission income.

*Source: Authors.*

### Box 1. Mitigating Financial Risks: A Stock-Take of Central Banks’ Approaches

### Box 1. Mitigating Financial Risks: A Stock-Take of Central Banks’ Approaches to Central Bank Digital Currency Design

### Central banks’ remuneration and holding-limit preferences
- Central banks largely favor nonremunerated CBDC. According to the Bank for International Settlements survey, more than half of central banks do not intend to remunerate CBDC, although some do not exclude the option a priori and consider doing more research into this aspect (Illes, Kosse, and Wierts 2025).
- Most central banks in advanced stages of CBDC research consider holding limits; similar proportions in advanced economies and in emerging and developing economies consider them. There is no consensus on how to calibrate holding limits.
- Research and communication gaps:
  - Calibration should be determined based on prevailing financial stability risks and the objectives of CBDC.
  - Research related to calibration of other quantity or price measures (for example, transaction limits, convertibility or holding limits during stress) is nascent.
  - Less than 30 percent of surveyed central banks in advanced economies are considering transaction limits (Illes, Kosse, and Wierts 2025).

### Two main approaches to calibrating holding limits
- Top-down approach:
  - Desired level of adoption determined at country level, then divided by eligible population.
  - Example: In the euro area it has been proposed limiting CBDC adoption to the current holding of banknotes in circulation (EUR 1–1.5 trillion), corresponding to a holding limit of EUR 3,000–4,000 per individual (Bindseil 2020; Bindseil and Panetta 2020).
- Bottom-up approach:
  - Desired level of adoption determined at the individual level, then aggregated.
  - Example: Bank of England proposes setting holding limits such that the Digital Pound can be used for day-to-day transactions and salary payments. A limit of GBP 10,000 (GBP 20,000) would allow 75 percent (95 percent) of income earners to do so (Bank of England 2023).

### Proposed holding limits in selected countries (as of July 2025)
- Note: Percents represent the share of CBDC holding limits in GDP per capita. TBD = to be determined.
- Table summary (country — project — project phase — Individuals (Lowest Tier) — Individuals (Highest Tier) — Merchants):
  - Sand Dollar (The Bahamas) — Live — B$ 500 (1%) — B$ 8,000 (18%) — B$ 8,000–B$ 1,000,000 (18–2227%)
  - JAM-DEX (Jamaica) — Live — None2 — None2 — None2
  - e-Naira (Nigeria) — Live — NGN 120,000 (3%) — NGN 5,000,000 (141%) — No limit
  - e-Cedi (Ghana) — Live (plans) — Yes (TBD) — Yes (TBD) — Yes (TBD)
  - Digital Euro (Eurosystem) — Live (plans) — TBD3 — TBD3 — 0
  - Digital Pound (United Kingdom) — Live (plans) — GBP 10,000–20,000 (22–43%)4 — — “Significantly higher than individuals”
  - D-Cash (ECCU) — Pilot — None — None — -
  - e-CNY (China) — Pilot — RMB 10,000 (6%) — None — -
  - Digital Rupee (India) — Pilot — INR 100,000 (16%) — -

- Table notes and clarifications:
  - 1 The project phase corresponds to the phase for which the specific design features are considered, not to the current phase of the project. For example, the D-cash design features were those considered during the 2021–24 pilot. Those for the Digital Euro and Digital Pound correspond to the proposed design features in the event these CBDCs were introduced in the future.
  - 2 Limits imposed by wallet providers. Lynk: J$ 0, JNPay: J$50,000 (2 percent of GDP per capita) (lowest tier) and J$100,000 (5 percent) (highest tier).
  - 3 Bindseil (2020) and Bindseil and Panetta (2020) suggest a limit of EUR 3,000 (6% of GDP per capita). Discussions with stakeholders on limits calibration are ongoing, and decisions will be made closer to the launch date.
  - 4 Tiered access will be considered.

### Policy considerations for mitigating financial stability risks
- Design features are a first line of action, but traditional safeguards remain relevant.
- Macroprudential policies can help mitigate many CBDC-related risks (examples):
  - Tools to limit excessive risk-taking.
  - Measures to reduce leverage.
  - Policies to manage maturity and currency mismatches or sovereign–bank interconnectedness.
- Crisis management policies provide a safety net to address run risks.
- Key vulnerabilities and limitations:
  - Gaps in supervision and regulation of the financial system are a key vulnerability to address.
  - Relying solely on traditional policies is likely insufficient: macroprudential policies work imperfectly.
  - Further effort needed to address risks related to market-based funding and NBFIs comprehensively.
- For extensive discussion of CBDC policies, see Bouis and others (2024).

### Channels through which CBDC can affect financial stability (conclusion highlights)
- CBDC, as a liquid, safe, and widely accessible instrument, can compete with bank deposits and other money-like assets.
- Six main interrelated channels (operate through liabilities and assets, fee income, run risk, information flows, payment system resilience) can affect:
  - Bank profitability, solvency, and liquidity ratios.
  - Broader financial sector characteristics: risk-taking behavior, interconnectedness, and resilience.
- Economic significance depends on:
  - Level of CBDC adoption.
  - Country characteristics.
  - CBDC design features.
- Quantitative evidence:
  - Studies are limited and model-dependent.
  - Under middle-of-the-road scenarios, bank profitability may decrease but the ultimate impact on financial stability is likely to be contained, especially in countries with low competition, low reliance on deposit funding, broad access to alternative funding, and innovating banks.
- Recommendation:
  - Carefully analyze CBDC impacts on a country-by-country basis using tools from balance sheet analyses to detailed counterfactual models.
  - Where CBDC may be a concern for financial stability, mitigate risks with well-calibrated CBDC design features and traditional financial stability safeguards.

### Models and tools for CBDC-related analyses (Annex highlights)
- Categorization of tools and models (extended):
  - Category #1 — Balance sheet analyses:
    - CBDC take-up assumptions mostly exogenous; scenarios ad hoc or informed by holding caps.
    - Sometimes address collateral scarcity qualitatively; data-informed integrated balance sheet approaches exist.
  - Category #2 — Financial models (IO, utility, consumer choice, SFC):
    - Examples: Chang and others (2023); Gross and Letizia (2023); IMF (2023a); Li (2023); Meller and Soons (2023); Whited, Wu, and Xiao (2023); Bouza and others (2024); León, Moreno, and Soramäki (2024); Li, Usher, and Zhu (2024).
    - CBDC volumes endogenous in several models.
    - Structural models anchored in micro/macro data; demand functions and price sensitivities estimated.
    - Deposit rate spreads to policy often endogenous (except some models).
    - Mostly nominal in nature; not yet covering full real economic and welfare implications.
  - Category #3 — Macro models:
    - NK and DSGE models: Barrdear and Kumhof (2021); Ferrari Minesso, Mehl, and Stracca (2022); Abad, Nuño, and Thomas (2025).
      - CBDC volumes mostly steered exogenously; deposit rates often exogenous; impulse response analyses common.
      - Conventional DSGE caveats apply.
    - OLG models: Andolfatto (2021); Banet and Lebeau (2022); Kim and Kwon (2023).
      - CBDC volumes mostly exogenous; restrictive regarding bank competition assumptions.
    - New Monetarist (NM) models: Jiang and Zhu (2021); Davoodalhosseini (2022); Williamson (2022); Assenmacher, Bitter, and Ristiniemi (2023); Chiu and others (2023); Keister and Sanches (2023).
      - Reference models may lack banks or deposits; some combined NM-DSGE frameworks exist.

### Model development needs and illustrative applications
- Suggested model developments:
  - Develop “macroeconomic shells” around nominal structural models to analyze pass-through to real economy (real GDP, employment).
  - Enhance liquidity focus to analyze banking system liquidity dynamics in more detail (example: Meller and Soons 2023 integrating liquidity coverage ratio).
  - Combine models with exogenous interest rates with frameworks that endogenize bank interest rates (for example, combine Meller and Soons 2023 with Gross and Letizia 2023).
  - Improve asset and liability segmentation to assess potential collateral shortages under anticipated CBDC demand.
  - Incorporate nonbank financial institutions (for example, money market funds) to examine their influence on balance sheets and liquidity risk dynamics under CBDC scenarios.
  - Adequately reflect banks’ money creation ability; many macro models surveyed do not reflect bank money creation, which is important for accurate assessment.

- Example applications:
  - Stock-flow consistent (SFC) model application (Bouza and others 2024) for Tunisia, Georgia, and Qatar:
    - CBDC demand endogenous; model matches pre-CBDC cash preference and deposit-policy rate spreads.
    - Selected Georgia findings:
      - Banks’ excess reserves get depleted or additional reserve borrowing needed when CBDC demand increases.
      - Deposit interest expenses for banks either rise or fall depending on volume and rate effects; for Georgia volume effects dominated so deposit expenses fall.
      - Banks’ net income (RoA) falls due to higher central bank reserve borrowing costs (assumed constant), but RoA remains positive so solvency ratios do not drop.
      - Central bank net income (seigniorage) follows a hump-shaped pattern as CBDC take-up and CBDC interest vary; net seigniorage remains positive throughout.
      - Monetary policy pass-through increases because of additional CBDC-induced competition; gain is more pronounced for countries with lower initial pass-through intensities.
  - New Keynesian model application (Abad, Nuño, and Thomas 2025) calibrated to the euro area:
    - CBDC volumes steered by changing CBDC utility parameter; both unremunerated and remunerated CBDC scenarios examined.
    - Distinguish operational monetary policy floor and ceiling frameworks.
    - Key quantitative result: assumed CBDC take-up amounting to 14 percent of GDP reduces bank deposits by 11 percent of GDP (remaining 3 percent of CBDC take-up subtracts from cash), while lending drops only by 0.6 percent relative to pre-CBDC outstanding stock.
    - System states by CBDC share:
      - CBDC levels below 4 percent of GDP: system rests in a floor system.
      - Beyond 4 percent of GDP: some banks start borrowing reserves, resulting in a corridor system.
      - CBDC take-up beyond 10 percent: system reaches a ceiling state; banking system structurally borrows from the central bank and interbank market rates are pushed against central bank lending facility rate.

*Source: ftnea2025008 - Box 1. Mitigating Financial Risks: A Stock-Take of Central Banks’ Approaches to Central Bank Digital Currency Design (IMF).*

### Annex Figure 2.3.    Selected Model Results from Abad, Nuño, and Thomas (2025)

### Annex Figure 2.3.    Selected Model Results from Abad, Nuño, and Thomas (2025)

### Model setup and notation
- Source: Abad, Nuño, and Thomas 2025.
- Note: CBDC = central bank digital currency.
- The CBDC take-up is varied by the model user (horizontal axes of the charts).

### Key model results (panel-by-panel summary)
- Panel 1:
  - Cash and deposits drop mechanically as CBDC take-up increases.
- Panel 2:
  - Banks’ excess reserves get depleted and, at some point, the net reserve borrowing position builds up instead.
- Panel 3:
  - Policy rates fall as CBDC take-up increases.
- Panel 4:
  - Deposit rates rise as CBDC take-up increases.
- Panel 5:
  - Total household wealth, defined as cash plus deposits plus CBDC holdings, declines because CBDC is not, in this example, remunerated, so households save less, the model suggests.
- Panel 6:
  - Bank equity exhibits a hump shape; for an explanation of why bank equity has a hump shape in panel 6, see page 26 in Abad, Nuño, and Thomas (2025).

### Interpretive note included in the figure
- The figure shows how varying CBDC take-up mechanically alters the composition of money balances and affects banking reserves, policy rates, deposit rates, and aggregate household wealth in the authors’ model.

*Source: Abad, Nuño, and Thomas 2025. (Annex Figure 2.3., as presented in the source PDF.)*

---


_Source: https://www.imf.org/-/media/files/publications/ftn063/2025/english/ftnea2025008.pdf_