## ftnea2023001

## Source details

**Canonical URL:** [ftnea2023001](https://www.imf.org/-/media/files/publications/ftn063/2023/english/ftnea2023001.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/ftn063/2023/english/ftnea2023001.pdf.md)
- [Structured JSON version](/-/media/files/publications/ftn063/2023/english/ftnea2023001.pdf.json)

---

### Overview of cross-border payment frictions
- Cross-border payments are described as "expensive, slow, and opaque," reflecting multiple frictions.
- Frictions are particularly acute for lower-income countries.
- Key causes of difficulty:
  - Multiple currencies and multiple intermediaries.
  - Thin foreign exchange markets for emerging market and developing economy currencies, leading to exchange rate volatility.
  - Transfers involve parties across jurisdictions subject to different laws and regulations, typically connected through correspondent banking relationships (CBRs).
- Central banks sometimes intermediate cross-border payments, but many central banks "do not trust each other enough to do so," so links between central banks exist mostly for countries that are geopolitically close or share historical connections.

### Purpose and scope
- Explain why frictions related to cross-border payment transactions are pervasive.
- Explore how digital money (stablecoins, tokenized commercial bank deposits, central bank digital currencies—CBDCs) may help overcome or mitigate some frictions.
- Provide a conceptual foundation for the design of markets and platforms to facilitate transfer of digital money across borders.
- Focus on trust networks essential to exchange money and how tokenization affects these networks.
- Complementary to Adrian and others (2022) which advances a vision for a cross-border payments, exchange, and contracting platform (XP); this paper concentrates on trust networks and tokenization for more efficient trading.

### Money and trust networks — key concepts and propositions
- Forms of money:
  - Central bank money: liability of the central bank; includes physical cash and electronic balances held by commercial banks at the central bank.
  - Commercial bank money: balances held by depositors in commercial bank accounts.
  - Nonbank money: balances kept at nonbank financial service providers, including eMoney, stablecoins, and money market funds.
  - Unbacked crypto assets (e.g., bitcoin) are not based on credit relationships and are not liabilities of any entity.
- All modern forms of money are credit-based and rest on claims/liabilities between parties; even central bank money is ultimately backed by public sector sustainability.
- Trust is directional and bi-lateral:
  - Holders must trust issuers (resilience, governance, legal soundness, asset safety/liquidity).
  - Issuers must trust holders (AML/CFT and compliance).
- Fixed and recurring costs of establishing trust: obtaining and monitoring information, vetting counterparties, specialized skills.
- Public policy tools that reduce trust costs:
  - Deposit insurance, regulation, licensing, supervision (to signal safety of issuers).
  - National digital identity systems, sanctions lists, and access infrastructure (to signal trustworthiness of holders).
- Propositions:
  - "Credit-based payment transactions can occur only among people in the same trust network."
  - Corollary: A payment between disjoint trust networks is not immediately possible because the payee will not want to hold the payer’s money and the payer’s issuer will not extend deposit credit to an unknown person.
  - "Payment transactions can occur only over common infrastructure."

### Interoperability in domestic payments — mechanism and implications
- Domestic interoperability mechanism:
  - Bank A debits payer i’s deposit Da and transfers central bank reserves Ra from its account at the central bank to Bank B’s account; Bank B credits payee j’s deposit Da on its balance sheet.
- Two essential pillars enabling domestic interoperability:
  1. A common settlement asset: central bank reserves.
  2. A common platform: the central bank’s real time gross settlement (RTGS) system providing settlement finality.
- Central bank acts as a trust-enhancing mechanism:
  - Both banks trust central bank money (void of default and liquidity risk) and the central bank supervises banks, reducing trust-building costs.
  - Trust-link reduction example:
    - Without a central bank, interoperability among n banks requires n(n-1) bilateral trust relations.
    - With a central bank, trust relations reduce to 2n (between each bank and the central bank in both directions).
- Historical note:
  - Before RTGS diffusion, deferred net settlement created reliance on interbank trust; RTGS reduced intraday credit exposure and shifted settlement to when reserve assets move.

### Cross-border challenges, correspondent banking, and implications for lower-income countries
- Cross-border interoperability is harder because trust links are typically bilateral among correspondent banks rather than anchored by a common settlement asset and platform.
- CBRs have been shrinking in several countries since the global financial crisis.
- Leading explanation for CBR retrenchment:
  - Increased compliance costs associated with AML/CFT regulations, including know-your-customer requirements for money transfer operators.
  - Regulatory uncertainty and perceived risks that banks view as outweighing benefits.
- Consequences and frictions for lower-income countries:
  - Withdrawal of CBRs increases cost of remittances.
  - Most cross-border payments settle using US dollar foreign exchange reserves through CBRs and balance transfers on the books of the US Federal Reserve.
  - During crises (e.g., 2008–09, 2011, 2020), trust erosion can cause interbank foreign exchange markets to show dislocation even among advanced economy banks—similar problems occur for lower-income banks routinely.
- Specific sources of higher costs and restricted access in lower-income countries:
  - Perception of weaker regulatory and supervisory capacity → higher perceived credit risk.
  - Higher compliance costs with AML/CFT due to lower Financial Action Task Force ratings.
  - Higher exchange rate volatility: "9 currencies of emerging markets and developing economies depreciated by more than 25 percent in 2020, and a further 21 fell by more than 10 percent."
  - Pre-funding of nostro balances ("split liquidity") because correspondent banks are not relied upon for credit; if borrowing is possible, risk premia/spreads tend to be high.

### The credit and pre-funding models in cross-border payments
- Credit model (correspondent banking):
  - Bank A debits person i’s account and issues an IOU in currency b to Bank B, which credits person j’s account; Bank B takes counterparty risk and Bank A assumes foreign exchange risk.
  - Balances can be normalized via reciprocal flows or by offloading to foreign exchange dealers that warehouse currency positions and use RTGS to transfer reserves.
  - Establishing bilateral trust links entails substantial sunk costs (counterparty information, monitoring, hedging capacity, compliance), leading to concentration.
  - Empirical example: the concentration ratio of the biggest four banks by turnover stayed above 80 percent throughout the 2010s in the correspondent banking market for the euro.
- Pre-funding model:
  - Bank A pre-funds a wholesale account (Wb) at Bank B.
  - For a payment, Bank A debits Da; Bank B debits Wb and credits Db.
  - Bank A incurs liquidity costs, foreign exchange risk, and counterparty risk.
  - Pre-funding similarly leads to market concentration due to sunk costs and trust requirements.

### Central banks as trust bridges: swaps and reserves
- Central banks can substitute for commercial correspondent banks by providing settlement and foreign exchange access:
  - Credit model replicated via swaps (collateral-backed IOUs) between central banks.
  - Pre-funding replicated via foreign exchange reserves held with foreign counterparts.
- Empirical insights:
  - Where swap lines exist, cross-border payments tend to be cheaper.
  - Swap lines are associated with lower bid-ask spreads in foreign exchange markets.
  - Lower spreads correlate with lower remittance costs:
    - For a $200 remittance, a 10 percent increase in the bid-ask spread level is associated with a $0.15, or 2.4 percent, higher cost of remittances.
    - A 10 percent increase in the bid-ask spread volatility is associated with a $0.19, or 3.0 percent, higher cost.
- Limitations:
  - Swap lines are costly and risky; they tend to be backstops among highly trusted central banks and cluster regionally.
  - Political and practical hurdles limit expansion of centralized swap arrangements.

### Options to improve cross-border payments: global CCP and marketplaces
- Global central counterparty (CCP) clearinghouse for central bank swaps (conceptual benefits):
  - Reduces number of bilateral trust relationships (each central bank trusts the global clearinghouse).
  - Nets swaps, holds loss-absorbing capital, requires collateral, and pools risks.
- Main hurdles:
  - Political difficulty of sovereign risk-sharing.
  - Collateral requirements may be prohibitive for less-liquid currency pairs.
  - Further feasibility, cost-benefit analysis required.

### Digital money, tokenization, and trust reconfiguration
- Tokenized money: property rights to a currency inscribed and traded on a common (permissioned) ledger.
  - Ownership transfers imply settlement trust shifts from counterparties to the network and its governance.
  - Money remains credit-based: redemption criteria (redeemability at face value) and financial integrity (customer due diligence, transaction monitoring) must be satisfied.
- Trust network cost dynamics:
  - Multiple private issuers imply high sunk costs: holding n monies across i users entails costs proportional to 2ni.
  - Gateways (digital wallet providers or issuer-owned wallets) can mutualize trust:
    - With a single gateway, costs reduce to 2(i + n).
    - Gateways verify issuers and users, lowering trust-establishment costs.
  - Public policies (licensing, supervision) and issuance by regulated entities or central banks reduce trust costs further.
- Analogy: gateways play a role similar to banks holding foreign currency, subject to host-country regulatory acceptance.

### Cross-country marketplace model for tokenized money (multiownership model)
- Interoperability problem:
  - If person i holds A-coins and person j holds B-coins, direct transfer requires shared trust (gateways that trust both coins) or ability of end users/gateways to hold keys to both assets.
  - The multiownership model: person j can receive A-coins and, if they or their gateway trust A-coins, be indifferent between holding A- or B-coins.
- Foreign exchange problem:
  - Marketplace enables conversion of coins (A to B) before delivery to recipient.
  - Mechanisms:
    - Coincidence of wants: matched orders executed via smart contracts (e.g., hash-time locked contracts).
    - Market makers: warehouse A- and B-coins and set exchange rates based on supply-demand; facilitate trades when no coincidence exists.
- Advantages over traditional correspondent banking:
  - Market makers are not bound by bilateral credit claims; they can freely trade and net positions using multiownership.
  - Market making becomes more competitive: coins are standardized contracts that any connected market maker can bid to hold—reducing reliance on preferred bilateral issuer relationships.
  - Lower costs due to reduced networking and verification costs.

### Extensions, design considerations, and risks
- Marketplace is agnostic to types of coins; trading CBDCs has advantages:
  - Greater market liquidity, lower settlement risk, fungibility at par.
- Currency pair selection:
  - Marketplaces may gravitate to fewer, more liquid pairs; vehicle currencies (c-currency) can emerge endogenously.
  - Liquidity is key; central bank participation could provide liquidity where markets are thin.
  - Decentralized finance (DeFi) techniques (e.g., liquidity pools) and specially designed auctions are potential tools.
- Fragmentation risk:
  - Digital marketplace liquidity could reduce liquidity in conventional FX markets and risk price fragmentation.
  - Solution: openness to market makers active in conventional FX markets to enable arbitrage.
- Governance and operator questions:
  - International financial organizations, public entities, or regulated private entities could operate marketplaces; public-orchestrated solutions offer coordination, trusted governance, and compatibility with financial integrity standards.
- Operational model examples:
  - Private settlement asset and marketplace (e.g., Ripple’s XRP model).
  - Open-source marketplaces (e.g., Stellar Foundation, DeFi networks).
  - Marketplace and settlement asset based on unbacked crypto assets (e.g., Strike leveraging Bitcoin and the Lightning Network).
- Public solution advantages:
  - Tackles coordination problems, centralizes participation and liquidity provision.
  - Offers trusted governance and operational stability.
  - Provides full compatibility with financial integrity standards.

### Box 2 — Liquidity Pools: setup, mechanics, and numeric examples
- Initial market expectation: 1 USD equals 4,000 KHR.
- Liquidity pool setup:
  - Liquidity suppliers add amounts of USDC and KHRC in the ratio of 4,000 KHRC per 1 USDC to the liquidity pool smart contract.
  - Liquidity providers receive Liquidity Pool (LP) tokens in proportion to their contribution size (example: 1 LP for every 1,000 USD or equivalent in KHR contributed).
  - A 0.3 percent fee on trades is proportionally distributed among all token holders.
  - When a liquidity provider cashes out, she receives her initial liquidity contribution and burns her LP tokens in exchange for the accrued fees.
- Automated market maker pricing rule:
  - QUSDC * QKHRC = constant.
  - Where QUSDC is the amount of USDC in the pool, and QKHRC is the amount of KHRC in the pool.
- Example quantities as presented:
  - Example pool holdings: 100k USDC and 400m KHRC.
  - Stated constant in example: 40t.
  - Swap demand in example: 40m KHRC.
  - The manufacturer must contribute X USDC, where X solves:
    - (100푘+푋)=
      40푡
      400푚−40푚
  - Required USDC contribution (example pool): 11,111.11 plus fees.
  - Implied exchange rate (example pool): 3,600 KHRC to 1 USDC.
- Slippage and pool size sensitivity:
  - The difference between the realized exchange rate and the initial proportions of the pool is called slippage.
  - The larger the size of the pool, relative to the size of the swap, the less the slippage.
  - If the pool is 100 times larger (multiply the amounts of each token by 10), the required contribution of USDC to swap out 40m KHRC is 10,101.01, an implied exchange rate of 3,960 KHRC to 1 USDC.
  - Generally, as the size of the pool increases the slippage goes to zero.
- Key numeric parameters preserved:
  - Initial expected exchange rate: 1 USD = 4,000 KHR.
  - Fee on trades: 0.3 percent.
  - Example pool holdings: 100k USDC and 400m KHRC.
  - Swap demand in example: 40m KHRC.
  - Stated constant in example: 40t.
  - Required USDC contribution (example pool): 11,111.11 plus fees.
  - Implied exchange rate (example pool): 3,600 KHRC to 1 USDC.
  - Required USDC contribution (100× larger pool): 10,101.01.
  - Implied exchange rate (100× larger pool): 3,960 KHRC to 1 USDC.

*International Monetary Fund — FINTECH NOTES Trust Bridges and Money Flows — Introduction (content unit: ftnea2023001)*

### Introduction ...........................................................................................................

### Introduction

### Overview of cross-border payment frictions
- Cross-border payments are described as "expensive, slow, and opaque," reflecting multiple frictions (FSB 2020; CPMI 2020).
- Frictions are particularly acute for lower-income countries.
- Key causes of difficulty:
  - Multiple currencies and multiple intermediaries.
  - Thin foreign exchange markets for emerging market and developing economy currencies, leading to exchange rate volatility.
  - Transfers involve parties across jurisdictions subject to different laws and regulations, typically connected through correspondent banking relationships (CBRs).
- Central banks sometimes intermediate cross-border payments, but many central banks "do not trust each other enough to do so," so links between central banks exist mostly for countries that are geopolitically close or share historical connections.

### Purpose and scope of the paper
- Explain why frictions related to cross-border payment transactions are pervasive.
- Explore how digital money (stablecoins, tokenized commercial bank deposits, central bank digital currencies—CBDCs) may help overcome or mitigate some frictions.
- Provide a conceptual foundation for the design of markets and platforms to facilitate transfer of digital money across borders.
- Focus on trust networks essential to exchange money and how tokenization affects these networks.
- Complementary to Adrian and others (2022) which advances a vision for a cross-border payments, exchange, and contracting platform (XP); this paper concentrates on trust networks and tokenization for more efficient trading.

### Money and trust networks — key concepts and propositions
- Forms of money:
  - Central bank money: liability of the central bank; includes physical cash and electronic balances held by commercial banks at the central bank.
  - Commercial bank money: balances held by depositors in commercial bank accounts.
  - Nonbank money: balances kept at nonbank financial service providers, including eMoney, stablecoins, and money market funds.
  - Unbacked crypto assets (e.g., bitcoin) are not based on credit relationships and are not liabilities of any entity.
- All modern forms of money are credit-based and rest on claims/liabilities between parties; even central bank money is ultimately backed by public sector sustainability.
- Trust is directional and bi-lateral:
  - Holders must trust issuers (resilience, governance, legal soundness, asset safety/liquidity).
  - Issuers must trust holders (AML/CFT and compliance).
- Fixed and recurring costs of establishing trust: obtaining and monitoring information, vetting counterparties, specialized skills.
- Public policy tools that reduce trust costs:
  - Deposit insurance, regulation, licensing, supervision (to signal safety of issuers).
  - National digital identity systems, sanctions lists, and access infrastructure (to signal trustworthiness of holders).

- First proposition:
  - "Credit-based payment transactions can occur only among people in the same trust network."
- Corollary:
  - A payment between disjoint trust networks is not immediately possible because the payee will not want to hold the payer’s money and the payer’s issuer will not extend deposit credit to an unknown person.
- Second proposition:
  - "Payment transactions can occur only over common infrastructure."

### Interoperability in domestic payments — mechanism and implications
- Domestic interoperability solution:
  - Bank A debits payer i’s deposit Da and transfers central bank reserves Ra from its account at the central bank to Bank B’s account; Bank B credits payee j’s deposit Da on its balance sheet.
- Two essential pillars enabling domestic interoperability:
  1. A common settlement asset: central bank reserves.
  2. A common platform: the central bank’s real time gross settlement (RTGS) system providing settlement finality.
- Central bank acts as a trust-enhancing mechanism:
  - Both banks trust central bank money (void of default and liquidity risk) and the central bank supervises banks, reducing trust-building costs.
  - Trust-link reduction example:
    - Without a central bank, interoperability among n banks requires n(n-1) bilateral trust relations.
    - With a central bank, trust relations reduce to 2n (between each bank and the central bank in both directions).

### Cross-border challenges, correspondent banking, and implications for lower-income countries
- Cross-border interoperability is harder because trust links are typically bilateral among correspondent banks rather than anchored by a common settlement asset and platform.
- CBRs have been shrinking in several countries since the global financial crisis (IMF 2016a, IMF 2016b, Rice, von Peter, and Boar 2020).
- Leading explanation for CBR retrenchment:
  - Increased compliance costs associated with AML/CFT regulations, including know-your-customer requirements for money transfer operators.
  - Regulatory uncertainty and perceived risks that banks view as outweighing benefits.
- Consequences and frictions for lower-income countries:
  - Withdrawal of CBRs increases cost of remittances.
  - Most cross-border payments settle using US dollar foreign exchange reserves through CBRs and balance transfers on the books of the US Federal Reserve.
  - During crises (e.g., 2008–09, 2011, 2020), trust erosion can cause interbank foreign exchange markets to show dislocation even among advanced economy banks—similar problems occur for lower-income banks routinely.
- Specific sources of higher costs and restricted access in lower-income countries:
  - Perception of weaker regulatory and supervisory capacity → higher perceived credit risk.
  - Higher compliance costs with AML/CFT due to lower Financial Action Task Force ratings.
  - Higher exchange rate volatility: "9 currencies of emerging markets and developing economies depreciated by more than 25 percent in 2020, and a further 21 fell by more than 10 percent."
  - Pre-funding of nostro balances ("split liquidity") because correspondent banks are not relied upon for credit; if borrowing is possible, risk premia/spreads tend to be high.

### Digital money, tokenization, and the evolving trust network
- The paper studies how trust networks evolve when money is expressed on ledgers commonly available to market participants (examples: stablecoins, tokenized commercial bank deposits, CBDCs).
- Consideration of a global clearinghouse that could eliminate the need for complete sets of bilateral trust relationships.
- The paper advances a model for a marketplace to trade tokenized money directly across borders, arguing efficiency gains:
  - Lower risks and costs.
  - Greater competition.
  - Greater transparency.
  - Improved scalability.
- Scope of subsequent analysis:
  - Outline nature of money and trust networks for domestic systems.
  - Extend discussion to cross-border payments and how information or knowledge gaps complicate trust and credit relationships.
  - Consider ways that digital money may simplify cross-border economic relationships.
  - Present conceptual model for trading/transferring digital money across borders and identify key efficiency gains.

### Key methodological and conceptual links to related work
- The paper complements a vision (Adrian and others, 2022) for an XP platform integrating tokenization, programming, and encryption into infrastructure for cross-border currency exchange and contract writing/trading.
- This paper narrows focus to trust networks essential to money exchange and specifically how tokenization affects and can be leveraged for more efficient trading.

_Italic: International Monetary Fund — FINTECH NOTES Trust Bridges and Money Flows — Introduction (content unit: ftnea2023001)_

### introduction of central banks and before the global diffusion of RTGS systems that began during the

### ftnea2023001 - introduction of central banks and before the global diffusion of RTGS systems that began during the

### RTGS, deferred net settlement, and trust in interbank payments
- Historically, transfers of reserves often occurred at the end of the day (deferred net settlement) while payee accounts were credited during the day, creating reliance on interbank trust.
- Real-time gross settlement (RTGS) changed this: a payee’s account is not credited until the settlement asset moves from the payor’s bank to the payee’s bank.
- In practice:
  - Some banks extend intraday credit by lending central bank reserves over intraday money markets.
  - Levels of interbank trust are volatile and can collapse in stress (interbank money markets nearly froze during the great financial crisis in 2008–09 and 2011).
  - Increased caution toward foreign bank credit exposures and balance sheet constraints contributed to violations of covered interest rate parity.

### Interoperability in cross-border payments: loss of a common settlement asset
- Cross-border payments replace a common settlement platform with a common messaging system (e.g., Swift) that does not move assets.
- Without a common settlement asset held by both banks:
  - Banks must either extend credit to each other (the credit model) or pre-fund cross-border payment needs (the pre-funding model).
  - Correspondent banking chains can become long and costly; some banks (e.g., Bank D in the text example) may lack direct cross-border trust links and must route through others.

### The Credit Model (correspondent banking)
- Mechanics:
  - Bank A debits person i’s account and issues an IOU in currency b to Bank B, which credits person j’s account; Bank B takes counterparty risk and Bank A assumes foreign exchange risk.
  - Balances can be normalized via reciprocal flows or by offloading to foreign exchange dealers that warehouse currency positions and use RTGS to transfer reserves.
- Market structure implications:
  - Establishing bilateral trust links entails substantial sunk costs (counterparty information, monitoring, hedging capacity, compliance), leading to concentration.
  - Concentration example: the concentration ratio of the biggest four banks by turnover stayed above 80 percent throughout the 2010s in the correspondent banking market for the euro.

### The Pre-Funding Model
- Mechanics:
  - Bank A pre-funds a wholesale account (Wb) at Bank B (via customer flows or using a foreign exchange dealer).
  - For a payment, Bank A debits Da; Bank B debits Wb and credits Db.
- Costs and risks:
  - Bank A incurs liquidity costs, foreign exchange risk, and counterparty risk.
  - Pre-funding similarly leads to market concentration due to sunk costs and trust requirements.

### Role for central banks: swaps and reserves as trust bridges
- Central banks can substitute for commercial correspondent banks by providing settlement and foreign exchange access; structures mirror credit and pre-funding models:
  - Credit model replicated via swaps (collateral-backed IOUs) between central banks.
  - Pre-funding replicated via foreign exchange reserves held with foreign counterparts.
- Empirical insights (summarized in the paper):
  - Where swap lines exist, cross-border payments tend to be cheaper.
  - Swap lines are associated with lower bid-ask spreads in foreign exchange markets (Annex A).
  - Lower spreads correlate with lower remittance costs (Annex B):
    - For a $200 remittance, a 10 percent increase in the bid-ask spread level is associated with a $0.15, or 2.4 percent, higher cost of remittances.
    - A 10 percent increase in the bid-ask spread volatility is associated with a $0.19, or 3.0 percent, higher cost.
- Limitations:
  - Swap lines are costly and risky; they tend to be backstops among highly trusted central banks and cluster regionally (United States, European Union, China).
  - Political and practical hurdles limit expansion of centralized swap arrangements.

### Options to improve cross-border payments: global CCP and marketplaces
- Global central counterparty (CCP) clearinghouse for central bank swaps (conceptual benefits):
  - Reduces number of bilateral trust relationships (each central bank trusts the global clearinghouse).
  - Nets swaps, holds loss-absorbing capital, requires collateral, and pools risks.
- Main hurdles:
  - Political difficulty of sovereign risk-sharing.
  - Collateral requirements may be prohibitive for less-liquid currency pairs.
  - Further feasibility, cost-benefit analysis required.

### Tokenized money, gateways, and trust reconfiguration
- Tokenized money: property rights to a currency inscribed and traded on a common (permissioned) ledger.
  - Ownership transfers imply settlement trust shifts from counterparties to the network and its governance.
  - Money remains credit-based: redemption criteria (redeemability at face value) and financial integrity (customer due diligence, transaction monitoring) must be satisfied.
- Trust network cost dynamics:
  - Multiple private issuers imply high sunk costs: holding n monies across i users entails costs proportional to 2ni.
  - Gateways (digital wallet providers or issuer-owned wallets) can mutualize trust:
    - With a single gateway, costs reduce to 2(i + n).
    - Gateways verify issuers and users, lowering trust-establishment costs.
  - Public policies (licensing, supervision) and issuance by regulated entities or central banks reduce trust costs further.
- Analogy: gateways play a role similar to banks holding foreign currency, subject to host-country regulatory acceptance.

### Cross-country marketplace model for tokenized money (multiownership model)
- Interoperability problem:
  - If person i holds A-coins and person j holds B-coins, direct transfer requires shared trust (gateways that trust both coins) or ability of end users/gateways to hold keys to both assets.
  - The multiownership model: person j can receive A-coins and, if they or their gateway trust A-coins, be indifferent between holding A- or B-coins.
- Foreign exchange problem:
  - Marketplace enables conversion of coins (A to B) before delivery to recipient.
  - Mechanisms:
    - Coincidence of wants: matched orders executed via smart contracts (e.g., hash-time locked contracts).
    - Market makers: warehouse A- and B-coins and set exchange rates based on supply-demand; facilitate trades when no coincidence exists.
- Advantages over traditional correspondent banking:
  - Market makers are not bound by bilateral credit claims; they can freely trade and net positions using multiownership.
  - Market making becomes more competitive: coins are standardized contracts that any connected market maker can bid to hold—reducing reliance on preferred bilateral issuer relationships.
  - Lower costs due to reduced networking and verification costs (link to Catalini and Gans 2019).

### Extensions, design considerations, and risks
- Marketplace is agnostic to types of coins; trading CBDCs has advantages:
  - Greater market liquidity, lower settlement risk, fungibility at par.
- Currency pair selection:
  - Marketplaces may gravitate to fewer, more liquid pairs; vehicle currencies (c-currency) can emerge endogenously.
  - Liquidity is key; central bank participation could provide liquidity where markets are thin.
  - Decentralized finance (DeFi) techniques (e.g., liquidity pools) and specially designed auctions are potential tools.
- Fragmentation risk:
  - Digital marketplace liquidity could reduce liquidity in conventional FX markets and risk price fragmentation.
  - Solution: openness to market makers active in conventional FX markets to enable arbitrage.
- Governance and operator questions:
  - International financial organizations, public entities, or regulated private entities could operate marketplaces; public-orchestrated solutions offer coordination, trusted governance, and compatibility with financial integrity standards.
- Three operational models noted:
  - Private settlement asset and marketplace (e.g., Ripple’s XRP model).
  - Open-source marketplaces (e.g., Stellar Foundation, DeFi networks).
  - Marketplace and settlement asset based on unbacked crypto assets (e.g., Strike leveraging Bitcoin and the Lightning Network).
- Public solution advantages:
  - Tackles coordination problems, centralizes participation and liquidity provision.
  - Offers trusted governance and operational stability.
  - Provides full compatibility with financial integrity standards.

*Italic: Content derived from the provided IMF PDF chapter.*

### Box 2. Liquidity Pools

### Box 2. Liquidity Pools

### Liquidity pool setup and incentives
- Initial market expectation: 1 USD equals 4,000 KHR.
- A liquidity pool for USDC and KHRC is established by liquidity suppliers adding amounts of USDC and KHRC in the ratio of 4,000 KHRC per 1 USDC to the liquidity pool smart contract.
- This ensures the pool has an equal dollar value (or Khmer Riel value) of both currencies.
- Liquidity providers receive Liquidity Pool (LP) tokens in proportion to their contribution size (example: 1 LP for every 1,000 USD or equivalent in KHR contributed).
- A 0.3 percent fee on trades is proportionally distributed among all token holders.
- When a liquidity provider cashes out, she receives her initial liquidity contribution and burns her LP tokens in exchange for the accrued fees.

### Automated market maker pricing rule
- Pricing formula: QUSDC * QKHRC = constant.
  - Where QUSDC is the amount of USDC in the pool, and QKHRC is the amount of KHRC in the pool.
- Whenever someone swaps one currency for the other, quantities put into and taken out of the pool must preserve the constant.
- Example statement in text: suppose at the time the swap is initiated, the constant equals 40t and the pool holds 100k USDC and 400m KHRC.

### Example swap calculation (preserving source expressions)
- The US manufacturer needs 40m KHRC.
- The manufacturer must contribute X USDC, where X solves:
  - (100푘+푋)=
    40푡
    400푚−40푚
- The required contribution of USDC is 11,111.11 plus fees for an implied exchange rate of 3,600 KHRC to 1 USDC.

### Slippage and pool size sensitivity
- Definition: The difference between the realized exchange rate and the initial proportions of the pool is called slippage.
- Relation to pool size:
  - The larger the size of the pool, relative to the size of the swap, the less the slippage.
  - If the pool is 100 times larger (multiply the amounts of each token by 10), the required contribution of USDC to swap out 40m KHRC is 10,101.01, an implied exchange rate of 3,960 KHRC to 1 USDC.
  - Generally, as the size of the pool increases the slippage goes to zero.

### Key numeric values and parameters (as presented)
- Initial expected exchange rate: 1 USD = 4,000 KHR.
- Fee on trades: 0.3 percent.
- Example pool holdings: 100k USDC and 400m KHRC.
- Swap demand in example: 40m KHRC.
- Stated constant in example: 40t.
- Required USDC contribution (example pool): 11,111.11 plus fees.
- Implied exchange rate (example pool): 3,600 KHRC to 1 USDC.
- Required USDC contribution (100× larger pool): 10,101.01.
- Implied exchange rate (100× larger pool): 3,960 KHRC to 1 USDC.

*Source: Box 2. Liquidity Pools, Trust Bridges and Money Flows — A Digital Marketplace to Improve Cross-Border Payments (FINTECH NOTES), INTERNATIONAL MONETARY FUND.*

---


_Source: https://www.imf.org/-/media/files/publications/ftn063/2023/english/ftnea2023001.pdf_