How Money Moves Today - The Plumbing of Global Finance

There are approximately 11,000 banks in the SWIFT network. If every bank needed a direct relationship with every other bank, that would require over 60 million bilateral relationships—each with legal agreements, technical connections, and ongoing maintenance.

This is obviously impractical. Instead, banks use correspondent banking: relationships with a smaller number of well-connected banks that can route payments to their destinations.

The Hierarchy:

Tier 1 - Money Center Banks: The largest global banks (JPMorgan, HSBC, Citibank, Deutsche Bank, etc.) maintain extensive correspondent networks. They have relationships with thousands of banks worldwide and can route payments almost anywhere.

Tier 2 - Regional Banks: Significant banks that aren't globally connected but maintain relationships with multiple Tier 1 banks. They can access the global network through their correspondents.

Tier 3 - Local Banks: Smaller banks with limited international capabilities. They rely on regional or money center correspondents for cross-border transactions.

The Pyramid Effect:

1. Alice's regional bank
2. A US money center bank (e.g., JPMorgan)
3. An Asian correspondent (or direct Thailand relationship)
4. Grandmother's local Thai bank

Each intermediary extracts fees. Each adds potential for delay. But without this network, Alice's small regional bank would have no way to reach a local Thai bank.

SWIFT (Society for Worldwide Interbank Financial Telecommunication) is often misunderstood. It's not a payment system—it's a messaging system.

When Alice's bank sends a SWIFT message to initiate her transfer, they're sending an instruction—like a very sophisticated email. The actual movement of money happens separately, through the correspondent banking network.

The Confusion:

Because SWIFT messages initiate payments, people often conflate "SWIFT" with "international wire transfer." Headlines about "SWIFT alternatives" or "bypassing SWIFT" often confuse messaging capability with settlement capability. You can replace SWIFT messaging relatively easily (there are alternatives like CIPS, SPFS, and blockchain-based messaging). Replacing the underlying correspondent banking network is much harder.

Let's trace Alice's $1,000 transfer step by step:

Step 1: Initiation (Day 1, Morning)

• Recipient's name and account number
• Receiving bank's SWIFT/BIC code
• Purpose of payment

Regional Bank of Boston debits Alice's account for $1,000 + $35 fee.

Step 2: SWIFT Message Creation (Day 1, Morning)

Regional Bank of Boston creates an MT103 payment message containing all transaction details. The message is encrypted and transmitted through SWIFT's network.

But here's the critical point: Regional Bank of Boston doesn't have a direct relationship with Bangkok Local Bank. The message must route through intermediaries.

Step 3: Correspondent Processing (Day 1-2)

Regional Bank of Boston has a correspondent relationship with JPMorgan. The payment routes:

1. Regional Bank → JPMorgan (USD)

2. JPMorgan → Bangkok Bank (their Thai correspondent)

3. Bangkok Bank → Bangkok Local Bank

Step 4: Settlement (Day 2-5)

Each step requires actual settlement of funds between the banks involved. This doesn't happen instantly—banks batch transactions and settle periodically. Weekend and holidays add delays.

Step 5: Final Credit (Day 3-5)

Grandmother receives notification that Thai baht has been credited to her account. The journey is complete.

• Alice's bank fee: $35
• Correspondent fees: $25-40
• FX spread: $15-30 (1.5-3% on $1,000)
• Total cost: $75-105 (7.5-10.5% of transfer)

Time: 3-5 business days

These Latin terms confuse many people, but the concept is simple:

Nostro (Latin: "ours") — Our money at another bank
Vostro (Latin: "yours") — Another bank's money at our bank

These are two perspectives on the same account.

Example:

JPMorgan opens a Thai baht account at Bangkok Bank to facilitate payments to Thailand.

• From JPMorgan's perspective: "Our baht account at Bangkok Bank" = Nostro account
• From Bangkok Bank's perspective: "JPMorgan's baht account with us" = Vostro account

It's the same pool of money, just named differently depending on whose books you're looking at.

Here's the crucial insight: for correspondent banking to work quickly, banks must pre-fund their nostro accounts with foreign currency.

When Alice sends $1,000 to Thailand, JPMorgan doesn't wire dollars to Bangkok Bank and wait for them to arrive. Instead:

Pre-funding enables speed at the cost of capital efficiency.

How much money sits in nostro/vostro accounts globally? Estimates vary, but the commonly cited figure is $10-28 trillion.

Conservative estimates: $10-15 trillion in clearly identifiable nostro/vostro accounts

Aggressive estimates: $25-28 trillion including all pre-funded accounts and related trapped capital

For this course, we'll use $15-20 trillion as a reasonable working figure.

This trapped capital isn't earning productive returns. Banks would rather lend money at 8-10% than have it sit in nostro accounts earning 0-3%.

Simple Calculation:

• Opportunity cost: 7% × $15 trillion = $1.05 trillion annually

Even if the real number is half this, we're talking about hundreds of billions in annual opportunity costs—capital that could be financing businesses, home loans, and economic growth.

Why Don't Banks Minimize Nostro Balances?

• Payments fail or delay
• Reputation damage
• Regulatory scrutiny
• Customer complaints

Banks maintain buffer balances above minimum needs. This buffer represents pure inefficiency but provides operational safety.

Not all banks have equal nostro needs. Capital concentrates:

• Top 30 global banks hold 60-75% of nostro capital

• These are the banks that facilitate most international transactions

• Smaller banks rely on correspondents and have minimal direct nostro needs

• USD accounts: $6-8 trillion (largest by far)

• EUR accounts: $3-4 trillion

• GBP, JPY, CNY: $1-2 trillion each

• Other currencies: smaller but significant

• Major financial centers (New York, London, Hong Kong, Singapore, Tokyo) hold disproportionate shares

• Emerging market currencies often route through these hubs

This concentration matters for disruption potential: you don't need to convince 11,000 banks to change. You need to convince the 30-50 banks that control most of the capital.

Let's be clear about what's wrong with the current system:

It's slow: 2-5 days for basic transfers when information moves in milliseconds

It's expensive: 5-10% costs for small transfers, 0.5-2% even for large ones

It's opaque: Senders don't know final cost until transfer completes

It's unreliable: 2-6% of international payments fail or require manual intervention

It traps capital: Trillions sit idle instead of being productively deployed

It excludes: Many countries and individuals lack access to the correspondent banking network

These problems have been documented for decades. Why hasn't the system been replaced?

It actually works. Trillions of dollars flow through the system daily. Yes, there are failures and delays, but the vast majority of transactions complete successfully. Compared to the chaos of international payments before modern correspondent banking, this is remarkable.

It manages risk. Pre-funding eliminates settlement risk—the danger that one party to a transaction defaults before completing their side. Settlement risk has caused major financial crises. The nostro/vostro system, while capital-inefficient, is risk-efficient.

It's regulated. Banks are licensed, audited, and supervised. When problems occur, legal frameworks exist for resolution. Customers have recourse. This regulatory infrastructure took centuries to build.

It handles complexity. International payments involve currency conversion, sanctions screening, tax reporting, anti-money laundering checks, and compliance with multiple jurisdictions. The current system handles all this (imperfectly but functionally).

It's integrated. Banks' systems are built around correspondent banking. Treasury management, risk systems, accounting, regulatory reporting—everything assumes the current model. Changing the payment rails means changing everything connected to them.

Even if everyone agreed that a better system existed, switching would require coordination:

Bank A won't switch until Bank B does. If Bank A adopts new payment rails but Bank B hasn't, Bank A still needs correspondent relationships with Bank B.

Regulators need to approve. Each jurisdiction has banking regulators who must sign off on new approaches. Coordination across 100+ jurisdictions is extremely difficult.

Standards must align. Technical standards, message formats, compliance procedures—all must be compatible across all participants.

Critical mass is required. A new system is only valuable if enough counterparties use it. Until critical mass is reached, everyone stays with the incumbent.

This coordination problem explains why improvements happen incrementally (SWIFT gpi improving tracking and speed) rather than revolutionary replacement.

The banks that run correspondent banking have little incentive to disrupt it:

It's profitable. Correspondent banking fees are significant revenue. Why cannibalize yourself?

It creates competitive moats. A bank with extensive correspondent relationships is valuable. Commoditizing payments reduces this advantage.

It requires expertise. Compliance, risk management, and operational knowledge are barriers to entry. Banks have invested heavily in these capabilities.

It's familiar to regulators. Regulators understand correspondent banking. Novel systems create novel risks that regulators may not approve—or may burden with excessive requirements.

The correspondent banking system isn't uniformly inefficient. Disruption potential varies by segment:

This segmentation suggests that disruption—if it comes—will start at the edges (remittances, emerging markets) rather than the core (interbank settlement, major currencies).

Any system hoping to improve on correspondent banking must:

Match or exceed reliability. The current system processes trillions daily with high success rates. Novel systems need to prove equivalent reliability before institutions will adopt.

Provide clear cost savings. Not 10% better—significantly better. The switching costs are high, so the benefits must be compelling.

Handle compliance requirements. AML, KYC, sanctions screening, tax reporting—all must be addressed. Many blockchain "solutions" ignore these requirements.

Achieve sufficient liquidity. As we discussed in Lessons 1-2, illiquid alternatives are worse than the incumbent, not better.

Manage counterparty risk. The nostro system exists to manage risk. Alternatives must provide equivalent or better risk management.

Navigate regulation. Regulators in multiple jurisdictions must approve. This takes time and relationship-building.

Several approaches are attempting to improve cross-border payments:

• Tracks payments end-to-end

• Aims for same-day settlement on many corridors

• Working within existing correspondent network

• Significant progress but not transformational

• Dollar (or other currency) value on blockchain rails

• Rapid transfer, low fees

• Growing usage for crypto-native transactions

• Regulatory uncertainty limits institutional adoption

• Government-issued digital currencies

• mBridge and similar projects exploring cross-border use

• Early stage, limited interoperability so far

• Massive resources but slow progress

• Various projects (JPM Coin, Fnality, Partior)

• Private or consortium blockchains

• Limited to participating institutions

• Progress but not disruptive yet

• Uses XRP as bridge asset

• Aims to eliminate nostro pre-funding

• Active in some remittance corridors

• Limited adoption, significant potential

We'll explore XRP's approach in detail in later lessons.

✅ Correspondent banking works at scale: Trillions flow through the system daily with high success rates.

✅ Nostro/vostro accounts contain massive capital: Estimates range from $10-28 trillion globally.

✅ The system has significant costs and delays: These are well-documented across corridors.

✅ Coordination problems protect incumbents: Network effects and switching costs create powerful barriers.

⚠️ Exact trapped capital figures: Estimates vary widely; $15-20 trillion is reasonable but not precise.

⚠️ Whether alternatives can meet institutional requirements: Compliance, reliability, and liquidity remain unproven at scale for most alternatives.

⚠️ Timeline for meaningful change: Could be 5 years, could be 50 years.

🔴 Dismissing the current system's strengths: It has survived for centuries because it solves real problems. "Disruption" requires solving those same problems better.

🔴 Assuming technology automatically wins: Superior technology has lost to inferior technology with better network effects many times.

🔴 Ignoring compliance requirements: Many blockchain proposals assume away the regulatory complexity that defines cross-border payments.

The correspondent banking system is genuinely inefficient, with trillions in trapped capital and billions in excess fees. But it's also robust, trusted, and deeply integrated into global finance. Disrupting it requires not just better technology but better solutions to risk management, compliance, liquidity, and coordination. This is possible but far from easy.

Assignment: Create a detailed diagram tracing a cross-border payment through the correspondent banking system.

Requirements:

• All institutions involved (sender's bank, correspondents, recipient's bank)

• SWIFT messages sent at each step

• Actual fund movements (debits/credits)

• Approximate timing for each step

• Sender bank fee

• Each correspondent fee

• FX conversion cost (research current spreads)

• Total cost as dollar amount and percentage

• Cite sources for your estimates

• Which banks likely maintain nostro accounts in the destination currency?

• Estimate how much capital might be pre-positioned for this corridor

• What happens if nostro balances are insufficient?

• How feasible would it be to disrupt payments in this specific corridor?

• What are the main barriers?

• What would success require?

• Diagram accuracy and clarity (30%)

• Cost breakdown research quality (25%)

• Nostro analysis thoughtfulness (25%)

• Disruption assessment realism (20%)

Time investment: 3-4 hours
Value: This exercise forces you to understand the actual mechanics rather than abstractions. The corridor-specific knowledge will be valuable when we discuss XRP adoption patterns.

1. System Architecture Question:

What is the PRIMARY function of SWIFT in international payments?

A) To hold funds in escrow until payment completes
B) To convert currencies at favorable exchange rates
C) To transmit secure, standardized messages between banks
D) To guarantee that payments will not fail

Correct Answer: C

Explanation: SWIFT is a messaging system that transmits payment instructions between banks using standardized formats. It doesn't hold funds (banks do, through correspondent accounts), convert currencies (banks or exchanges do), or guarantee payment completion (that depends on the correspondent banking network). This distinction is crucial—you can replace SWIFT's messaging function relatively easily, but that doesn't solve the underlying settlement challenges.

2. Nostro/Vostro Concept Question:

JPMorgan maintains a Euro account at Deutsche Bank. From JPMorgan's perspective, this account is:

A) A vostro account
B) A nostro account
C) A reserve account
D) A settlement account

Correct Answer: B

Explanation: "Nostro" means "ours" in Latin—it's JPMorgan's money sitting at Deutsche Bank. From Deutsche Bank's perspective, the same account would be called a "vostro" (yours) account. This terminology is counterintuitive but important: the same account has different names depending on whose books you're examining.

3. Pre-Funding Purpose Question:

Why do banks maintain pre-funded nostro accounts instead of transferring money for each individual payment?

A) Government regulations require nostro accounts for all international payments
B) Pre-funding eliminates settlement risk and enables faster payment completion
C) Currency markets are only open during limited hours, requiring advance positioning
D) Banks earn higher interest on funds held in nostro accounts

Correct Answer: B

Explanation: Pre-funding enables banks to credit recipients immediately using funds already positioned at the destination. Without pre-funding, every payment would require actual settlement of funds across borders—taking days and creating risk that one party defaults before the other completes. Pre-funding sacrifices capital efficiency for operational speed and risk management. Option D is actually backwards—nostro accounts typically earn lower returns than alternative uses of capital.

4. System Persistence Question:

Despite obvious inefficiencies in correspondent banking, the system persists. Which explanation is MOST accurate?

A) Banks are too incompetent to build better systems
B) The system provides risk management, compliance infrastructure, and reliability that alternatives haven't matched
C) Government regulations prohibit any changes to international payment systems
D) Customers don't care about fees or delays in international payments

Correct Answer: B

Explanation: The correspondent banking system has genuine strengths: it manages settlement risk through pre-funding, handles complex compliance requirements across jurisdictions, and processes trillions reliably daily. Alternatives must match these capabilities, not just offer better technology. Option A is unfair—banks have invested heavily in improvements like SWIFT gpi. Options C and D are factually incorrect.

5. Disruption Potential Question:

Based on this lesson, which payment segment has the HIGHEST potential for disruption by new technologies?

A) Large interbank settlements between major global banks
B) Corporate payments between established banking customers
C) Remittances and small payments to emerging markets
D) Domestic payments within developed countries

Correct Answer: C

Explanation: Remittances to emerging markets face the highest fees (7-10%), serve populations with weak banking relationships (lower switching costs), and involve illiquid currency pairs (where alternatives might offer genuine improvement). Large interbank settlements are already relatively efficient, and domestic payments have been largely modernized. The edges of the system are more vulnerable to disruption than the core.

For Next Lesson:
We'll examine the case for a neutral bridge currency—why some argue the dollar's dominance as vehicle currency creates problems, what alternatives have been proposed, and what characteristics an ideal neutral bridge would need to have.

End of Lesson 3

Total words: ~5,500
Estimated completion time: 55 minutes reading + 3-4 hours for deliverable