Why Cross-Border Lending Is So Hard—and How Portable Credit Could Fix It

Cross-border lending is hard because creditworthiness does not travel cleanly. A borrower may have strong income, assets, payment history, and credit at home, yet a foreign lender cannot access or interpret the evidence, fit it into local policy, or manage the legal and capital risk. Portable credit aims to connect those missing layers.

This article distills my conversation with Charlie Conner, founder of OpenCastle, about non-resident home buyers, open banking, cross-border underwriting, bank partnerships, and capital risk absorption. Watch the full episode: The Future of Global Credit—How OpenCastle Unlocks Cross-Border Lending.

Key takeaways

• Credit history is usually local even when the borrower’s life is global.
• A lender’s problem is not only missing data; it is whether the data can support a local, enforceable decision.
• Portable credit requires data, analytics, and risk execution—not a universal score alone.
• Open banking improves visibility but does not automatically change a credit committee’s risk appetite.
• A partnership model can let banks, brokers, data providers, and capital partners keep doing the work they already perform well.
• Each market requires local regulatory, data, and underwriting adaptation.
• The system must protect consent, privacy, accuracy, explainability, and the borrower’s path to correct an error.

What is cross-border lending?

Cross-border lending describes credit where important parts of the transaction sit in different countries.

For example:

• A U.S. resident buys a home in Spain.
• A British retiree receives income in the United Kingdom but seeks a mortgage in Portugal.
• A remote worker lives in one country, earns from companies in several others, and applies locally.
• A business owns assets in one market and wants financing in another.

The lender may need to evaluate foreign income, bank accounts, investments, debts, payment history, identity documents, tax records, collateral, and currency exposure. It must then decide how those facts fit local underwriting policy and what remedies exist if the loan defaults.

That is why “show the lender my home-country credit score” is rarely enough.

Why can a qualified borrower become invisible abroad?

Traditional credit infrastructure is organized around local institutions and rules.

A borrower may be easy to understand at home because the lender knows:

• Which data sources are reliable.
• What a particular account or credit record means.
• How income is documented.
• How local employers, banks, and tax systems work.
• Which scoring model fits the product.
• How collateral and collections work.

Abroad, those assumptions break.

The foreign lender may see only fragments. A strong salary can be difficult to validate. An investment account may not connect through the local data provider. A home-country credit score may use a different scale or model. The lender may not know how to compare obligations denominated in another currency.

Conner describes the common result as a credit “haircut”: the lender reduces the loan amount, demands more equity or collateral, prices for uncertainty, or declines the borrower.

The decision may not mean the borrower is weak. It may mean the system cannot prove that the borrower is strong in a form the lender can use.

What is portable credit?

Portable credit is the ability to carry usable evidence of creditworthiness from one market into another.

The word “usable” matters. Portability is not merely exporting a PDF or copying a score. The receiving lender needs to know:

• Who the borrower is.
• Where the information came from.
• Whether it is current and complete.
• Whether it was altered.
• How foreign income, assets, and liabilities map to local policy.
• What currency and timing risks apply.
• What the lender can do if the loan fails.
• Whether the borrower consented to the use.

Conner’s long-term vision is that a person’s access to credit should reflect financial capacity rather than where the person happens to live or apply.

That ideal requires infrastructure, not a single model.

What are the three layers of a portable-credit system?

The OpenCastle model discussed in the episode can be understood as three connected layers.

If any layer is missing, the lender can still say no.

Excellent data with no local underwriting interpretation is a file, not a decision. A positive model output with no approved policy is a recommendation, not a loan. A lender that understands the borrower but cannot absorb or recover from the risk may still reduce the amount.

This is why cross-border credit is a systems problem.

How does OpenCastle’s partnership model work?

Conner explains that OpenCastle chose not to rebuild the entire mortgage value chain.

Banks already know how to fund, originate, service, and hold loans. Mortgage and real-estate brokers already help customers find properties and navigate financing. Data providers already connect particular accounts or markets.

The missing piece, in OpenCastle’s view, is a specialized analytical and capital-support layer for non-resident borrowers.

The simplified flow is:

1. A borrower seeks property and financing through the existing broker and bank ecosystem.
2. The bank identifies a non-resident case that does not fit the normal local file.
3. A cross-border partner obtains permissioned financial evidence.
4. The data is normalized and analyzed against an underwriting framework.
5. The bank receives a clearer view of the borrower.
6. Where required and approved, capital support helps absorb defined risk.
7. The bank makes the lending decision and operates the loan under the applicable structure.

This is a B2B2C model: infrastructure companies serve the financial institutions and intermediaries that serve the borrower.

The advantage is focus. Each participant can remain accountable for a defined part of the transaction instead of one startup trying to become a broker, bank, data aggregator, underwriter, insurer, servicer, and capital provider at once.

Why is open banking important?

Open banking can allow a customer to authorize a regulated provider to access account information through secure interfaces. For lending, that can make recent income, spending, balances, and payment behavior easier to assess.

The United Kingdom’s Open Banking organization notes that these data can support more timely affordability and credit decisions. Its own summary for SME lenders also makes the limit clear: open banking improves the quality and timeliness of information, but it does not change credit policy overnight.

That is exactly the distinction Conner makes.

A bank can gain visibility into foreign cash flow and still hesitate because:

• The data covers only some accounts.
• Asset and debt information is incomplete.
• The borrower spans multiple currencies or legal entities.
• The credit committee has no approved policy for the case.
• The lender is uncertain about recovery in default.
• Regulatory capital treatment is unattractive.
• The institution lacks an experienced partner.

Open banking is a data rail. Lending still requires policy, capital, controls, and execution.

Why can’t one global score solve the problem?

A universal score sounds simple, but it compresses context that different lenders may need.

The same person can represent different risk depending on:

• Loan product.
• Currency.
• Loan-to-value ratio.
• Property and collateral jurisdiction.
• Income stability and source.
• Existing obligations.
• Legal enforceability.
• Interest-rate structure.
• Residency status.
• Data completeness.

A score can support a decision, but it cannot replace the evidence or local policy behind it.

The companion lesson from FreshCredit is relevant here: credit report accuracy matters before a credit score. A portable-credit system must preserve the source, context, and correction path—not merely transport a number.

What value does portable credit create for each participant?

For borrowers

• A fuller assessment of income, assets, debts, and payment capacity.
• Less dependence on liquidating investments to fund a purchase entirely in cash.
• A potential path to a larger or more appropriately structured loan.
• Fewer repeated document requests if evidence can be verified and reused lawfully.

For brokers

• More financeable customers.
• Fewer transactions lost after a property is selected.
• Clearer expectations about documentation and likely loan structure.

For banks

• Access to customers who may be attractive but difficult to interpret with local tools.
• Better evidence and specialized underwriting support.
• A defined partner for risk, data, or capital gaps.

For data and capital partners

• A specific, high-value workflow in which their capability becomes actionable.
• A way to participate without owning the full customer relationship or loan lifecycle.

These are potential benefits, not guaranteed outcomes. Every decision remains subject to the lender, product, borrower, jurisdiction, and risk structure.

What makes cross-border data difficult?

The technical challenge is not simply connecting an API.

Coverage

One provider may see transaction accounts but not investments, pensions, business income, liabilities, or assets held elsewhere.

Normalization

Account types, dates, currencies, identifiers, and reporting conventions differ. Duplicate accounts or transfers can overstate assets or income if not handled correctly.

Identity and entity matching

The system must connect the right person to the right accounts and distinguish personal, joint, trust, and business ownership.

Data quality

Missing descriptions, inconsistent categories, or stale records can change an affordability calculation.

Consent and purpose

The borrower should understand which data is being accessed, for what decision, for how long, and by whom.

Explainability

The lender and borrower need a usable explanation of what affected the decision, especially when an automated model is involved.

Correction

If the source data is wrong, the workflow needs a path to challenge, correct, and recalculate the result.

The best model cannot compensate for an untraceable input.

Why does every country need local adaptation?

Data availability, regulation, credit policy, collateral, servicing, and consumer rights vary by market. Conner describes using a consistent analytical objective while partnering with different providers and adapting the inputs country by country.

That is a practical alternative to pretending one technical stack can be copied everywhere unchanged.

A market launch should answer:

• Which institutions and data sources are authoritative?
• What customer authorization is required?
• Which accounts and assets are accessible?
• Which underwriting variables are lawful and relevant?
• Who is regulated for each activity?
• How are adverse decisions explained?
• Which currency risks apply?
• How are collateral and recovery handled?
• Which party holds the credit risk?
• What audit and retention evidence is required?

This is one reason partnership quality can matter more than raw model sophistication.

What are the major risks?

• Privacy risk: collecting more financial data than the decision needs.
• Security risk: creating a valuable cross-border concentration of identity and account information.
• Accuracy risk: matching the wrong source, borrower, currency, or obligation.
• Model risk: applying a framework outside the population or market for which it was tested.
• Bias and fairness risk: treating missing or unfamiliar data as evidence of poor creditworthiness.
• Regulatory risk: performing a regulated function without the required role, permission, or control.
• Capital risk: underestimating default, recovery, or correlation across markets.
• Execution risk: approving a loan without a workable servicing and enforcement path.
• Partner risk: depending on a data, bank, or capital provider whose coverage or policy changes.

Risk cannot be removed by calling the product “infrastructure.” It has to be allocated, controlled, monitored, and explained.

A builder’s checklist for portable credit

• Define the exact borrower and loan use case.
• Map every jurisdiction and regulated activity.
• Collect only decision-relevant data with clear permission.
• Preserve provenance and update time for each input.
• Normalize currencies, transfers, ownership, and duplicates.
• Validate the underwriting approach in the target market.
• Separate model recommendation from lender decision.
• Define human review and exception handling.
• Give the borrower a correction and appeal path where required.
• Identify who holds and funds the risk.
• Design servicing, monitoring, and recovery before origination.
• Test what happens when a data or capital partner becomes unavailable.

Portable credit is infrastructure, not a feature

The cross-border borrower does not need one more dashboard showing a foreign score. The borrower needs financial evidence to survive a chain of local decisions.

That requires permissioned data, accurate normalization, contextual underwriting, bank policy, risk capital, and execution to work as one system. It is the same reason connected infrastructure matters in revenue operations: a strong component cannot repair a broken handoff by itself.

If your B2B company is building a complex revenue process across disconnected data, CRM, outreach, and reporting tools, see whether Enhance AI Partners is a fit.