Skip to content
Vuong Nguyen avatar Vuong
About
Browse
Archives
Explore
Areas
Leadership
Technology
Fintech
Engineering
Blockchain
Stablecoin
AI/ML
Life
Learning
Creativity

The Invisible Credit: Why Remittance Senders Are the Best Borrowers No One Will Lend To

• 9 min read
Fintech , Blockchain , Stablecoin

Hands holding a phone showing a mobile banking app on a busy market street

Before I knew I’d marry their daughter, I was occasionally sitting in the back of my future in-laws’ boutique money transfer company in Boston, watching them move money between families separated by an ocean. Vietnamese families would come in with cash, sometimes a few hundred dollars, sometimes a few thousand, and my future mother-in-law would arrange for an equivalent amount to reach relatives in Ho Chi Minh City, Hanoi, or Da Nang, where someone on a moped would deliver the cash to the door. She tracked every transaction in a composition notebook, and every transfer landed.

I wasn’t a stranger to what that money meant on the other end. I grew up in post-war Vietnam, where resources were scarce and the economy was still finding its footing. Overseas remittances from family abroad weren’t just helpful. For many households they were the difference between stability and crisis, covering school fees, medical bills, and the kind of financial emergencies that formal institutions weren’t equipped to absorb. I understood the sender sitting across from my future in-laws in Boston. I also understood the family waiting on the other side.

What my in-laws were running is called a hawala network, one of the oldest financial systems in the world. It works on a single principle: trust between nodes substitutes for formal institutional infrastructure. It’s efficient, low-fee, and deeply embedded in community. It also leaves absolutely no record that any credit system can read.

They weren’t credit invisible because they were risky. They were invisible because the credit infrastructure was built for a different kind of financial life than the one they were actually living. The problem isn’t that remittance senders lack creditworthiness. The problem is that the credit system has no infrastructure for reading the financial behavior they already demonstrate. This has bothered me ever since.

What Creditworthiness Actually Measures

Credit scoring is a prediction problem. A lender wants to know: given everything observable about this person, how likely are they to repay? FICO, the dominant consumer credit scoring model, answers that using five inputs: payment history at 35%, amounts owed at 30%, length of credit history at 15%, new inquiries at 10%, and credit mix at 10%.

The model works reasonably well within the system it was designed for. Tradelines, the account-level records that populate credit bureau files, exist only when a regulated financial institution reports your payment behavior under the Metro 2 standard, the data format that bureaus require for tradeline reporting. Your mortgage counts. Your credit card counts. Your monthly $500 transfer to Oaxaca doesn’t. Six years of never missing rent in Pilsen doesn’t. Your entire financial history before you arrived counts for nothing.

From a credit perspective, a monthly remittance is indistinguishable from any other recurring financial obligation. The difference is that one reports to bureaus and the other disappears into a payment system designed for transfers, not credit histories.

Underwriters have tried to work around this. Cash flow underwriting analyzes bank statement inflows and outflows directly and has gained ground in fintech lending, but it only works if the borrower is fully banked in the US. Community development financial institutions (CDFIs) and Latino-serving credit unions do relationship-based underwriting that informally accounts for transfer behavior. It’s the right approach at the individual level. It doesn’t scale.

The model adoption problem compounds everything. VantageScore 4.0, the bureau-owned alternative to FICO, and FICO 9 have both made real progress on alternative data. Many mortgage lenders still run FICO 2, 4, and 5, models from the late 1990s built before the alternative data conversation existed. A remittance tradeline that meaningfully lifts a VantageScore 4.0 may do nothing for a borrower applying for a Federal Housing Administration (FHA) loan through a lender running FICO 5. The data problem and the model adoption problem are separate failures that have to be solved in sequence.

LATAM Corridor: Where Credit Invisibility Lives at Scale

The US-to-Mexico corridor is the largest bilateral remittance corridor in the world, with over $63 billion sent in 2023 and a record $64.7 billion in 2024, exceeding Mexico’s tourism revenue and foreign direct investment combined. The US-to-Guatemala corridor sends approximately $20 billion annually, roughly 20% of Guatemala’s entire GDP. El Salvador, Honduras, and the Dominican Republic are similarly dependent at 20 to 25% of GDP.

The senders aren’t a marginal population. Many are concentrated in construction, agriculture, hospitality, and domestic care, industries where income is real and employment is stable but documentation structures don’t produce the W-2 paper trail that bureau-dependent underwriting requires. A credit card issuer will approve a borrower with six months of repayment history but ignores eight years of remittance payments. The system just can’t see it.

The receiving side is even further in the dark. Buró de Crédito and Círculo de Crédito in Mexico skew toward formally employed urban populations. Neither systematically ingests remittance inflow data from US operators. The woman in Oaxaca who has managed household finances across two currencies for a decade, timing peso conversion, absorbing FX volatility, keeping a family operational on a monthly wire, has demonstrated real financial sophistication. I grew up around people like her. That competence is invisible to every credit system she’ll ever interact with.

Three Infrastructure Chasms, Not One

Most proposed solutions treat this as a single problem. It’s three.

Who’s Building, and What’s Still Missing

Nova Credit’s Credit Passport translates foreign bureau records into US-equivalent formats American lenders can underwrite against. For migrants with formal urban employment history, it works. For those from countries with thin bureau coverage, it reaches the people who need it least.

Esusu built around rent reporting, getting the largest recurring obligation most Americans carry finally onto a tradeline. For immigrant communities in renter-heavy urban markets, it directly addresses the first domestic barrier to building credit history.

Kredete is the most structurally interesting player, even though its primary focus is the African diaspora. The flywheel is worth studying: stablecoin settlement rails using USDC that push transfer costs under a dollar, active bureau furnisher relationships, and a secured credit card that also reports to bureaus. They’ve processed $500 million in remittances and report average US credit score improvements of 58 points. No equivalent company is operating at that scale in the US-to-Latin America (LATAM) corridor. That company doesn’t yet clearly exist.

Meanwhile, Remitly and Wise sit on years of verified transfer history across the largest corridors in the western hemisphere. Neither has made a serious move into credit reporting. They’re competing on fees while the credit access opportunity accumulates in their transaction databases.

The most intriguing long-term possibility is using onchain transaction history directly as a credit signal. Blockchain-based remittance rails generate immutable, timestamped records that a borrower could theoretically permission a lender to verify without any intermediary’s attestation. The infrastructure argument is sound. My concern is that onchain data, without guardrails, isn’t inherently reliable. The history of volume stuffing, wash trading, and manufactured transaction activity on public chains means raw onchain history carries real signal noise. The data format is promising. The verification layer still needs work before any serious underwriter should trust it at face value.

What That Composition Notebook Was Actually Tracking

My mother-in-law’s notebook was a credit ledger. It tracked who sent money, how much, how reliably, over how many years. It tracked who in the community could be trusted with a large transfer and who needed more careful handling. It was, functionally, a better behavioral credit database for that community than anything the formal system maintained.

Growing up in Vietnam, I saw remittances arrive and keep families afloat. Years later I watched my future in-laws facilitate those same flows from the other side of the corridor, one composition notebook entry at a time. I’ve never had to argue abstractly that this financial behavior is real, consistent, and creditworthy. I lived it on both ends.

The tragedy isn’t that my in-laws kept that notebook instead of using a bureau-connected system. The tragedy is that the data dissolved when those networks wound down, leaving no credit trail behind, for them or for the millions of people running similar informal ledgers across the United States today.

The companies working on this are doing genuinely important work. But the remaining gaps aren’t the result of insufficient fintech ambition. They’re infrastructure problems that require regulatory coordination and standards-setting that no single company can deliver. The data has always been there. The system just wasn’t built to read it.

If the repayment history already exists, what exactly is the credit system protecting by refusing to read it?

References