Cash-Funded Cross-Border P2P: What Happens to the First Mile Over the Next Five Years

Cash is not leaving remittances. It is changing jobs. Funding a transfer with bills at an agent is in structural decline. Collecting bills at an agent on the other end is declining more slowly, and in some corridors it will still be the product that pays for the network in 2031. Providers that treat “cash” as one thing will mis-price both ends.

September 2026  ·   8–10 minute read

Is Cash Is Still King?

Separate cash-in from cash-out before you forecast anything

Most “cash is dying” slides collapse three different journeys into one number:

•      Cash-funded, cash-paid. The classic agent-to-agent remittance. Still half of Central American transactions in 2025, and only 9% of that region’s remittances were digital at both ends.

•      Cash-funded, digital-paid. Sender walks into a store with wages in an envelope; receiver gets a bank deposit, wallet credit, or card load.

•      Digitally funded, cash-paid. The growing hybrid: app or card on the send side, agent pickup on the receive side because the grandmother does not have—or does not want—an account.

Inter-American Dialogue data for 2026 put Latin American remittance initiation at 53% digital and 47% cash. That sounds like a tipping point. It is not a funeral. The same customers switch channels month to month.

U.S. Treasury and IRS estimates still put 30–36% of U.S. money-service-business remittances as cash-funded. Digital has won a large share of the interface. It has not removed cash from the transaction.

Mexico shows how fast the receive side can move when the plumbing exists. CEMLA, using Banco de México data, reports that account deposits overtook cash payout for electronic remittances: 50.4% of inflows were deposited to accounts in 2025, rising to 53.3% in the first half of 2026. Cash payout fell by $5.3 billion in 2025 and another $728 million in H1 2026 versus the prior-year period. That is the receive-side story in the most competed corridor on earth. It is not the story in rural Guatemala, Honduras, or large parts of the Sahel in Africa.

Why cash funding is expensive—and still used

World Bank Remittance Prices Worldwide for Q3 2025 is unambiguous on price. Funding a $200 transfer with cash averaged 7.01%. Card funding averaged 4.39%. Non-digital remittances as a group averaged 7.30%; digital remittances averaged 4.59%; digital-only MTOs averaged 3.54%. Banks, often used as a stand-in for “formal,” remain the worst of all at 14.99%. A decade ago digital was the expensive option. That relationship has reversed.

The cost is not in the middle of the pipe. Intermex’s 2025 disclosure is the cleanest public look at the stack: $23.8 billion moved across 53.9 million remittances, $26.6 million in bank charges (about 49 cents a transfer), and $352 million paid to sending and paying agents—more than thirteen times the bank bill. Western Union has said working capital parked in destination markets can cost 6–8% a year. Agent commission plus vault cash plus AML at the counter is the P2P cost structure.

Replacing SWIFT does not fix it.

Replacing the envelope at the counter does.

People still fund with cash for reasons that do not yield to a better app:

•      They are paid in cash. Informal and mixed-formal work remains common among the classic remittance sender.

•      They fail, or fear, KYC. An agent counter with a matricula, passport, or familiar clerk is still easier than a liveness check that rejects the ID they have.

•      They want the receipt and the human. First-generation senders, larger tickets before a school term, and emergency sends still skew retail.

•      The receiver needs cash tomorrow morning in a place where the nearest bank is a bus ride and the nearest agent is the pharmacy.

•      Policy can cut both ways. A 1% U.S. federal tax on certain cash remittances took effect on 1 January 2026. Digital and most crypto-network transfers were largely carved out. That is an explicit price penalty on the cash-funded first mile.

How cash funding looks in each region

Region	Cash-funded first mile today	What changes by 2031
North America	30–36% of U.S. MSB remittances still cash-funded. Retail agents (grocery, post, pharmacy) remain the on-ramp for unbanked and cash-paid workers. 1% cash remittance tax now live.	Cash-in share falls fastest here. Tax + enforcement at retail + prepaid-card substitution accelerate the shift. Cash-in survives as a smaller, higher-priced specialty for the remaining unbanked and for senders who will not pass remote KYC.
European Union	Intra-SEPA P2P is already account-to-account. Cash funding concentrates on extra-SEPA migrant corridors (North and West Africa, parts of South Asia) and on senders outside the banking system.	Cash-in becomes a thin corridor product. SEPA Instant and EMI accounts take almost all intra-EU volume. Extra-SEPA cash-in declines with account penetration among the second generation, not with a new brand.
APAC	Two different “cash” stories: MTO agent cash in Gulf/ANZ send markets, and mobile-money cash-in/cash-out inside receive markets (PH wallets, South Asian agents). India receive is increasingly account/UPI, not an envelope.	Wallet cash-in/out stays relevant longer than MTO cash-in. Philippines and parts of ASEAN keep a cash last mile. India stays a licensed-fiat, account-first receive market. GCC send cash-in erodes as wage digitalization spreads.
LATAM	The cash core. 47% cash initiation region-wide; 50% cash-at-both-ends in Central America (2025). Mexico payout has already flipped to accounts. Highest take rates in C2C sit where physical cash still dominates.	Mexico and urban South America keep shifting payout to accounts and PIX-like rails. Central America remains the longest cash-at-both-ends holdout. Intra-LATAM and non-Mexico U.S. corridors are where cash economics stay attractive longest.

North America: policy is doing what product did not

The United States is the world’s largest remittance sender. It is also the market where cash funding is being taxed and policed at the same time it is being out-competed. Analysts covering Western Union in 2026 pointed to the cash remittance tax and immigration enforcement at retail locations as accelerants of an already-running digital shift. Western Union itself reported a pickup in prepaid-card sales at newly enabled agent locations in early 2026 and tied part of that to the tax: more than 1,000 locations selling cards, with 60% of newly loaded cards used to send a Western Union remittance.

That is the tell. Cash funding is not only losing to Remitly and Wise. It is being rerouted through a stored-value product that still uses the agent counter. The sender still walks into Kroger or Vallarta. The transfer is no longer a cash-in money transfer in the legal or tax sense. Over five years, expect the visible “cash at the register” volume to keep falling while the physical location remains useful as a KYC desk, card loader, and receive-side cash vault.

Western Union’s answer has not been to abandon retail. It signed exclusive multiyear deals with Deutsche Post, Canada Post, Vallarta Markets, and renewed Kroger, and it is buying Intermex—the cash-heavy U.S.–LATAM specialist—for about $500 million. That is a bet that the remaining cash pool is still worth consolidating, even if the pool is shrinking. New York’s review of the deal focused on retail access and prices precisely because the cash send market in some LATAM corridors is still concentrated enough to matter.

European Union: cash is already the exception inside the bloc

For euro-to-euro household transfers, cash funding is a rounding error. SEPA Instant across 41 countries made the default journey an IBAN and a phone. Cash-funded extra-SEPA flows still exist—Paris to Dakar, Madrid to Rabat, Milan to a secondary city in South Asia—but they are a corridor business, not a regional one. MiCA and EMI licensing pull the next generation of those senders into accounts, not into a better agent. The five-year cash-in question in Europe is how fast the remaining extra-SEPA migrant senders bank, not whether a new MTO rebuilds a cash network.

APAC: do not confuse mobile-money cash with MTO cash

GSMA’s 2026 mobile-money report is a reminder that “cash” in Africa and parts of Asia often means cash-in and cash-out of a digital wallet, not a Western Union form. Globally, cash-based transactions were still 37% of mobile-money value in the December 2025 snapshot, with cash-in 21% and cash-out 16%. That is domestic and regional wallet plumbing. It will fade more slowly than MTO cash-in from a Gulf salary office, because it is how the receive-side informal economy still works.

On formal remittance corridors into India, the first mile is already moving toward accounts and UPI-linked products; the receive side is not waiting for an envelope. The Philippines is the mixed case: GCash and similar wallets take a rising share of payout, while OFW senders in cash-heavy host markets still use agents. Direct crypto remittances remain about 1% of formal Philippine flows. Cash-out of a wallet or an MTO remains the relevant “cash” product, not a Bitcoin send.

LATAM: the region that decides whether cash funding has a second act

FXC Intelligence notes that C2C take rates are highest in Latin America and Africa, where physical cash still dominates. That is why the region is both the biggest digital-migration story and the last large cash-funding franchise. Mexico’s payout flip to accounts is real and continuing in 2026.

Central America is not Mexico. Half of remittances there were still cash at both ends in 2025. Honduras, Guatemala, El Salvador, and Nicaragua also posted the fastest inflow growth in 2025, which means the cash-heavy countries were adding volume even as Mexico’s mix went digital.

U.S. labor-market and immigration policy is the swing factor. If outbound undocumented and mixed-status senders shrink, cash-in shrinks with them—those senders are over-indexed to agents. If they stay and keep earning, Central America remains a cash-at-both-ends market into the early 2030s, with a rising hybrid layer on top.

Five-year outlook for cash-funded transfers

Metric	2025–26 snapshot	2031 directional view
Cash as a funding method (global mix)	Minority in NA/EU/high-income APAC; ~47% of LATAM initiations; 30–36% of U.S. MSB remittances	Clear minority everywhere except parts of Central America and selected SSA/fragile corridors. Global cash-in share likely in the low-to-mid 20s of remittance transactions, lower by value.
Cash as a payout method	Just lost majority in Mexico electronic remittances (53.3% to accounts in H1 2026). Still dominant in much of Central America and rural SSA.	Account and wallet payout become the default in Mexico, urban Andes, and PH cities. Cash pickup remains a paid feature, not the standard path, except in thin last-mile markets.
Cash-at-both-ends journeys	50% of Central American remittances (2025); already uncommon in Mexico and intra-EU	Becomes a specialty corridor product. Still material in absolute dollars in CA and SSA. No longer the industry’s center of gravity.
Price of cash-funded $200	7.01% cash funding vs 4.39% card; 7.30% non-digital vs 4.59% digital (WB Q3 2025)	Spread stays wide. Cash-in priced as a premium service (tax + agent commission + AML). Hybrid digital-send/cash-pickup cheaper than cash-at-both-ends but still above pure A2A.
Agent economics	WU ~360k locations with activity in the prior 12 months. Intermex paid $352M to agents vs $26.6M to banks. WU buying Intermex and locking retail exclusives.	Fewer, larger retail partners. Agents sell transfers plus prepaid, bill pay, and card load. Commission per transfer under pressure; revenue per square foot defended by adjacent products.
Stablecoins vs cash-in	On-chain remittance share still low single digits. Incumbents using tokens for agent settlement and cards, not to retire the counter.	Stablecoins attack prefunding and agent-to-agent settlement more than they attack the sender’s envelope. A cash-in customer in 2031 may still hand over bills and receive a transfer that settled on a token rail.

The base case is not zero cash in 2031. It is a smaller, more expensive, more consolidated cash-in market sitting underneath a much larger digital-origination market—and a cash-out market that shrinks later than cash-in because receive-side inclusion lags send-side smartphones. Predictions from 2018 that agents would be gone by 2023 were wrong for the same reason a 2026 prediction of “cash gone by 2029” will be wrong: the last mile is a labor and trust market, not an app market.

What does change is the profit pool. Ken Research–style industry models already have digital channels rising from roughly half of provider revenue in the mid-2020s toward about two-thirds by 2031. MoneyGram’s experience is the cautionary chart: digital mix from 30% of transfers in 2021 to 70% in 2025, while MoneyGram Online still produced only 25–30% of revenue and money-transfer fee revenue fell 10% in 2025. Western Union in Q2 2026 saw branded-digital transactions up 25% and revenue up 6%. Cash-funded transfers are the high-take-rate slice. Losing them faster than you replace the revenue is the incumbent problem of the next five years.

Strategies for financial institutions

•      Do not run cash-in on the same P&L as A2A. A 7% cash-funded product and a 2% account-to-account product are different businesses. Bundle them and you will either overprice the banked customer or starve the agent network.

•      Decide whether you are an on-ramp or a last mile. Banks that want the unbanked sender should partner for retail cash-in rather than pretending a branch teller line will compete with Kroger on hours. Banks that want the receive-side deposit should fight for account payout, which is what already flipped Mexico.

•      Treat the 1% U.S. cash remittance tax as a product brief. Prepaid load, account opening at the agent, and debit funding are the legal substitutes customers are already using. If you do not offer one, the MTO will.

•      Use cash-in as a KYC moment, not only as a transfer. The sender who still brings cash is the customer most likely to need a first account, a card, and a receive-side product for the family. That is the FI’s edge over a pure MTO—if the handoff exists.

•      Assume cash-out demand outlives cash-in demand. Even as U.S. senders digitize, receivers in thin markets will want bills. Correspondent and vault-cash partnerships in those markets remain a bank-relevant service.

Strategies for payment platform providers

•      Build the hybrid ticket as the default SKU. Digital fund / cash pickup and cash fund / account deposit will outnumber cash-at-both-ends well before 2031. One transaction ID, both rails, or you lose the switcher—the actual LATAM customer.

•      Price cash-in as a feature. Show the cash-funded option. Make it more expensive on purpose. Use the spread to pay agents without hiding it in FX. Customers who can fund by card already know they are being charged for the envelope.

•      Consolidate agents rather than adding them. The winning cash network of 2031 is fewer doors, longer hours, and more SKUs per door—transfer, card load, bill pay, pickup. WU’s exclusive grocery and post deals and the Intermex purchase are that strategy. Copy the logic in the corridors you own.

•      Put stablecoins behind the agent, not in front of the sender. Prefunding and agent settlement are where tokens change cash economics (WU’s own framing for USDPT). A cash-in customer should not have to learn a chain. They should get a cheaper, faster ticket because you no longer park 6–8% working capital in every payout country.

•      Do not harvest the cash-in customer and strand them. Every cash-funded send is a chance to issue a card or wallet for the next send. If the second transaction is still cash because you never offered the substitute, you paid agent commission twice.

•      Watch Central America and SSA separately from Mexico and India. Portfolio-level “cash is 40% and falling” averages hide the corridors where cash-at-both-ends is still half the book. Those corridors still justify agent capex. Most others do not.

What would falsify this view

Three developments would make cash funding decline faster than the base case: a broader U.S. remittance tax that hits prepaid and debit the way the 2026 tax hits cash; a sharp drop in new migrant inflows into the U.S. and southern Europe; or receive-side account mandates that make cash payout legally awkward. Three developments would make cash funding stickier: a retreat from agent-level KYC that pushes senders back to cash, a failed FedNow/FPS cross-border rollout that leaves digital last-mile patchy, or a new wave of cash-paid informal work in sending markets.

None of those change the structural point. Cash-funded P2P over the next five years is a shrinking share of a still-growing C2C pool, a rising share of remaining MTO margin, and a logistics problem that digital specialists still rent from someone with a vault and a night deposit. The envelope at the counter is not the future of cross-border P2P. It is also not gone.

The firms that price it as a last-mile utility—and migrate the second transaction off it—will take what profit is left. The firms that wait for cash to disappear will keep paying 13 times more to agents than to banks and call it tradition.

Notes and sources

Cash vs digital mix: Inter-American Dialogue (LATAM 53% digital / 47% cash initiation by 2026; Central America 2025 cash-at-both-ends 50%, digital-at-both-ends 9%); U.S. Treasury/IRS estimates of 30–36% cash-funded MSB remittances. Mexico payout: CEMLA / Banco de México—account deposits 50.4% of electronic remittances in 2025 and 53.3% in H1 2026. Costs: World Bank RPW Issue 54, Q3 2025—cash funding 7.01%, card 4.39%, non-digital 7.30%, digital 4.59%, digital-only MTO 3.54%, banks 14.99%. Agent vs bank cost: Intermex 2025 ($352M agent commissions vs $26.6M bank charges on $23.8B / 53.9M transfers). WU network and strategy: FY2025 10-K (~360k active locations; USDPT for agent settlement); Q4 2025 earnings (retail exclusives, prepaid lift, Intermex). U.S. 1% cash remittance tax in effect 1 Jan 2026. Mobile-money cash-in/out: GSMA SOTIR 2026. Take-rate geography: FXC Intelligence 2026 C2C mapping.

Companion piece to “The Next Five Years of Cross-Border P2P” (September 2026). Directional, not investment or legal advice.