Indonesia’s QRIS Boom: One Bank Figure Says 13.66 Billion Payments, a Trade Brief Says 2.5 Billion

Auditing the denominator, the vintage and the fee schedule behind Indonesia’s QRIS numbers, 2025-2026

In the first half of 2026, Bank Indonesia recorded 12.55 billion QRIS transactions worth Rp 1,120 trillion. In a trade brief published in June 2026, New Zealand’s Ministry of Foreign Affairs and Trade put Indonesia’s QRIS volume for calendar 2025 at 2.5 billion transactions worth Rp 262.1 trillion. The second document describes the same national QR system, in an earlier year, at roughly a fifth of the central bank’s count. Both numbers are wearing official clothes. Only one of them can describe what it claims to describe.

This article audits the disagreement rather than picking a winner. The method is the one regulators’ own statisticians use: take each published figure, reconstruct its denominator, its time window and its definition, then test whether the numbers could coexist under any honest reading. The answer here is not fraud or a data error in the usual sense. It is something more common and more costly: a stale vintage circulating under a current-year label, plus a fee schedule whose headline contradicts its own fine print. The distinction matters, because a reader who mistakes an old vintage for a growth collapse will make a different decision from one who understands the labels.

What each document actually says

Start with the central bank, which is the system operator and the natural source of record. At its December 2025 Board of Governors Meeting, Bank Indonesia reported full-year 2025 QRIS figures: approximately 13.66 billion transactions, around 59 million users, and roughly 42 million merchants, of which about 90 percent are micro, small and medium enterprises. The transaction count ran at more than double the bank’s own annual target of 6.5 billion. These figures were reported publicly and picked up across the payments press; they supersede the bank’s earlier mid-year data, which had shown 10.31 billion transactions through September 2025 alone.

The other document is narrower. MFAT’s June 2026 trade brief, in the row its authors labeled 2025, carries two QRIS numbers: 2.5 billion transactions and Rp 262.1 trillion in value. The brief is a trade-promotion document, not a payments statistics bulletin, and its payment rows cite aggregated sources. Bank Indonesia’s own published series, from its Laporan Pelaksanaan Sistem Pembayaran (payment system report) and its August 2026 release, gives a different and internally consistent trail: 55.43 million registered users at the end of 2024, 59.53 million at the end of 2025, 65.77 million by June 2026, and merchant counts of 42.75 million at end-2025 rising to 44.86 million by June 2026, with the MSME share at 96.68 percent in the mid-2026 reading.

Deputy Governor Filianingsih Hendarta put the bank’s forward frame on the record in January 2026: “In 2026, we will continue to drive the expansion of QRIS by embracing the theme of independence, represented by 17, 8, and 45,” referring to targets of 17 billion transactions, cross-border acceptance in eight countries and 45 million merchants (Marketing-Interactive, January 2026). Nobody announcing a 17 billion target for 2026 believes the system did 2.5 billion in 2025. The two documents are not describing the same year with different methods. Something else is going on.

Table 1: Three QRIS readings, their labels and their implied average ticket

Source, publication date Period label Transactions Value Implied avg ticket
MFAT trade brief, June 2026 Calendar 2025 2.5 billion Rp 262.1 trillion Rp 104,840
Bank Indonesia RDG, December 2025 Calendar 2025 13.66 billion not published in this release not derivable
Bank Indonesia release, August 2026 H1 2026 12.55 billion Rp 1,120 trillion Rp 89,243

Source: Bank Indonesia 🔗 releases and December 2025 RDG figures as compiled in the payments press; MFAT trade brief via MFAT 🔗, June 2026

Taking the denominators apart

The ticket column does most of the audit work here, so it deserves its own paragraph.

An average ticket is value divided by count, and each pair of numbers implies one. MFAT’s pair implies Rp 104,840 per transaction. Bank Indonesia’s first-half 2026 pair implies Rp 89,243. These are close enough to be the same system with the same mix; a QR payment rail’s average ticket does not swing 80 percent in twelve months without an announced structural change, and no such change occurred. So the disagreement is not about the size of a transaction. It is about how many transactions there were, and the gap between 2.5 billion and 13.66 billion is a factor of 5.5.

Now test the vintage. QRIS volume by year, from Bank Indonesia’s historical reporting: roughly 2 billion transactions in 2022, about 3.5 billion in 2023, and by 2025 a run rate that had already passed 10 billion through September alone. MFAT’s 2.5 billion, paired with Rp 262.1 trillion, sits cleanly in the 2022 range. A figure that belongs to 2022, wearing a 2025 label in a document published in 2026, is a stale vintage, not a conflicting measurement. This is the audit’s first finding, and it is the mundane kind: a trade brief recycled an old row without updating it, and the label traveled further than the data.

The third check is the bank’s own projection table, because projections are where well-run institutions mislead themselves most honestly. In the governor’s annual meeting remarks published in late 2025, Bank Indonesia’s table projected QRIS volume of 15,374 million transactions and Rp 1,486.8 trillion in value for 2025. The actual outturn was 13.66 billion, about 11 percent under the projection. That is a normal miss, and it says something the raw growth numbers do not: even the operator that runs the rail, with real-time visibility, overestimated by a margin a casual reader would never notice between “13.66 billion” and “15.37 billion.” Projection and actual must never be merged into one series, and in the bank’s own documents they are marked separately. In derivative documents they frequently are not.

Table 2: Users and merchants, two official readings of the same quarters

Period Users (BI payment system report) Merchants MSME share
Q4 2024 55.43 million not in this row not in this row
Q4 2025 59.53 million 42.75 million not in this row
December 2025 RDG ~59 million ~42 million ~90 percent
June 2026 65.77 million 44.86 million 96.68 percent

Source: Bank Indonesia 🔗 payment system report 2025 and August 2026 release; December 2025 RDG figures as compiled by the payments press

Here the audit finds agreement rather than conflict. The RDG round numbers sit within rounding distance of the report’s precise ones, for the same periods. When two official products from one institution agree to within a rounding error, the vintage question is settled for those cells: they are the same year, honestly labeled.

Where the numbers cannot coexist

So far the audit has found one stale row. The deeper incompatibility sits in the fee schedule, and it is the kind that survives even perfect data.

From October 1, 2026, Bank Indonesia extended the zero merchant discount rate, MDR, to every merchant category on QRIS transactions up to Rp 100,000. Before that change, the zero band applied to micro merchants (UMI) up to Rp 500,000, and regular merchants paid 0.7 percent on transactions above Rp 100,000, with education at 0.6 percent, fuel retailers at 0.4 percent, and designated public-service categories at zero. These are the bank’s published category rates. The 2026 change did not abolish the schedule above Rp 100,000; it removed the charge inside the band for everyone.

Now put the average ticket next to the band. The first-half 2026 average ticket is Rp 89,243, inside the zero band. But an average inside the band does not mean most value is inside the band, and this is where the “QRIS is free” framing breaks. If the transaction distribution has a long right tail, which tourist and large-ticket behavior guarantees it does, then the count-weighted median can sit at Rp 60,000 while the value-weighted majority sits above Rp 100,000, where the 0.7 percent rate still applies. The inbound tourist data makes the tail concrete: in January 2025, foreign visitors made 24,121 inbound QRIS payments in Bali worth Rp 7.1 billion, an average of roughly Rp 294,000 per transaction, three times the national average and well inside the charged tier. A kiosk selling coffee and a hotel booking share the same QR standard, the same statistics, and opposite fee treatment.

The honest statement of the fee regime is therefore count-free, value-priced: by transaction count, the zero band covers a large share of Indonesian QRIS activity; by transaction value, the charged tiers still carry the bulk of the money. Neither half of that sentence is a contradiction. Announcing only the first half, as much of the coverage did, is the second audit finding: the denominator switched from transactions to value without an announcement.

The third incompatibility is smaller and worth naming because it recurs across the region’s payment statistics. The 2025 quarter-on-quarter growth figures in the payment system report, 139.99 percent for volume and 107.22 percent for value in Q4 2025, are quarter-over-quarter rates. Annualized naively, they would imply a system several times larger than the bank’s own annual count, which of course they do not describe. Quarterly growth and annual totals answer different questions, and the audit’s rule is to never let a quarter’s percentage borrow an annual denominator.

What the cross-checks establish

Run the findings together and the audit supports four statements with defined denominators.

  • The 13.66 billion figure is the system operator’s actual count for calendar 2025, cross-checked against its own mid-year run rate (10.31 billion through September 2025) and its user and merchant series, which match the RDG release to within rounding.
  • The 2.5 billion figure is internally coherent, carries a plausible 2022-era average ticket, and matches QRIS volume for calendar 2022; its 2025 label in the 2026 brief is a vintage error, not a measurement dispute.
  • The fee schedule is count-free on tickets up to Rp 100,000 from October 2026 and value-priced above it, at 0.7 percent for regular merchants, with the national average ticket (Rp 89,243, H1 2026) below the band and tourist tickets (Rp 294,000, Bali, January 2025) far above it.
  • The bank’s 2025 projection (15,374 million transactions) exceeded its actual by roughly 11 percent, a reminder that operator projections and operator actuals are different series even when both come from the same institution.

The fraud line belongs in this accounting too, because scale and loss move together. The Indonesian payment industry’s own anti-fraud body recorded Rp 4 trillion in social-engineering fraud losses for calendar 2024 (ASPI/IASC annual report, via Bank Indonesia). A system that grew from roughly 2 billion transactions in 2022 to 13.66 billion in 2025 grew its fraud surface at the same time, and the 2026 zero-fee band, by pushing more low-value commerce onto the rail, grows it further.

The other rail and the other rows

A QRIS audit that stops at the QR code misses the second half of Indonesia’s payment stack, and the second half carries its own pricing lesson.

BI-FAST, the central bank’s real-time interbank transfer rail launched in December 2021, handled 1.359 billion transactions in Q4 2025 alone, a 30.44 percent increase, and carried 82.6 percent of retail payment volume that quarter (Bank Indonesia payment system report, 2025). Its pricing is published in pieces a reader can actually check: the infrastructure charges participating banks Rp 19 per transaction for interbank credit transfers, and the fee a sending bank may charge its customer is capped at Rp 2,500 per transfer. Two rails, two pricing logics: QRIS prices the merchant by ticket size and category, BI-FAST prices the sender by a flat cap with a known wholesale cost underneath. An operator pricing Indonesia payments without separating the two will quote a “free” QR acceptance story against a capped transfer story and conclude the stack is cheaper than its own invoice shows. The denominator here is the transfer, not the ticket, and the cap is per transfer regardless of amount, which makes BI-FAST regressive in exactly the way QRIS’s schedule is progressive.

The account-ownership rows in the same trade brief deserve the same vintage treatment as the QRIS rows. MFAT’s brief carries adult bank account ownership of 60 percent for 2025. The World Bank’s Global Findex, the standard reference, measured 56 percent for 2024, up from roughly 52 percent in 2021. The three numbers form a plausible trend line rather than a contradiction, but they are three different instruments: a national promotional target row, a survey estimate with a defined sample and questionnaire, and an earlier survey vintage. Unlike the QRIS case, where the factor-of-5.5 gap ruled out honest trend reading, here a reader could merge the rows into one series and get an answer that is roughly right and methodologically wrong. The audit distinction is between numbers that cannot coexist (QRIS 2025) and numbers that coexist only because their definitions drift (ownership 2021 to 2025). Both belong in the same caution, for different reasons.

Cross-border expansion adds a third family of numbers with an unstated denominator. QRIS is now accepted in Malaysia, Thailand, Japan, China and Singapore, with UAE links under exploration, and the bank’s 2026 target is acceptance in eight countries (Deputy Governor Hendarta, January 2026). The January 2025 Bali figures, 24,121 foreign-visitor transactions worth Rp 7.1 billion, are the only published slice of what cross-border volume actually looks like at ground level: tiny in count, large in ticket, concentrated in one tourist island. A “five countries” acceptance claim and a “24,121 monthly transactions” usage figure describe the same program from opposite ends, and neither headline carries the other’s denominator. The new QRIS Tap feature shows the same pattern in miniature: launched in October 2025, it reached 508,000 transactions across 14 provinces within months, a growth rate of roughly 1,200 percent month-on-month that is arithmetic on a very small base. Percentage growth on small bases is the most quotable and least informative number in payments statistics.

What this section adds to the verdict is a rule about families of numbers. The QRIS vintage error was one bad row; the ownership drift, the cross-border count and the Tap growth rate are structurally recurring: every number that travels from an operational system into a promotional document tends to shed its denominator on the way. The fix is not distrust of the operator. It is treating every transferred number as a claim about a specific period, a specific unit and a specific base, and refusing to quote it without those three attachments.

Verdict

The two headline numbers are not a mathematical contradiction; they are two different datasets, one current and one roughly four years old, sharing a label. On the evidence of ticket consistency, series continuity and the bank’s own targets, the 13.66 billion reading stands as the description of 2025 and the 2.5 billion reading stands as a stale row that should carry a 2022 label or be dropped. The fee story, which survives the data dispute untouched, is a denominator switch: free by count inside the band, charged by value above it, with the average tourist transaction paying three times over. A reader who carries one sentence out of this audit should carry this one: before quoting a payments statistic, ask which year the number belongs to and which denominator the fee applies to, because in Indonesia’s QRIS both questions have answers that the headlines do not print.

Methodology

Data as of November 2026. This audit used only published figures and did no fieldwork; every number carries its source in the text or the tables. The core computation is the implied average ticket: transaction value divided by transaction count for each labeled period, computed without rounding the inputs (MFAT pair: Rp 262.1 trillion over 2.5 billion; Bank Indonesia H1 2026 pair: Rp 1,120 trillion over 12.55 billion). Vintage attribution compares the disputed pair against Bank Indonesia’s public annual volume series and its mid-year 2025 run rate, treating a factor-of-5.5 gap with consistent ticket sizes as evidence of a period mismatch rather than a definitional one. Time frame: figures labeled calendar 2025, Q4 2025, H1 2026 and January 2025, all as published between June 2026 and August 2026 plus the December 2025 RDG. The denominator definition for the fee analysis is explicit: count-weighted share inside the Rp 100,000 zero band versus value-weighted share, with Bali inbound transactions as the tail illustration. Limitations: the bank did not publish a full-year 2025 value figure in the December RDG release, so the 2025 implied ticket cannot be computed from the operator’s own pair; the transaction-size distribution is inferred from the tourist tail and the published average, not observed; and the MFAT brief’s underlying source chain for its payment rows is not disclosed in the document itself.

Sources

  • PaymentBrief 🔗, “Indonesia Payments: QRIS, Local Acquiring and BI-FAST,” 2026 (transposing Bank Indonesia December 2025 RDG full-year figures and the MDR schedule effective October 1, 2026)
  • Marketing-Interactive 🔗, “BI unveils cross-border expansion and merchant growth plans for QRIS,” January 2026 (quoting Deputy Governor Filianingsih Hendarta; transposing Bank Indonesia 2025 outturn and 2026 “17-8-45” targets)
  • Bank Indonesia 🔗, Governor’s Annual Meeting remarks and payment system projections, late 2025; payment system report figures for Q4 2024, Q4 2025; QRIS release of August 2026 for H1 2026 (site home)
  • MFAT 🔗, Indonesia trade brief, June 2026, payment system row labeled 2025 (site home)
  • ASPI/IASC annual report, social-engineering fraud losses, calendar 2024 (no public link)

About Global Infinity Labs

Global Infinity Labs is an AI innovation company specializing in MCP data services. This article was prepared by our expert team using insights and data provided through our MCP platforms.

Contact: support@globalinfinitylabs.com