Reporting thresholds — four designs, four consequences
“Are we above the threshold?” is a question with four different meanings depending on which return you are asking about, and the four are not interchangeable. Some reporting thresholds decide whether you report at all. Some decide how often. Some decide whether what you report is shared onward. And some are not thresholds but derogations a supervisor grants. Confusing them reliably produces over- or under-reporting across a multi-market footprint. Below: the four designs, the measurement bases underneath them, three worked cases and the controls that keep threshold logic defensible.
1. The four designs of statistical reporting thresholds
Every threshold in a statistical or credit-register return fits one of four boxes, and the box tells you what your systems must still do on the “small” side of it.
| Design | Decides | Worked example |
|---|---|---|
| Scope threshold | Whether an item is reportable at all | German external-sector reporting exempts payments up to EUR 12,500; below it, the item is simply not reportable |
| Frequency band | How often you file — you are always in scope | The Spanish external-transactions survey (ETE): monthly at EUR 300m, quarterly at EUR 100m, annual below, and on request only where the amount does not exceed EUR 1m |
| Disclosure threshold | Whether data you already report is shared onward | The Spanish credit register (CIR): the consolidated report returned to declaring entities covers holders at or above EUR 3,000 of accumulated risk, stepping down to EUR 1,000 from 2 January 2027 |
| Derogation | Whether the supervisor releases you from part of the obligation | The ECB payment statistics regime, where national central banks may grant derogations guided by proportionality |
Porting a scope-threshold mental model into a frequency-band regime produces a population that is simply missing — and nothing in the file itself will tell the reviewer that it is.
2. What the threshold is measured on
Just as important as the design is the measurement basis, and three variants recur:
- Per item. The German EUR 12,500 exemption applies to the individual payment, not to a counterparty total.
- Aggregated at a defined level. The AnaCredit threshold under Article 5 of Regulation (EU) 2016/867 is a debtor’s commitment amount equal to or larger than EUR 25,000, calculated as the sum of commitment amounts for all the debtor’s in-scope instruments in relation to the observed agent — not group-wide and not per instrument. Aggregating at the wrong level changes the population, not a field.
- Across a period, not at a point. The same article tests the threshold on any reporting reference date within the reference period, so an exposure that peaked mid-period and fell away is still in scope. Testing only at period end produces a short population.
A fourth variable is the look-back: the ETE bands use the preceding year’s transactions or prior 31 December balances, and the payment statistics derogation test uses the 12 months before the assessment.
3. The comparison operator is part of the rule
“Above EUR 1,000” and “EUR 1,000 or more” are different rules. In the CIR, the declaration rule in Orden ECO/697/2004 refers to accumulated risk above EUR 1,000, while the return-report rule, once stepped down, refers to risk equal to or above it — so a holder at exactly EUR 1,000 falls on different sides of the two lines. AnaCredit also says “equal to or larger than”. Store the operator with the amount, and write one test case per threshold that sits exactly on the boundary.
4. Derogations are asked for, not assumed
Regulation (EU) No 1409/2013 on payment statistics, as amended by Regulation (EU) 2020/2011, is the clearest illustration of how a derogation differs from a threshold. Under Article 4, national central banks are guided by the principle of proportionality and may grant derogations to:
- payment institutions meeting the conditions in Article 32(1) and (2) of PSD2;
- e-money institutions meeting those in Article 9(1) and (2) of Directive 2009/110/EC;
- other payment service providers meeting both sets of conditions.
They may do so regardless of whether the firm has actually been waived or exempted from prudential requirements nationally. The test is the conditions, not the registration status.
Availability then depends on a national-materiality test in Article 4(3): either the total value contributed by all providers that could benefit does not exceed 5% at national level for each listed payment service (and cumulatively across them), or, for account information services, the total number of clients of those providers does not exceed 5% nationally. The derogation is only available where the reporting burden would be disproportionate given the agent’s size. Three further mechanics matter in practice:
- A derogated agent does not stop reporting — it reports the reduced Tables 4b and 5b of Annex III.
- The national central bank re-verifies the conditions annually, based on the 12 months before the assessment, and grants or withdraws with effect from the start of the second successive calendar year.
- The ECB publishes a list of entities granted derogations, so your status is checkable.
5. Worked case: a frequency band hidden in a group structure
Facts: a fintech group has a licensed e-money institution registered with the Banco de España and a separate Spanish holding company. The holding company has almost no cross-border payments but carries a EUR 140 million loan from its foreign parent.
What the rule says: the ETE under Banco de España Circular 4/2012 excludes payment service providers registered in the Banco de España’s official registers, which report cross-border flows through the separate declaration regime for payment service providers. The holding company is not a PSP, so it is in the ETE, and the band is set by the higher of transactions or balances. On the balance limb, EUR 140 million sits in the EUR 100–300 million band.
What the practitioner does: treats the holding company as a quarterly filer, sources the population from the general ledger rather than the payments system, and names an owner, because teams organised around the licensed entity leave this one unclaimed.
Outcome: the threshold never removed the holding company from scope; it only set the cadence, driven by a balance nobody was watching.
6. Worked case: an exposure that peaked mid-period
Facts: a credit institution reporting to AnaCredit grants a corporate debtor a EUR 15,000 overdraft and a EUR 12,000 term loan, both booked at the same observed agent. Both are in force on an earlier reporting reference date in the reference period; the overdraft is then repaid and cancelled before the last one.
What the rule says: the commitment amounts are summed across the debtor’s instruments at that observed agent — EUR 27,000 — and the threshold is met if reached on any reporting reference date within the reference period, not only the last one.
What the practitioner does: computes the threshold on every reporting reference date in the period and flags the debtor as in scope, rather than running a single last-date snapshot that would see only EUR 12,000. If the same debtor also borrows from a branch that is a different observed agent, that exposure is tested separately.
Outcome: the debtor is reportable. A period-end-only filter would have produced a file that validates cleanly and is still short.
7. Worked case: a small e-money institution and the payment statistics derogation
Facts: a small e-money institution meets the conditions of Article 9(1) and (2) of the e-money directive but is fully authorised rather than waived. It has been filing the full payment statistics tables.
What the rule says: the derogation turns on the conditions, not the waiver, so the institution is eligible in principle. Whether a derogation is actually available depends on the 5% national test across all providers that could benefit, and on the national central bank’s assessment of proportionality.
What the practitioner does: writes to the national central bank asking whether derogations are granted in that market and on what basis, rather than unilaterally switching to the reduced tables. Meanwhile it keeps filing the full return and tracks volumes against the conditions.
Outcome: a granted derogation, or a documented reason the full return continues. Both are defensible; self-help is not.
8. Building threshold logic that holds up
The recurring mistake is one module that filters out items below a value before they enter the reporting population. It is right for a per-item scope threshold and wrong almost everywhere else. The fix is to model the threshold as an attribute of the obligation rather than a filter in the pipeline, recorded per return in a threshold register:
| Field | Example values |
|---|---|
| Design | Scope / frequency band / disclosure / derogation |
| Amount and operator | EUR 12,500, “up to”; EUR 25,000, “equal to or larger than” |
| Measurement basis | Per item / aggregated per debtor per observed agent / per entity |
| Test timing | Point in time / any date in period / prior-year look-back |
| Source of the rule | Instrument and article, with the date last checked |
| Review trigger | Annual band reset, dated step-down, derogation re-verification |
Those fields are enough to implement every variant above. Add a dated change log, because thresholds move: the CIR return threshold has a fixed step-down in January 2027, and derogations are re-assessed every year.
FAQ
Does being below a reporting threshold always mean no reporting?
No. Only a scope threshold removes the item. Frequency bands change cadence, disclosure thresholds change onward sharing, and derogations must be granted and reduce rather than remove the return.
At what level is the AnaCredit threshold measured?
On the debtor’s commitment amount summed across all in-scope instruments in relation to the observed agent, and it is met if EUR 25,000 or more is reached on any reporting reference date within the reference period.
Can a firm claim a payment statistics derogation itself?
No. National central banks grant them, guided by proportionality, and availability depends on the 5% national-materiality test. The right step is to ask the national central bank what its policy is.
Does a derogated firm stop filing payment statistics?
No. It reports the reduced Tables 4b and 5b of Annex III to Regulation (EU) No 1409/2013, and the national central bank re-verifies eligibility every year.
Does a Spanish payment or e-money institution file the ETE?
Generally not: the ETE excludes PSPs registered with the Banco de España, which report cross-border flows under the separate PSP declaration regime. Non-PSP entities in the same group remain in scope.
Which CIR threshold changes in 2027?
The threshold for the consolidated report returned to declaring entities, which moves from EUR 3,000 to EUR 1,000 on 2 January 2027. The declaration threshold is a separate rule.
What to do, today
- Head of regulatory reporting: classify every threshold you apply into one of the four designs and record its operator, measurement basis and test timing.
- Data engineering: replace pre-population filters with threshold logic evaluated per obligation, and add a boundary test case for each threshold.
- Group finance: list every non-licensed entity in each market and test it against the local external-sector bands on both the transaction and balance limbs.
- Compliance or COO: where your firm may meet the small-institution conditions, ask the national central bank in writing whether payment statistics derogations are granted.
Related: ECB payment statistics · The Spanish ETE bands · German AWV reporting · The CIR threshold step-down · External-sector reporting compared · The DTE for PSPs


