Home/ Online Tools/ Track Record

Our Rules Come From Data, Not From Feeling

Many forwarders say “we prepare documents carefully.” We’d rather show the work: our inspection-mitigation and tax-adjustment rules were derived line by line from parsed CEISA declaration records and full-text analysis of official SPTNP tax adjustment notices. This page lays out the data sources, sample sizes and measured findings — judge for yourself.

Traceable data sources Anonymised aggregate statistics No client-identifying data

How Much Real Paperwork These Rules Are Built On

Two sources: declaration line items from Indonesia’s CEISA 4.0 customs portal (read-only extraction), and the official text of SPTNP tax adjustment notices, parsed one by one.

535declarations

Declaration records parsed individually (lodged from 2024 onward), after expansion

Initial sample: 140 declarations

5,807lines

Declaration lines covering 771 Indonesian HS8 tariff lines

Initial sample: 1,320 lines / 83 tariff lines

73notices

SPTNP tax adjustment notices, all parsed in full, with tax lines traced back to declarations

First-hand record of enforcement behaviour

Why expand the sample: a few dozen declarations produce wrong rules. Switching the extraction dimension from “broker view” to “importer dimension” multiplied the sample by 3.8× (declarations), 4.4× (lines) and 9.3× (tariff lines) — and immediately overturned several small-sample conclusions. See section 3.

Six Findings That Came Out of the Data

Every figure below traces back to specific records. No “generally speaking” — only “what this batch actually showed”.

66 / 66

Declarations containing a declared-value error: every single one was fined — no exceptions

41 of them fell into one fixed statutory bracket (a fixed statutory amount, not proportional)

0 / 7

Declarations with a rate error only and correct value: zero fines

Fines target value, not rate — this dividing line sets where our defences focus

+17.3%

Median uplift applied by customs to the customs value (66 uplifted declarations)

Measured in this batch, not a statutory rate; of the 45 pure value-error cases, 42 saw no rate change at all

1.88×

Red-lane (physical inspection) declarations show a tax adjustment rate 1.88× that of green-lane ones

Red 75.5% (40/53) vs green 40.2% (33/82)

83.5%

Share of declaration lines filed at an explicit 0% duty rate (1,102 / 1,320)

Most shipments ride on a preferential agreement — if it fails, there is zero tolerance

63.0%

Ten declarations where 0% was revised to a positive rate account for 63.0% of all duty adjustments

Adjustments are not spread evenly — they cluster on a few points

The One-Line Rule: Tax Adjustment and Fines Are Two Different Problems

Lumping “adjustment” and “fine” together puts your defences in the wrong place. The data is unambiguous.

Fines: only for a wrong declared value

All 66 declarations with a value error were fined — not one exception. The 7 with only a rate error and a correct value were fined zero. A clean dividing line from the same batch.

So we pulled “value-declaration defence” out as its own block: supporting evidence on file, mandatory explanation for deviations from historical prices, customs-value build-up review, and a three-tax cross-check — four gates all on value.

Adjustments: mostly because the preference does not hold

83.5% of lines were filed at an explicit 0% duty — riding on preferential agreements such as FORM E / ACFTA. The biggest driver of adjustment is neither a classification error nor an unfilled field, but the preference being disallowed and duty collected at MFN rates.

So the review gate centres on “four-way consistency for preference eligibility”: agreement code, certificate number, origin and rate must all line up — plus field-by-field FORM E checking and pre-clearance zeroing.

One counter-intuitive point: “declaring low saves money” is wrong. Under-declaring does not just cost duty — it drags VAT and income tax along with it. Those two alone are 57.5% of all adjustment amounts.

Two Pieces of Evidence Worth Your Time

① Inspected shipments lose their preferential rate far more often

Red lane (inspection)

75.5%

Green lane (release)

40.2%

Red 40/53, green 33/82 — a 1.88× gap. The reason is simple: once a container is opened, customs checks documents against the goods item by item, and preference eligibility, amounts or marks that would not otherwise be examined get rejected on the spot. In other words, not being inspected has value in itself — which is why we put document review first.

② Where the adjustment money actually lands

VAT (PPN)

46.4%

Import duty (BM)

23.1%

Fine (Denda)

19.4%

Income tax (PPh)

11.1%

Duty is under a quarter of it — the bulk is the VAT and income tax recomputed alongside it. Watching only the duty rate means missing the most expensive part.

⚠️ The flip side: historical declared prices cannot serve as a safety line. We tested it — five lines that customs revalued in one declaration were all above both historical safety thresholds for that category; not one was caught. And the reference value customs actually applied was 51% higher than our own historical P90. The only usable interception line is a value customs has already determined, not one we filed ourselves.

The Findings Became Clauses, Not Just Documentation

Findings that are not written as “stop and fix when X” will be missed again. So they are fixed as numbered clauses, each with four parts: trigger, action, evidence, release decision.

Stage	Clauses	What it governs
Filing side	R1 – R13	Self-checks during PIB entry: value-declaration defence, three FORM E checks, document-completeness gate, single rate per HS8, complexity gate, zero tolerance on 0% duty
Review side	S1 – S34	Pre-filing review: three-document consistency, naming and specification wording, marks, twelve-field FORM E check, complexity escalation gate, high-risk category points, interception at customs-determined values
Remedy side	Valuation response / SPTNP handling	Deadline countdown and appearance discipline after a customs valuation notice; cause analysis and reduction paths after an SPTNP

How clauses get revised: when new evidence overturns a clause, the evidence field changes before the action does. For example, our original “freeze when intra-declaration price spread for the same HS8 ≥1.5×” rule was killed by the control data — declarations that were not revalued showed wider spreads (median 2.50 vs 1.69). A hard freeze would generate heavy false alarms, so it was downgraded to “flag plus mandatory line-by-line explanation”, with a separate interception line anchored on customs-determined values.

What This Means For You

You don't need to know the rules

The rules run on our side; you just hand over the cargo

Documents reviewed before filing

If we find a “must fix”, we stop and confirm with you — never file it broken

Values backed by evidence

Deviations from historical prices require an explanation — we don't casually lowball

A remedy path if it goes wrong

If a valuation notice or SPTNP does arrive, the deadline plan and response path already exist

Send us your documents for review Check the code and permits first

Basis of Figures (read alongside the numbers)

All figures are anonymised aggregate statistics drawn from declarations and tax adjustment notices we handled. No client names, tax IDs, registration numbers or document numbers are included. The statistical basis and sample period are traceable in our internal working papers.
The “median +17.3% customs value uplift” is a measured value from this batch, not a statutory rate applied by Indonesian customs, and cannot be used to predict the outcome of any specific shipment.
The fixed fine bracket mentioned (Rp 5,000,000) is a statutory amount under UU 10/1995 Pasal 114 ayat (1), not a measured amount in this batch. What was measured is the proportion: 41 of 73 notices fell into that bracket.
Any determination involving duty rates or import restrictions (Lartas) is always based on a same-day check on Indonesia’s INSW system. The figures here are historical patterns, not predictions for any individual shipment.
This page discloses methodology and historical data. It is not a commitment or guarantee regarding inspection rates, adjustment amounts or clearance outcomes. Actual handling follows the current rules and practice of Indonesian customs (CEISA / INSW).