Methodology

Last reviewed: 20 April 2026

Data sources

Every piece of COE data published on this site originates from the Land Transport Authority (LTA) of Singapore. We do not use proprietary data feeds, private dealer sources, or paid subscription services. Primary sources:

  • LTA OneMotoring (onemotoring.com.sg) — the authoritative portal for quota premiums, bidding schedules, quota announcements, and PQP.
  • LTA deregistrations dataset — monthly deregistration counts by category, published via OneMotoring.
  • LTA quota announcements — quarterly quota allocations, published approximately one month before each quarter begins.
  • LTA road tax, ARF, and PARF schedules — formula tiers published on the OneMotoring rates page.

Running-cost inputs such as fuel, insurance and parking are editable assumptions. They are not live supplier quotes. Check the source and observation date on any cited article before using a historical example.

A known difference between official copies

For the second January 2010 exercise, the data.gov.sg copy of LTA's COE bidding results lists Category D (motorcycles) at S$20,090, the same figure as Category C in that exercise. The LTA DataMall file that COEkaki imports lists S$852, in line with the S$889 Category D premium of the first January 2010 exercise. COEkaki shows the DataMall figure. The data.gov.sg record was checked on 17 October 2026.

Update cadence

Bidding results: the published LTA bidding calendar determines exercise dates. COEkaki schedules imports at 14:30 and 16:30 SGT on result days. An import schedule is not a guarantee of successful retrieval or a live bid feed; check the displayed exercise date and source status.

Deregistration data: monthly imports are scheduled. Displayed observation periods and retrieval dates indicate the data actually available.

Vehicle population and registrations: these datasets have their own publication and retrieval dates. They need not be as recent as the latest COE exercise.

PQP: estimates are recalculated when new results arrive. The renewal month, rather than the latest single exercise, sets the calculation window.

How PQP is computed

The Prevailing Quota Premium (PQP) varies monthly. For the selected applicable month, the calculation uses all exercises in the three preceding months in which bidding took place; months without bidding are skipped. It is not a fixed last-six-exercise calculation. See LTA COE renewal guidance.

PQP(category, applicable month) = sum of premiums / number of exercises
in the three preceding bidding months

The site rounds its calculated estimate to whole dollars and shows unavailable when fewer than three bidding months exist. Confirm the official applicable PQP, renewal timing and vehicle eligibility with LTA before paying; the site estimate is not an official payable quotation.

Historical data

COEkaki's archive extends back to the April 1999 bidding exercise, the point at which the current Category A–E scheme began. Everything from that exercise onward comes from LTA DataMall: results up to 2009 from the historical COE Bidding Results 1990–2009 dataset, and 2010 onward from the live COE Bidding Results dataset, refreshed after every exercise. We publish no reconstructed or estimated premiums.

Between June 2001 and March 2002 the Closed and Open bidding systems ran in parallel, so those months carry two exercises; we record them as rounds 1 and 2 respectively. Before April 1999 a seven-category scheme applied (by engine capacity bands rather than today's A–E), and LTA has never restated those premiums under the current categories — so we do not show them rather than approximate a mapping. For pre-1999 context in guides we cite government and academic publications with inline footnotes.

Where LTA's own data has been revised post-publication (rare but it happens), we use the most recent revised figure and note the revision date.

Calculator formulas

Calculator pages describe their assumptions and results. All calculators share the same tax and rate rules, each dated by when it takes effect, so a correction applies to every tool at once.

  • ARF — applies the marginal OMV bands for the selected policy period and shows the bands it uses.
  • PARF rebate — uses age and the applicable vehicle policy cohort. Percentages and caps vary by cohort; select the correct period and use ARF actually paid. Refer to LTA vehicle tax structure.
  • Road tax — LTA's tiered formula by engine capacity (petrol) or motor power (EV).
  • Loan — calculates instalments and interest from principal, flat annual rate and term. The repayment estimate is not a loan approval.
  • Total cost of ownership — combines modeled upfront registration costs, including applicable taxes, with the selected years of running costs. Editable inputs are estimates; read the stated exclusions before comparing them with a dealer price.
  • Renew vs Scrap — four-way comparison: renew 5y, renew 10y, scrap + new, carless; scrap value = PARF + COE rebate + body shell.
  • Quota Outlook — holds each category’s latest complete quota flat for the next two periods. Its range comes from how much quotas have changed over the same span in the past; it is not a statistical confidence interval.
  • Weighted Premium Index — a fixed weighted basket of category premiums, normalized against January 2015 round one. The index page displays its weights, calculation and missing-data policy. It is not a whole-car price index.

Use the calculation explanation and data table on each tool to interpret its output. Formulas, policy tables and observation data are versioned separately; unchanged inputs can produce different results when the selected policy period changes.

Known limitations and error bars

We try to be explicit about where the numbers stop being reliable:

  • Forecasts are scenarios, not predictions. The Quota Outlook is a transparent supply model, but actual COE premiums depend on both supply and demand — and demand is harder to forecast than supply.
  • Depreciation is modeled. The visualizer provides illustrative scenarios rather than a resale valuation or guaranteed future price.
  • Flat and effective interest rates differ. The loan tool accepts a flat annual rate. Do not substitute an effective rate or use a universal conversion multiplier; fees and repayment timing affect the comparison.
  • Policy changes require reviewed updates. Each rule is dated by when it takes effect, so each calculation uses the rule currently in force. Policies not yet announced are not modelled.

Version history

  • 2026-04-20 — Initial published methodology. Calculator formulas, data sources, and cadence documented.

13 September 2026 — partial methodology correction: updated the PQP window, policy-cohort and loan descriptions, scenario terminology, and import-status limitations to match the implemented tools. Original publication and unrelated historical statements are preserved.

Keyboard shortcuts

Search
/
Show shortcuts
?
Home / Results / Trends / Calculators
G then H / R / T / C

Quick tour

The latest results show COE prices for all five categories. Open a result to see the full bidding details.

Results Archive

Use the trend charts to explore price history, and the calculators to estimate your vehicle costs.

Trends

Welcome back!