Official EV Data Lab (2026): Original Calculations from Government and EU Statistics

The EV Data Lab is an evidence-based research index, not a generic EV-news feed. We examine official statistical releases and transport datasets, reproduce clearly defined calculations, and explain where often-repeated claims become misleading. Each original analysis links to the primary agency, names the time period, separates raw figures from our own arithmetic, and identifies which comparisons are not justified by the available data. Research set last reviewed November 13, 2026.

Ten primary-source EV and vehicle-data investigations

What these figures can and cannot be compared with

A new registration is not a vehicle in the fleet. Denmark has an 88.2% electric share in the month’s new passenger cars and a 24.9% BEV share in the existing fleet; subtracting those shares gives an interesting within-country descriptive gap, not a forecast. The Finnish 60.1% figure also describes a single month’s new registrations. Croatia’s 0.61% and Bosnia and Herzegovina’s 0.062% describe registered stock in 2025. Placing all these figures into an international league table would compare unlike measures.

Alternative fuels are not only battery-electric vehicles. European Alternative Fuels Observatory figures can contain plug-in hybrids, LPG, CNG, LNG and hydrogen, in addition to fully electric vehicles. Bulgaria’s 6.32% and Estonia’s 4.51% in this collection are all-alternative-fuel shares, not BEV ownership shares. Separate the fuel category and vehicle class before calculating any BEV ratio.

Adjusted totals are not interchangeable with actual registrations. Denmark’s official release includes a seasonally adjusted monthly total alongside the actual registrations by fuel. We used the actual total when calculating the actual BEV share. France’s SDES release likewise publishes adjusted counts in one statistical series and fuel proportions in another presentation; we do not multiply these together to create a potentially invalid estimate.

Ownership registration is not user identity. In the Netherlands CBS source, 54.6% means that 377,001 of a stated 690,922 fully electric cars were registered to legal entities in a specific dataset. It does not mean 54.6% of EV driving kilometres were driven for business or that 54.6% of newly sold BEVs were company purchases.

Source-by-source calculations and reproducible boundaries

1. Netherlands: Registered BEV stock by owner type

Published insight: 54.6% legal-entity registrations. Source period: 2026* CBS StatLine selection. Method: 377,001 / 690,922 × 100. Important boundary: Ownership in a specified StatLine population; do not infer new EV sales. Read the detailed local-language analysis or open the primary data source.

2. France: New-car registrations: BEV + PHEV

Published insight: 47.7% plug-in share, 40.7% BEV alone. Source period: October 2026. Method: 40.7 + 7.0 percentage points. Important boundary: BEV+PHEV is not BEV-only; avoid mixing unadjusted shares with adjusted totals. Read the detailed local-language analysis or open the primary data source.

3. Bosnia and Herzegovina: Registered pure-electric passenger cars

Published insight: 0.062% of passenger cars. Source period: 2025 annual stock. Method: 702 / 1,138,056 × 100. Important boundary: 40,339 alternative-fuel cars include LPG and hybrids; not all are BEVs. Read the detailed local-language analysis or open the primary data source.

4. Croatia: Registered pure-electric passenger-car fleet

Published insight: 0.61% BEV and 2.07% hybrid. Source period: 2025 stock in 2026 MUP guide. Method: 12,618 / 2,058,313 × 100. Important boundary: Registered stock, not 2025 sales; hybrid definition is separate. Read the detailed local-language analysis or open the primary data source.

5. Serbia: Vehicle age among first registrations

Published insight: 82.9% to 74.8%, change −8.1 pp. Source period: Q1 versus Q2 2026. Method: 34,312 / 41,391 versus 37,441 / 50,054. Important boundary: All first-time-registered cars, NOT an EV-only measure. Read the detailed local-language analysis or open the primary data source.

6. Bulgaria: All alternative-fuel light vehicles

Published insight: 6.32% AF share in 2025; 222,040 AF vehicles in Q2 2026. Source period: 2025 share, Q2 2026 stock. Method: Compare periods only after aligning them. Important boundary: Alternative fuel includes LPG/CNG/LNG/BEV/PHEV/H2; Q2 count is not EV count. Read the detailed local-language analysis or open the primary data source.

7. Czechia: BEV share of new car registrations

Published insight: 7.6% versus 5.6%, +2.0 pp. Source period: January–October 2026 vs 2025. Method: 7.6 − 5.6 percentage points. Important boundary: +59% imports of used BEVs is growth, not a 59% import share. Read the detailed local-language analysis or open the primary data source.

8. Denmark: New registrations versus existing stock

Published insight: 88.2% of new cars; 24.9% of fleet. Source period: October 2026. Method: 16,673 / 18,901 × 100 vs fleet share 24.9. Important boundary: Use actual 18,901, not seasonally-adjusted 19,321 as new-reg denominator. Read the detailed local-language analysis or open the primary data source.

9. Finland: New car BEV share and BEV/petrol ratio

Published insight: 60.1% BEV; 2.24 BEVs per petrol. Source period: October 2026. Method: 3,417 / 5,685 × 100; 3,417 / 1,527. Important boundary: Monthly new registrations are not whole-fleet figures. Read the detailed local-language analysis or open the primary data source.

10. Estonia: Alternative-fuel vs national vehicle register

Published insight: 4.51% alternative-fuel share, not BEV share. Source period: EAFO 2025/2026-Q2; Statistics Estonia 2025. Method: No cross-population division performed. Important boundary: EAFO M1+N1 metric and the national traffic-register stock have different definitions. Read the detailed local-language analysis or open the primary data source.

Editorial standards for using official EV statistics

1. Start at the public dataset. Confirm that the agency is responsible for the relevant geographical area and statistic; search-engine summaries alone are insufficient.
2. Keep a dated snapshot. Record publication date, reference period, population, raw number, and whether the result is preliminary or revised.
3. Choose the right denominator. Match vehicle class, fuel definition, geographic region, time period, registration status and statistical method.
4. Compute before rounding. Ratios use the published counts. Percentage-point changes are different from percentage growth rates.
5. State an exclusion. List at least one important conclusion the original data do not support, such as an EV-only statistic derived from all first registrations or all alternative fuels.
6. Link the original record. Every investigation carries a primary source link, and corrections will identify substantial changes if newer official figures supersede older releases.
7. No fabricated novelty. The underlying numbers are public data. The independent contribution is a documented calculation, combination or methodological insight; we cannot claim no one else on the internet has ever performed the same arithmetic.

How often are these figures updated?

There is no uniform monthly update cadence across countries. Each publisher has its own statistical schedule. A new release does not retroactively change the historical source date shown in an older investigation; it can, however, warrant a documented correction or a separate updated analysis. We do not infer unseen years or manufacture forecasts from single-month samples.

How this relates to our charging and ownership guides

Statistics cannot tell an individual driver whether a particular charger works now or which tax notice applies to their vehicle. For practical, country-specific ownership, charging and regulatory instructions, use our electric vehicle guides by country. For research and correction practices see the editorial guidelines. Both collections intentionally contain text and tables without featured images.

About the work: Research and drafting were assisted by AI tools, with cited public-agency data and explicit arithmetic as the evidence trail. This is not a claim of field testing, personal interviews, native-speaker legal certification or independent authority endorsement. Language-level editorial review and a check against any later revisions remain important.