There is no shortage of writing about the G2V2 smart tachograph, ours included. The same goes for the Mobility Package, ETS2, CountEmissionsEU, waste-transport tracking and urban access restrictions. On any single obligation, everything you need is out there.
What is missing is the sum. What does staying compliant cost, per vehicle per year, between 2026 and 2028? And more to the point: what is that cost made of?
We built a bottom-up model to find out. The headline is not one number but four: €186 to €1,593 per vehicle per year, depending on which regime the fleet falls under. The gap between the lightest and the heaviest profile is 8.6 times, and it has nothing to do with how well the fleet is run. It depends on where it goes.
There is a second result, less obvious and more useful. Look at a steady-state year, one with no hardware to buy, and 97% of the cost is people’s time.
A note on scope before the numbers: this is an Italian perimeter. The EU regulations apply across the single market, but the labour costs, the workshop prices and the urban access rules are Italian, so read the levels as Italian and the structure as general.
Why four profiles instead of an average
An average compliance cost would be clean to read and impossible to use, because it would blend companies subject to entirely different obligations. Worse, plenty of operators are not certain which regime they are in, which is exactly why our map of the ten regulatory regimes starts with that question rather than ending with it.
So we modelled four profiles, each at 30 vehicles to keep the comparisons readable. Each one isolates a single variable:
- A. Vans under 3.5 t, own-account, domestic only. The profile that assumes it is exempt from nearly everything. It is the baseline.
- B. Vans 2.5–3.5 t, hire-and-reward, international. Same vehicle, but inside the scope of Regulation 561 and the tachograph mandate from 1 July 2026.
- C. Mixed fleet, heavy vehicles plus vans. Two regimes running in parallel, with the heavy vehicles already fitted with tachographs.
- D. Domestic vans delivering into historic city centres. Profile A plus the urban access layer.
What is in and what is out
In scope: only the incremental costs introduced or changed by the 2026–2028 regulatory wave. G2V2 tachograph hardware, installation, cards, two-yearly recalibration and data handling; Mobility Package obligations, meaning posting declarations, cabotage and driver return; ETS2 from 2028; urban access permits.
Out of scope: road tax, insurance, roadworthiness testing, driver CPC and routine maintenance. Those are running costs, not compliance costs. Folding them in would inflate the total and make the comparison between profiles useless. If you want full cost of ownership instead, the method is in our piece on fleet TCO.
One case deserves its own note. CountEmissionsEU is not in the base cost, and not by oversight. Regulation (EU) 2026/1030 is voluntary in the choice and binding in the method: it obliges anyone who decides to report transport emissions to do it a particular way, it does not oblige them to report. It becomes a cost when a customer asks for it, and at that point it is a commercial cost before it is a regulatory one.
The numbers
Every cell is the median of 20,000 Monte Carlo draws on a 30-vehicle fleet.
| Profile | Three years (fleet) | Cash | Recurring | Time | € per vehicle/year |
|---|---|---|---|---|---|
| A domestic own-account | €16,699 | 0% | 100% | 0% | 186 |
| B international hire-and-reward | €143,000 | 38% | 17% | 45% | 1,593 |
| C mixed fleet | €121,000 | 29% | 18% | 53% | 1,350 |
| D domestic with city-centre delivery | €37,300 | 0% | 95% | 5% | 418 |
For profile B the uncertainty band runs from €1,387 to €1,826 per vehicle per year between the tenth and ninetieth percentiles. That is narrow next to the jump between profiles, which is why the A-to-B gap is the result we stand behind most firmly.
The profile that barely pays, and why that is the story
Profile A, vans under 3.5 tonnes running own-account without crossing a border, carries an incremental compliance cost of zero in 2026 and zero in 2027. The entire regulatory wave the industry has been talking about for two years does not touch it.
Then 2028 arrives, and so does ETS2, which does not care who you are because it prices the fuel rather than the operator: €557 per vehicle in 2028 alone, which spread across three years gives the €186 in the table.
That figure is worth converting into something you can check at the pump. Using the diesel emission factor derived from the regulatory coefficients (2.613–2.692 kg of CO2 per litre) and an allowance price between €45 and €90 per tonne, ETS2 is worth 12 to 24 cents per litre if the allowance cost passes through to the price in full. For the €557, the model passes through between 85% and 100% of it, which gives 10 to 24 cents. Anyone budgeting 2028 fuel off a historical series is under-forecasting by that much. The same calculation appears in the 2025 report of Italy’s Freight Insights Observatory (MOST and the CSELI Foundation, published by the Conftrasporto trade confederation): with 2.68 kg of CO2 per litre, €0.121 per litre at €45 per tonne and €0.268 at €100.
One clarification, because comparing this against other published estimates is easy to get wrong. Those cents are the carbon component alone, the cost of the allowance passed into the price. Trade-association estimates of 30 to 32 cents per litre, reported in our piece on ETS2 and road haulage, also include the handling margins fuel suppliers apply downstream. They are two different quantities and they do not contradict each other: the first is what the CO2 costs, the second is what reaches the pump.
A higher figure for the same vehicle class circulates in Italy. The CER-Confcommercio report on the European emissions market, presented in Rome on 19 May 2026, puts ETS2 alone at roughly €660 per light commercial vehicle per year at average 2025 carbon prices, rising to about €960 in its €107.5 per tonne scenario. For heavy vehicles above 26 tonnes the same estimates exceed €7,800 and €11,300, and for road transport as a whole they reach €4.7–11.3 billion a year from 2028.
The two figures do not contradict each other, because almost all of the gap is the assumed allowance price. That report sets out two scenarios, €45 per tonne as the optimistic one and €107.5 as the pessimistic one for 2028, and the number in circulation is the second; our range starts from the same lower bound and stops at €90, the assumption we flag as the most contestable in the model. The report’s methodology is not published: the full PDF is not linked from the Confcommercio page, and mileage, fuel consumption and emission factor appear in none of the sources that carry it, so the figure cannot be reconstructed.
A result that lets part of the market off the hook is also what makes the part that condemns the rest believable: taking the same vans across a border multiplies the cost by 8.6.
The steady-state year: the 97% nobody budgets for
Across the three years, time accounts for 45% to 53% of the cost in the tachograph profiles. More than the hardware, but that is not the interesting figure, because the three-year window contains both the installation year and the year ETS2 switches on. Two exceptional events.
The interesting year is 2027: no hardware to buy, ETS2 not yet live. An ordinary year. For profile B it costs €745 per vehicle, and 97% of that is people’s time. For profile C, 98%.
The difference between those two lines is not only size, it is nature. Hardware is a cash event. An invoice arrives, it is visible, it goes in the budget, it gets depreciated, and whoever signs it knows exactly what was spent. Time produces no invoice at all. Nobody receives a €745-per-vehicle note: they have a manager who spends a few hours a week pulling cards, filing data, re-checking infringements and completing declarations. That cost exists in full and appears on no line of the accounts under its own name.
It is also why the industry conversation has fixed on the price of installation. It is the only visible part.
Which lever to pull first
Sensitivity analysis answers a question no benchmark can: not what the total is, but where it is worth intervening. We put concrete actions on the same scale, measuring the effect on profile B’s cost per vehicle per year.
| Action | Effect |
|---|---|
| Ten minutes less on every data download | -5.91% |
| €300 cheaper installation per vehicle | -5.00% |
| Two minutes less paperwork per operation | -4.52% |
| ETS2 allowance price €20/t lower | -3.50% |
| Five minutes less infringement review per driver per month | -2.13% |
Ten minutes saved on each data download is worth more than three hundred euros negotiated off the installation price, and more than a twenty-euro collapse in the carbon price. The first depends on how the work is organised. The other two depend on a supplier and on a market.
There is a structural reason. The installation is negotiated once. The minute on the data download repeats thirteen times a year for every driver and four times for every vehicle, indefinitely.
Incentives cover the wrong part
Italy offers several schemes that offset part of this spend over the same period: the Nuova Sabatini subsidised-finance scheme on vehicles and on software, enhanced capital allowances, regional grants, and the de minimis rules that govern stacking them.
They share one characteristic: they act on capital expenditure. In this model capital expenditure is the “cash” line, which is 38% of the three-year total for profile B and 0% for profiles A and D. The theoretical ceiling on what any incentive can offset is that share, and the share that weighs most has no incentive attached to it at all, because it is people’s time.
This is not an argument that incentives are useless. It is an argument that they solve the part you pay once and can see, and leave untouched the part you pay every year and cannot. Where the dominant line is organisational, the answer has to be organisational too.
What this model does not know
The sensitivity analysis produces an uncomfortable result that belongs up front: four of the six parameters that move the answer most are our own assumptions, not public data.
| Parameter | Swing | Nature |
|---|---|---|
| Administrative hourly cost | 17.7% | our assumption |
| G2V2 installation cost | 16.7% | our own workshop survey |
| Minutes per data download | 13.6% | our assumption |
| Number of international operations | 12.9% | assumption |
| Minutes of paperwork per operation | 9.0% | our assumption |
| ETS2 allowance price | 7.9% | our assumption |
There is no ministerial table for back-office hourly cost the way there is one for drivers, and there is no public measurement of how long tachograph administration actually takes. We declared wide ranges and left them wide.
The consequence is not the obvious one. The level of these numbers is sensitive to those assumptions. The structure is not. If the administrative hourly cost were higher than our range, the total would rise and the share taken by time would rise further still: the 97% in a steady-state year would become 98%. If it were lower, the total would fall and the ranking of the levers would be unchanged. The conclusion about the nature of the cost is more robust than the cost itself.
The urban permit assumption behind profile D is the weakest of the lot, for a reason worth naming: no national census of commercial-vehicle permit costs exists. Read that profile as an order of magnitude, not a measurement.
Method and sources
The model is a 20,000-draw Monte Carlo with a fixed seed, so it reproduces bit for bit. The code is analytics/scripts/model_costo_compliance_flotta.py, every parameter carries its source in a comment on the line, and the Monte Carlo and the sensitivity analysis share one calculation path by construction, so they cannot drift apart.
The regulatory and cost sources:
- Regulation (EC) 561/2006, as amended by Regulation (EU) 2020/1054, for the scope of driving and rest times.
- Regulation (EU) 165/2014, as amended by 2020/1054, for the G2V2 tachograph and the mandatory data-download intervals.
- Directive (EU) 2020/1057 for the posting of drivers. The declaration is filed per driver and per host Member State, valid for up to six months. It is not per operation, and that is an easy place to get the arithmetic wrong.
- Directive (EU) 2023/959 for ETS2, with payment starting in 2028 and a price-containment mechanism that releases allowances from the stability reserve above €45 per tonne in 2020 euros, a threshold that runs to the end of 2029.
- Commission Implementing Regulation (EU) 2018/2066, Annex VI, for the diesel emission factor: 74.1 tCO2/TJ with a net calorific value of 43.0 TJ/Gg. Combined with the density range in EN 590 (820–845 kg/m³) these give the 2.613–2.692 kg of CO2 per litre used in the model.
- Regulation (EU) 2026/1030 CountEmissionsEU, for the boundary of what is voluntary.
- The Italian transport ministry’s Table A of operating costs (March 2026) for employer cost per driver and the reference annual mileage.
- The Italian industry ministry’s weekly fuel price survey for diesel, aligned with our model of the full cost of a pharmacy delivery, because two of our own studies cannot run on different fuel prices.
- Installation and recalibration costs come from our own survey of quotes from authorised workshops, published in G2V2 tachograph on the Ducato, Master and Sprinter. That is our primary data, not a public tariff, and should be read as such.
To cite this study: Optivo, The cost of compliance for an Italian commercial vehicle fleet, 2026–2028, September 2026.
FAQ
Why does profile A cost nothing in 2026 and 2027?
Because the obligations that dominated the debate have a narrower scope than most people assume. The tachograph mandate and the driving-time rules apply to vehicles between 2.5 and 3.5 tonnes in international transport or cabotage, not to operators staying inside national borders on own account. If that is your fleet, the three-year window presents a bill only in 2028, and it arrives through the price of diesel.
Does the model hold for a fleet smaller or larger than 30 vehicles?
Most items scale with the number of vehicles or drivers, so the per-vehicle figure stays indicative across a wide range of sizes. Two do not scale linearly: the company card is a single card regardless, and the coordination time for running two regimes grows less than proportionally. On very small fleets the cost per vehicle therefore tends to rise.
Why is the ETS2 allowance price range so wide?
Because the €45 per tonne mechanism is not a hard ceiling. When the price goes above it the system injects extra allowances from the stability reserve, at most twice a year, which dampens volatility without capping the price. The threshold is also expressed in 2020 euros and expires at the end of 2029. We used €45–90 per tonne and flagged it as the most contestable assumption in the model.
Can the administrative time simply not be counted, since the staff are already on the payroll?
That is the most common position, and it is precisely why the line stays invisible. Not counting it is defensible only if those hours would have produced nothing else. If the manager filing tachograph data could otherwise have spent that time planning rounds better or looking after a customer, the cost is real, and it is equal to what those hours would have returned elsewhere.
Did you include penalties?
Not in the base cost. Modelling an expected penalty cost requires an enforcement probability, and no public measurement in Italy is granular enough to estimate one without inventing it. The amounts for individual infringements are documented in our piece on driving and rest times, but adding them here would mean presenting an estimate as a measurement.
If I redo the calculation with my own figures, do I get the same answer?
Almost certainly not, and that is the point. The script is open and every parameter can be changed: if your data-download times differ from ours, your cost differs. What we expect to hold is the structure, namely that in a steady-state year the dominant line is time rather than hardware.
How to use these numbers
The wrong way is to take the profile closest to yours and put the number in the budget. The useful way has three steps. Establish with certainty which regime you are in, because that is what decides whether the bill is €186 or €1,593 per vehicle. Measure how much time compliance activity actually absorbs, because it is the dominant line and the only one that never arrives as an invoice. Then act first on the organisational lever, because that is the one that repeats every year.
If you want to see how automated round planning builds Regulation 561 constraints in rather than checking them after the fact, book a demo. If you first need to work out which of the ten regimes apply to you, start from the fleet compliance map.