road user charge

how much does road traffic pay for itself?

A widespread assumption is that motorised traffic pays for itself through vehicle and fuel taxes. It does not. How large the gap is depends on which costs are counted, so every figure here states what it includes.

Two German reference points, both from credible sources, illustrate the range:

The two figures do not conflict; they draw different cost boundaries. Both establish the same structural fact: a large majority of the cost of motorised road traffic is not borne by those who generate it, and the remainder falls on the general public.

A global reference point. A 2026 study in Transportation Research Part A puts the full cost of a car kilometre worldwide at about 0,72 USD. Of this, 0,16 USD is borne by the driver and 0,55 USD by everyone else, so 77 % of the cost of automobility is externalised. Scaled globally, that is roughly 9,3 trillion USD a year, equivalent to 8,3 % of world GDP, with the unpriced share alone at 6,1 %. The estimate is conservative, because congestion and travel-time effects are excluded from the headline totals on methodological grounds.

To be checked : the driver and external shares add up to 0,71 rather than 0,72 USD, and the relationship between the 8,3 % and 6,1 % figures and the 77 % share should be stated precisely before the figures are used elsewhere.

Two independent estimates, one order of magnitude. The Munich calculation used in Socio-spatial equity and living lab research puts the hidden public cost at roughly 0,60 € per car kilometre; the global study puts the externalised cost at 0,55 USD. The two exercises use different methods and different geographies and arrive at the same order of magnitude. That convergence is considerably stronger evidence than either figure alone.

One finding shapes the assessment metrics directly. Collisions are by far the largest single cost component: close to 4 trillion USD a year, more than half of all social costs, and more than climate, air pollution and noise combined. The authors conclude that vehicle safety systems, speed regulation and infrastructure safety standards offer the highest return in reducing the social burden of motorised traffic.

Schweiggart, N., Gössling, S., Nieuwenhuijsen, M., Creutzig, F. (2026): The global cost of automobility. Transportation Research Part A: Policy and Practice 214, 105235. Open access: https://doi.org/10.1016/j.tra.2026.105235

The consequence is hard to overstate. A 2 t car carries about 1 t per axle. A 30 t lorry on three axles carries about 10 t. Ten times the axle load means something like 10 000 times the damage per axle, so a single heavy lorry can do the road damage of many thousands of cars.

The exponent is disputed. Measured values range from almost linear to considerably steeper than the fourth power, depending on road construction, temperature and moisture. The law gives the order of magnitude and the shape of the relationship, which is extremely steep, but not a precise figure. Local values are derived from national road authority data.

What follows for the charge:

• a maximum axle load of 10 t is one of the most cost-effective rules for city vehicles;
• charges for commercial vehicles must rise more steeply with mass than those for cars;
• an electric lorry that is heavier than its diesel equivalent increases this cost.

Space. The cost the charge is named after: street area taken up by moving vehicles and, in most cities a larger share, by parked ones.

Who bears it: everyone who could otherwise have used the space, and the city, which forgoes the value of an asset it owns.
How it is measured: with a lower bound from land value and the cost of capital, and an upper bound from revenues achieved where comparable space has been put to other use.

This cost is almost entirely missing from conventional cost accounts, which is one reason those accounts understate the urban case.

Emergency access: the space that must stay free. Every European country obliges planning authorities to keep parts of the street permanently clear for the emergency services: the approach to a building, the access to its entrance, and the standing areas for aerial rescue vehicles and portable ladders.

The dimensions are not small. A standing area for an aerial ladder is 5,50 m wide and 10 to 12 m long. A normal fire and rescue response needs some 50 m of standing space. And a carriageway has to keep 3,00 to 3,50 m clear wherever heavy fire engines must pass a narrow point.

Two consequences follow, and neither appears in any cost account we know of:

• A reserved claim on street space. These areas cannot be used for anything else. Where parked vehicles occupy them, the requirement is simply not met, and the failure stays invisible until the day it matters.
• A building-law consequence with a price. Where the standing area for the second escape route is missing, the building authority must prohibit using the affected rooms as living space. Parked vehicles thereby remove usable floor space from the housing stock, a cost borne by owners and tenants.

There is also a question of liability. Where a blocked approach causes an avoidable delay, and keeping it clear would have been possible, the question arises whether the responsible authorities have met their duty. That question is uncomfortable. It is also one reason why a systematic allocation of street space may be easier for an administration to defend than today’s situation, in which space goes to whoever occupies it first.

what is already paid, and what is not

Before any rate is set, this table is completed for each country. It is the difference between a fair charge and a double charge.

The pattern is the point: what is already priced is mostly what is easy to measure. What is not priced mostly falls on people who are not part of the transaction.

what the charge prices, and what it does not

A charge that claimed to cover every cost would be neither credible nor workable. It acts in three ways, and it leaves some costs to other policies.

The last three rows are deliberate limits. Whether any of them should become part of the charge, and how to avoid counting twice where one feature such as mass stands for several costs at once, is still to be evaluated.

how the costs are calculated

questions still to be decided

1. What is counted. Which costs are in, which are out, and why, stated once and applied everywhere. Most disagreement about cost totals is disagreement about this, not about the arithmetic.
2. Average or additional cost. The additional cost caused by one more vehicle for anything meant to change behaviour; the average only where the additional cost cannot be estimated, and marked wherever it is used.
3. No double counting, especially between air pollution, climate and health, and between space and congestion.
4. European or local values. The relative weights should hold across Europe; the absolute values should not be assumed to.
5. What remains public subsidy. Once the total is known and the charged share is decided, the difference is the subsidy the public continues to pay. It should be stated openly.

What this chapter does. It sets the convention on which the assessment metrics are built: what zero means, why deviations are counted as they are, and how every deviation is paired with the action that removes it. It also states a hypothesis about how that convention affects acceptance, which the research tests (RQ4).

The principle is simple: 0 is the standard, and every deviation from it is a measured gap. It was developed by Holistic-Governance.org as a general approach to stewardship metrics, and is documented at holistic-governance.org/stewardship-governance/metrics. In the general framework, gaps are named by urgency (minor, significant, critical) rather than scored, because there the question is what has to be done first. This document uses numerical gap points instead, because here the gaps have to translate into a price.

Zero as the reference point

Most rating schemes start at zero and let participants accumulate points towards a higher grade. That suits competitive settings. It suits a public duty less well, for three reasons.

Reference points shape behaviour. People evaluate outcomes relative to a reference point, and they weigh a shortfall below it more heavily than an equivalent gain above it. This loss aversion is one of the most robust findings in behavioural economics (Kahneman, D., Tversky, A. (1979): Prospect theory: an analysis of decision under risk. Econometrica 47(2), 263–291; Tversky, A., Kahneman, D. (1991): Loss aversion in riskless choice: a reference-dependent model. Quarterly Journal of Economics 106(4), 1039–1061). A scheme in which zero is the expected state makes any distance from it visible as something to close, not as a bonus that could additionally be earned.

Safety is a duty, not an achievement. Not endangering others in public space is the expected standard, not a voluntary extra. An accumulation scheme implies the opposite: that basic responsibility is something to be rewarded.

A fixed anchor allows comparison over time. A zero point stays where it is. In an accumulating scheme, the top of the scale tends to move as technology and expectations change, which makes comparisons between years and between vehicles difficult.

Why not an accumulation score

Points carry no sign

Deviations from Baseline 0 are counted as plain positive numbers, summed, and converted into a charge afterwards. A vehicle at Baseline 0 has nothing to sum and pays nothing; every point above zero is a measured distance from that state.

This is a communication decision as much as an arithmetic one. Negative scores would carry the loss-aversion effect described above, but they would also force the research to argue in negative numbers with manufacturers, with vehicle owners, and with the people least inclined to accept the instrument in the first place. “Your vehicle is at −47” invites a fight. “Your vehicle carries 47 gap points, and this is what removes each of them” invites a conversation. The behavioural effect comes from zero being the reference point, not from the minus sign in front of the deviation, and the reference point is preserved either way.

The convention throughout this document is therefore:

• 0 is the state of balance: the standard, not the floor, and achievable with vehicles available today.
• Gap points measure distance from it. They are positive and additive, and they are never described as a score to be maximised.
• Pathways to alignment are the actions that bring a vehicle back towards zero.

The vocabulary matters as much as the sign. The rating tables do not award marks for good behaviour and they do not issue penalties. They measure how far a vehicle sits from a state that can already be reached, and every entry has a corresponding action that closes the gap.

[Figure: zero is the standard — image to be re-inserted]

In the general framework, gaps are named by urgency (minor, significant, critical) rather than scored, because there the question is what has to be done first. This document uses numerical gap points instead, because here the gaps have to translate into a price.

The four states in the general stewardship framework. In this document, gaps are expressed as numerical points instead.

Applying the principle to a criterion

Each criterion in the assessment metrics is built in three steps:

1. Define the zero-state. What is the standard a vehicle meets when it uses street space responsibly? For direct vision, for example: a driver can see a child standing close to the front of the cab without relying on mirrors or cameras.
2. Measure the gap. How far does a given vehicle deviate from that standard, and how many gap points does that deviation carry? For example: mass beyond what the vehicle’s task requires, acceleration capability with no urban function, emissions that others breathe.
3. Name the pathway to alignment. For every gap, state the action that closes it: a lighter vehicle, a certified urban driving mode, a redesigned cab, a retrofit. Where the owner cannot reasonably take that action alone, the pathway is paired with funded support introduced at the same time as the charge (UMGA § 11(3)).

The third step is what distinguishes the instrument from a penalty. A gap without a pathway is only a charge; a gap with a pathway is an offer.

A hypothesis to test

The argument above is a claim about how people respond to a framing, and claims of that kind should be tested rather than assumed. The research therefore treats it as a hypothesis: presenting a charge as gap points with named pathways to alignment increases its perceived fairness and acceptance, compared with presenting the same charge as a score or as a penalty, while leaving the amount unchanged.

The hypothesis can be tested at low cost through a framing experiment in the proposed acceptance module of the Urban Mobility Barometer (see Socio-spatial equity and living lab research). Respondents are randomly assigned to one of three presentations of the same vehicle and the same charge, and perceived fairness and acceptance are compared across groups and countries. The result informs the design of the instrument’s communication directly, and it is a contribution in its own right to the literature on the acceptability of road pricing.

Usage and contextual evaluation

Alongside the vehicle-based rating, the Contribution varies with how and when a vehicle is used. Every verification mechanism below is designed to confirm an entitlement without collecting the underlying personal data (see UMGA § 10).

Open question. The combined share of journeys falling into these categories has to be modelled before rates are set. Exemptions that are individually justified can together erode the base to the point where the earmarked revenue does not materialise, and a charge that raises too little to fund visible local improvement loses the political case that justifies it.

Baseline 0

Baseline 0 describes vehicles whose presence in urban space imposes so little risk, and occupies so little space, that no Contribution is due. It is a state of the vehicle, not an exemption: a Baseline 0 vehicle carries no gap points at all (see The zero-baseline principle).

Which vehicles are at Baseline 0. Light electric vehicles with a maximum design speed of up to 50 km/h, acceleration in the low band and mass below the Baseline 0 threshold. They offer affordable, space-efficient and socially inclusive mobility.

What Baseline 0 is not. Vehicles operated by police, fire and rescue, municipal sanitation, public transport and certified medical transport are exempt from the charge under UMGA § 2(3), but they are not Baseline 0 vehicles. They are rated on the same scale as every other vehicle, their ratings are published, and new procurement must meet The safe city vehicle standard. A refuse vehicle is exempt from payment because of the service it performs, not because it imposes no risk.

Defining the mass threshold: open question. The current draft uses 450 kg. Two things must be specified before that number means anything:

• Which mass is measured. Kerb mass or technically permissible maximum laden mass, and, for electric vehicles, whether the traction battery is included. For a light electric vehicle the battery can be a large share of total mass, so the answer changes which vehicles qualify. The EU vehicle categories for light and heavy quadricycles (L6e, L7e) set their mass limits excluding the traction battery, which is where figures of around 425–450 kg come from (limits to be verified against Regulation (EU) No 168/2013). A threshold taken from those categories but applied to mass including the battery would exclude most of the vehicles it was meant to admit. Provisional position: kerb mass including battery, excluding driver and payload; to be confirmed.
• Step or taper. A single hard threshold creates a cliff edge, and cliff edges are where product design and lobbying concentrate. As currently drafted, a 449 kg quadricycle and a 25 kg e-bike are treated identically, while the same quadricycle at 460 kg falls into the full regime. Courts have also found such steps problematic (see the Freiburg case in Dynamic digital curb management).

Proposed structure, for discussion:

Speed and acceleration are independent gates. A vehicle capable of more than 50 km/h, or of more than 1,5 m/s² without a certified urban mode, enters the full regime regardless of its mass.

Open question: where the upper limit of the taper sits.

Vehicle-based metrics

The metrics below cover safety for vulnerable road users (Table 1); climate, exhaust and non-exhaust emissions (Table 2); and mass, speed capability and spatial footprint (Table 3). Direct vision for heavy vehicles and the access criteria derived from the city truck concept are developed further in The safe city vehicle standard.

Acceleration bands

Acceleration capability is assessed in three bands, separated by two thresholds:

A certified, tamper-evident urban driving mode counts: a vehicle that is limited to the low band within the urban area is assessed in the low band there.

Three acceleration values appear in this document, and they serve different purposes. They need not be identical:

• 1,5 m/s² is the upper edge of the low band and a gate for Baseline 0 (below).
• 3,0 m/s² separates the medium from the above-medium band in Table 1.
• 1,2 m/s² is the comfort and energy target for public service vehicles in normal operation (see A modular public autonomous vehicle platform). It lies inside the low band and is a service specification, not a rating threshold.

Table 1: Safety for vulnerable road users and active protection

Draft. The structural problems flagged in the notes matter more than the point values.

Note A: protection of vulnerable road users cannot be expressed as a star rating. Euro NCAP does not issue a star rating for vulnerable road user protection. It issues one overall star rating plus percentage scores across assessment areas, of which vulnerable road users is one. From 2026 the scheme is restructured into four new areas (safe driving, crash avoidance, crash protection and post-crash), so a metric keyed to the previous structure would be obsolete on arrival.

Two further problems have to be solved rather than assumed away:

• Ratings are not comparable across years. A five-star result from 2015 reflects a materially weaker protocol than a five-star result from 2025. Any use of assessment scores must be adjusted for the year of assessment.
• Most of the fleet in use was never assessed. A scheme that assigns the worst score to “unrated” would penalise an untested small hatchback more heavily than an assessed large SUV, the exact opposite of the intended effect. Untested vehicles need a fallback rating derived from measurable physical characteristics (frontal geometry, bonnet leading-edge height, mass, bumper profile) that can be read from type-approval data or measured directly.

The evidence for a geometry-based fallback is strong. Bonnet height is the single physical characteristic most clearly linked to pedestrian fatality risk:

• A 2023 study by the Insurance Institute for Highway Safety, based on some 18 000 collisions with pedestrians, found that vehicles with a bonnet height above 102 cm are 44–45 % more likely to kill an adult pedestrian.
• A 2024 study by Tyndall, based on some 3 400 collisions, found that each additional 10 cm of bonnet height raises the probability of death by 22 % for pedestrians aged 18 to 65, by 31 % for those over 65, and by 81 % for children.
• A 2023 Belgian study by the road safety institute Vias, based on national crash data, found the risk of death for a pedestrian or cyclist struck by a pickup truck to be almost 200 % higher.

For comparison, a typical US pickup has a bonnet height of about 126 cm, while a small car such as the Renault Clio has 70 cm or less. Sources as summarised in the civil society letter on individual vehicle approval of 4 July 2024 (T\&E, ETSC and others); to be cited from the original studies.

Research task: define a rating for vulnerable road user protection that is usable across the whole in-use fleet, based on assessed scores where they exist and on physical fallback characteristics, led by bonnet leading-edge height, where they do not. Relevant expertise: Euro NCAP, PACTS, Loughborough University.

Note B: direct vision is not yet rated for most vehicles. The Direct Vision Standard applies only to heavy goods vehicles over 12 tonnes entering Greater London, where a minimum three-star rating or a progressive safe system has been required since October 2024. The EU General Safety Regulation introduces a direct vision requirement for new heavy goods vehicles and buses, applying to new types from 2026 and to all new vehicles from 2029 (dates to be verified), but that is a pass-or-fail design requirement, not a rating. For cars and vans no direct vision rating exists anywhere.

This is an opportunity rather than an obstacle. Loughborough University, where the Direct Vision Standard methodology was developed, has indicated willingness to extend it to all motorised vehicles above a speed threshold still to be defined, in the region of 30–50 km/h (subject to confirmation). A direct vision rating for cars and vans (M1 and N1) would be novel, transferable and immediately usable by any city operating access restrictions. It is one of the outputs of The safe city vehicle standard.

Note C: rate performance, not fitment. Assigning points for the absence of specific safety systems was considered and rejected, for two reasons:

• Several of these systems are already mandatory for new type approvals under the EU General Safety Regulation (2019/2144). Pricing their absence would price only the age of the vehicle, which the emissions and mass tables already capture.
• Fitment is not effectiveness. Emerging, not yet published evidence from UK research indicates that blind spot information systems and moving-off information systems perform poorly in practice, with high false-positive rates. A metric that rewards fitting such a system rewards the wrong thing, and manufacturers would comply by fitting weak systems. To be cited once the findings are published.

Provisional position: blind spot information and moving-off information systems are held out of the scoring until published findings are available, and remain on the research agenda. Systems are scored on measured performance (results of assisted driving tests and emergency braking tests for vulnerable road users) where such results exist, and not scored at all where they do not.

Table 2: Climate, exhaust and non-exhaust emissions

Exhaust emissions are scored separately for cars and light vans, for powered two-wheelers and light vehicles, and for commercial and heavy-duty vehicles, because the emission standards, the substitution options and the social incidence all differ. Values are provisional, and the scales are to be replaced by a common scale based on real-world emission factors (see What remains open in this chapter).

Exhaust emissions: passenger cars and light vans (M1, N1)

Non-exhaust emissions, noise and time of day

Non-exhaust emissions deserve emphasis. They scale with vehicle mass and do not fall with electrification. As the fleet electrifies, tyre and brake wear become the dominant source of traffic particulates (see The cost base, category 6), which is one reason the mass table (Table 3) carries as much weight in this scheme as the exhaust table.

Open question. Tyre abrasion limits are still being developed at EU and UNECE level. Until they apply, the medium and high bands need a measurable definition, for example based on tyre width and vehicle mass.

Powered two-wheelers and light vehicles (L category: L1e–L7e)

Low-cost electric alternatives are widely available in this segment, which is what makes a strong price signal effective. But the same fact that makes substitution easy also makes the incidence uneven. Powered two-wheelers are owned disproportionately by younger and lower-income users, and in several member states, such as Italy, Greece and Spain, they are a mass mode rather than a niche. A charge in this segment without a matching support measure is regressive and will be quoted against the project. The support measure is therefore part of the table, not a separate paragraph elsewhere.

The two-stroke rule, and the fairness condition attached to it. Two-stroke engines emit disproportionately high levels of fine particulates and volatile organic compounds relative to their spatial footprint, and the environmental case for a high rate is strong. The 250-point figure is provisional and, at roughly seventeen times the rate for a Euro 5 four-stroke, at the top of what is defensible. Two conditions apply before it is adopted:

• the scrappage grant is available before the charge starts, not promised for later;
• the rate is checked against income data in the member states where two-stroke ownership is highest, so that the design is known to be progressive rather than assumed to be.

Commercial and heavy-duty vehicles (N1–N3, M2–M3)

Commercial vehicles (trades, logistics, municipal services, buses) need a framework that recognises the established clean standards while addressing ageing diesel fleets, whose high annual mileage makes them disproportionate contributors to urban air pollution.

The 150-point rate for the oldest tier falls hardest on sole traders and small businesses running a single old van, a constituency whose opposition has defeated urban charging schemes elsewhere. The transition period and the support measure are not concessions to be negotiated away; they are what makes the measure implementable.

Open question: consistency. A light van at Euro 6d carries 0 points here, while a passenger car at the same standard carries 3 in the table above. Heavy vehicles at Euro VI carry 0, the same as zero-emission vehicles. Both can be justified, for example by the essential function of commercial vehicles and by the limited availability of zero-emission alternatives in some segments. But the justification has to be stated, and it has to be revisited as zero-emission alternatives become available segment by segment. Otherwise zero stops meaning “achievable today” (see The zero-baseline principle).

Open question: legal. Charges applied to heavy goods vehicles interact with the Eurovignette Directive (1999/62/EC as amended by Directive (EU) 2022/362), which governs how member states may charge heavy goods vehicles for infrastructure use and external costs, including reference values for the external-cost element. Whether, and how, a municipal space-and-safety contribution on N2 and N3 vehicles can sit alongside that framework needs a legal opinion before the heavy-duty rates are proposed publicly. It is the first point a freight association will raise.

Table 3: Vehicle mass, speed capability and spatial footprint

Three design choices shape this table.

Mass alone is the wrong proxy. The safety argument of this instrument is a kinetic energy argument, and kinetic energy scales with mass × speed². A 2 400 kg vehicle limited to 30 km/h is considerably less dangerous to a pedestrian than a 1 600 kg vehicle capable of 70 km/h. Mass therefore interacts with a speed capability term. This also opens a compliance route that does not require manufacturers to build lighter vehicles: a certified, tamper-evident urban speed and acceleration mode (see Acceleration bands). That makes the measure far more politically survivable, and it is something the vehicle industry can deliver in software.

Footprint belongs in the calculation. This is a charge for the use of public space, so pricing only mass would miss the thing being charged for. Length × width is readily available from type-approval data and is the most direct measure of space occupied. It is therefore a separate term, and Koblenz already uses exactly this measure for its parking charges (see Dynamic digital curb management). The evidence for why it matters is published: the Clean Cities carspreading analysis (June 2026) puts the consequences of growing vehicle size at the equivalent of 10 % of UK car parking spaces consumed, and a 40 % increase in the number of child pedestrians who could be killed in collisions (headline figures from the published summary; to be cited from the full report). That is the case for a footprint term and for the direct vision work in one study; what is missing is an instrument that acts on it.

Tier boundaries must match the fleet. A lowest tier for multi-seat vehicles set below 1 000 kg would be empty in practice, because essentially no current five-seat car falls below it. At the other end, several large electric SUVs and imported pickups now exceed 3 500 kg. Boundaries are therefore set against the actual distribution of the European fleet.

Provisional structure

Speed capability factor (to be derived). A multiplier below 1,0 for vehicles with a certified urban speed and acceleration limitation, rising above 1,0 with maximum design speed. Its functional form follows from the kinetic energy relationship rather than being chosen for convenience.

Commercial vehicles. Gap points for mass reflect both kinetic energy and road wear, which rises very steeply with axle load (see The cost base, category 2). The reduction for meeting The safe city vehicle standard is the mechanism by which a heavy vehicle can be made acceptable in a city rather than simply priced out. A refuse vehicle or a delivery truck cannot be replaced by something light, so the instrument must reward making it safe.

What remains open in this chapter

The tables show the shape of the instrument, not a finished metric. Five questions have to be answered before they produce a charge, and they are answered together with The cost base and Socio-spatial equity and living lab research:

1. A common scale across vehicle classes. Each table is internally consistent, but the classes are not comparable with each other, and cannot be as long as points are attached to the name of an emission standard. Euro standards were introduced at different times and with different limit values for each vehicle category: “Euro 5” means a standard from 2020 for motorcycles, from 2009 for cars, and a different limit set again for heavy-duty vehicles. The research therefore derives exhaust points not from the label but from the real-world emissions it corresponds to, using the established European emission factor databases (COPERT, used by the European Environment Agency for national inventories, and HBEFA, used in Germany, Austria, Switzerland and Sweden), weighted by the damage cost per unit of emission from The cost base. The emission standard then serves only as the key for looking up the right factor, and the tables above become a readable summary of the result rather than its source.
2. The combination rule. How the three tables and the usage factors combine into one assessment. The speed capability factor already multiplies the mass points, so the scheme is partly multiplicative; whether the rest is added or multiplied changes the result substantially.
3. The value of a point. How many euros one gap point is worth in a given city. That is the political level, set in Socio-spatial equity and living lab research, not here.
4. The mass basis and the Baseline 0 threshold (see Baseline 0).
5. Tier boundaries fitted to the fleet, using registration data from the partner member states.