In the past, like any other country in the world, Europe is also charged fuel taxes and some fixed taxes to recover expenditure on its highway system [2]. The aim of the introduction of Eurovignette in the EU was to recover construction, maintenance, repair and environmental costs on their road network. More importantly, ensuring unbiased competition and stopping discrimination among hauliers in the member states. However, this was a time-based charging system that is jointly operated by Belgium, Denmark, Luxemburg, the Netherlands and Sweden [39]. The main problem found with the Eurovignette system was the charges do not consider the distance travels by a vehicle, but the time, and therefore poorly correlated with infrastructure costs. On the other hand, distance-based tolls found to be corresponding much more closely to the cost of infrastructure use [40].

Later they changed to “polluter pays” principle from the “user pays” principle to allow for the internalization of the external costs. Many countries came up with their charging mechanisms considering different aspects. According [3], in Europe, road generated revenues exceed road expenditures in all the countries studied. This means road charges have subsidized other policies. To understand the basic principles few major policies in Europe are first reviewed here.

1) German heavy goods vehicles charging system

Heavy good vehicles over 12 tonnes on German motorways have to pay a distance-based charge since 2005, January [40] [41]. This charge was imposed on top of existing taxes and hauliers have continuously argued that the charge was too high.

The German HGV tolling scheme is aimed at goods vehicles over the gross vehicle weight of 12 tonnes to recover the full cost. This includes all costs of construction, maintenance and operation of motorways and also charging scheme operating cost. The charge level is determined based on vehicle characteristics such as weight, number of axles, and the emission class [10]. The introduction of this charging scheme in Germany in 2005, known as LKW-Maut, resulted in a large number of heavy vehicles avoiding tolls by means of diverting from Western Germany to France where the main highways are not tolled yet [5].

The latest advanced technology is used for the German HGV charging scheme. Satellite positioning and mobile telephone communication technology cover all the required aspects of such a charging system. A mobile telephone connection was made between the vehicle on-board unit (OBU), which combines a GPS receiver and a digital map and charging centre. This acts as the transponder for the charging scheme.

2) UK Lorry charging scheme

The British government had planned to introduce a lorry road user charge system (LRUC) in 2008 but was abandoned the introduction in the year 2005 [5] [40]. The charges would have applied to all trucks over the gross weight of 3.5 tonnes, over the entire road network and offer the varying charges by the time of day, road type and zone. The reason for withdrawal was the complexity of the proposed project and the government decided to introduce a simple distance-based user charge system. A device called tachograph was installed to all heavy vehicles over 3.5 tonnes, operating within the EU, to measure the distance. This installation was compulsory and distance travelled during each year was observed at the point of registration and tolls are calculated accordingly. The British government did not charge for the mileage run in other countries by vehicles registered in the UK but engaged in international haulage. A fuel rebate was given based on a benchmarked fuel efficiency level to omit double taxation. This benchmarked system has certainly penalized vehicles underperforming in terms of efficiency and encouraged vehicles with high fuel efficiencies [5]. This toll charging system was purely focused on charging a fair share for using the UK roads by overseas hauliers rather than generating additional revenue for transport investments [40].

Since 1985 flat fee for Heavy Goods Vehicles (HGV) was in existence and it was changed to a nationwide distant related fee, named as the LSVA, since Swiss government predicts a 100% increase in HGV on the Swiss road network as a result of bilateral treaties made with the European Union to extend the maximum weight limit of HV from 28 tons to 34 tons, and to 40 tons later on. After a referendum and a subsequent vote, the LSVA started on January 1^st 2001.

The LSVA applies to all domestic and foreign heavy vehicles transporting either goods or passengers with a maximum laden vehicle weight of over 3.5 tons. For domestic vehicles installation of an On-Board-Unit (OBU) was mandatory. Foreign vehicles are predominantly using a ticket fetched at self-service machines. The distances travelled by each HGV, on all public roads (not only motorways) in the country are measured using the tachograph. A tachograph is switched on and off when crossing the border. Charges are calculated based on the distance travelled maximum permissible vehicle weight and emission class of the vehicle [42].

Trucks and buses on Austrian roadways were subjected to a time-based toll system until 2004, where distance-based pricing was introduced [43]. In Austria, there’s an OBU called “Go Box” and all trucks over 3.5 tonnes of gross weight are mandatory to fix an OBU. Light vehicles still have to pay the time-related user fee based on Vignette. Gross weight and axle configuration (3 classes) are considered when determining road user charges per kilometer travelled, but not emission parameters unlike in Switzerland. Only motorways and express roads are subjected to toll, as a result, diversion of traffic to local and regional parallel roads was inevitable. However, the proportions are found to be insignificant (less than 2% of the total traffic on motorways).

In the United States, highway users pay for their roadway use through indirect user fees that are levied at the federal, state, and local levels. Federally, the primary sources of truck user fee revenues are fuel taxes. Other federal truck user fees include sales taxes on trucks, tractors, trailers, and tires, and an annual fixed-rate heavy vehicle use tax (HVUT) [3]. These tax rates vary depending on gross vehicle weight (GVW). Other than that weight-distance tax (WDT), overweight and over-dimension permit fees try to recover some components of the road damages in some states.

The revenue generated primarily through fuel taxes found to be insufficient to maintain the federal highway system (public sector subsidizes the road system) and many suggestions were proposed in the past as measures to increase revenue. Increase in fuel taxes, at least to keep pace with inflation, spreading the use of toll highways, imposing new taxes and fees, and encouraging private investments such as PPP’s are among them [3] [14] [44].

Vehicle Miles Travelled (VMT) based fee has been identified as the most suitable option in the long run in the US because it could establish a reliable source of funding, reduce traffic congestion, promote more efficient use of vehicles with higher loading capacities, and discourage unladen miles travelled. However, implementing a usage-based pricing scheme such as VMT based faced several challenges. For example, public and political acceptance, technology, administration and financial feasibility [18] are a few of them which are discussed in detail later in this paper. As a result, fuel taxes still exist in the United States [3].

At present, there are more than 300 toll roads in the USA [45]. Most toll roads define their base toll rate structures according to the vehicle number of axles and some HVs are tolled according to registered GVW. Taken together, the analyses conducted by [45] have strongly concluded that toll policies across the United States have penalized large commercial vehicles disproportionately. In his study, the toll charges were compared for various types of vehicles considering the following parameters. Vehicle lengths, load equivalency factors, consumption of road space and the damage caused to the pavement are such parameters.

Congestion has emerged as one of the greatest challenges faced by transport planners and authorities in developed economies. It’s a major social cost to the society and the cost was estimated as high as $78 billion in 2005 in the USA and is growing rapidly as a result of speedy increases in travel delays experienced by users every day. The cost was estimated not only in terms of lost time but productivity, air pollution, and energy wasted is considered. As a solution to these severe levels of congestion, the U.S. Department of Transportation has recently started a program to initiate congestion pricing in five metropolitan areas [46].

State Highway Cost Allocation Studies (HCAS) have a long history in the United States spanning over 70 years. More than 80 studies have been performed in at least 30 states during that time [47]. Federal Highway Cost Allocation Studies (HCAS) was completed and concluded in 1997 in the United States [48] and no major changes have occurred since then. According to [18], this self-financing scheme was designed to attain maximum efficiency followed by horizontal equity. This comprehensive usage-based charging system will enable self-finance construction, operation, and maintenance cost of the highway system and calculates user fees based on vehicle class, weight, configuration and distance travel on each facility. However, externalities produced by vehicles are not considered in this user charging scheme such as congestion and emissions and thus authors have stated that future studies need to introduce such cost components into the scheme [18] [48]. In the year 2000, HCAS was revised to reflect social costs calculated by the Environmental Protection Agency (EPA) in the US. As a result, crash, congestion, air pollution and noise costs were introduced and user fee based on vehicles type and environment condition (urban/rural) was re-calculated [49]. However, any of these HCAS were not fully implemented in United States mainly due to political setting that focuses on painless revenue generation from established sources rather than instigating equity among highway user classes [47].

Heavy vehicles (any vehicle or trailer with a mass over 4.5 tonnes) in Australia are subjected to two pricing charges. Namely; a road user charge (RUC) and an annual registration fee. The RUC is a levy on each liter of diesel an HV consume and presently it’s about 38.14 cents per liter. The National Transport Commissions (NTC) provides recommendations for setting these heavy vehicle charges which are based on a charging framework known as pay-as-you-go (PAYGO). This was introduced in Australia in 1992 [15] [50].

The PAYGO system tries to recover the road usage cost from each heavy vehicle type and to recover a share of common road costs, such as street lighting, rest bays and signage. PAYGO scheme follows a simple mechanism to calculate heavy vehicle charges using the latest heavy vehicle and trailer population data available at NTC. They have considered seven-year averages for both road expenditure and vehicle usage data assuming that charges do not change significantly in response to short-term changes in expenditure or vehicle use [16] [50].

In Australia, revenue recovered through heavy vehicle charges helps the government to provide better and safer roads. According to NTC, around 40 percent of heavy vehicle costs are recovered as state and territory registration fees, with the balance paid through a fuel-based road user charge which is collected by the Commonwealth Government [50].

In Australia approximately three-fourths of the HV fleet is owned by ancillaries and rest is by for-hire. It is obvious that for-hire use more road kilometers than ancillaries and registration cost is determined based on a cost-recovery scheme which considers historic expenditure on-road services. Therefore, the registration fee determined for vehicle type irrespective of its ownership or real usage leads to very high inaccuracy. In addition, there are substantial variations in registration costs between states in Australia for no valid reason. Therefore, cross-subsidization is very high in Australia.

As mentioned before, Australia became commonplace for PPP and has a growing number of toll roads in main cities and claim to be the country having the largest (in terms of lane kilometers) urban tolled road network in the world [20]. Based on the study carried out by [51] it was found that freight movements in the Melbourne metropolitan region is higher than interstate freight in Australia, which explains why toll roads with PPP are popular in cities of Australia. In general, toll charges in Australia are based on the distance of travel, vehicle type (very broad classification) and Time of Day. However, toll charges vary over facilities and sections with in the same facility as highlighted by [52] [53] in their studies. Australia firstly introduced the ETC system in the CityLink project for toll collection [54]. Despite the availability of the latest technology in CityLink appropriateness of toll charges (representing true user charge) stipulated in CityLink road is deplorable. As a result, a noteworthy social dialog has begun among the truck community and stakeholders recently in Australia [55].

It is a well-established fact that an introduction of a new tax or toll is not usually welcomed with open arms by the public. As a result, any transportation solution that involves such a levy has a political implication and complicated to choose [58 & 59]. Usually political and public acceptance is linked together where public opinion is mostly concerned by politicians. As a result, it could be stated that road pricing schemes are politically accepted only if it gets sufficient public support. On the other hand, it has been found that strong political decision is a positive measure of higher public acceptance [59].

Road pricing schemes, known as the first-best solution in theory [60], could be limited to the drawing board unless political and public acceptance is achieved. Even though economists are certain about its intended benefits [61] [62], many past proposals have proven that public and political acceptance is a must for successful implementation of such a scheme regardless of how technically well it sounds or the extent to which socio-economic benefits it could generate [24] [59] [60] [61] [63] [64].

Nevertheless, the most refined and reflective policy design will create some losers and induce opposition. In fact, from past failures, it can be stated that a lack of public acceptance, correlated with political acceptance, has been the most important obstacle to implement road pricing charges. Among many such failures reported in the world, cordon schemes were rejected by public referenda in Edinburg and Manchester [63], combined cordon and zonal scheme for Manhattan stopped by the New York state legislature in 2008 [63] and congestion metering proposal rejected in the city of Cambridge are few of them. According to [65], the Cambridge proposal was considered as a concept in the right place at the wrong time reasoning out the public refusal.

As mentioned above, a systematic economic rationale does not always assure political acceptability [66]. Politicians are always keen to apply known or well-established measures such as fuel taxes and registration fees since these measures are more acceptable than that are less known or unknown ones and generate revenue effortlessly [41]. Studies from Europe have found that the acceptance of tolls by stakeholders is high when it applies to all types of vehicles across all countries in Europe. This indicates that equity is greatly concerned by stakeholders. In conclusion, policymakers should endorse the perception of fairness into road pricing to gain stakeholders’ acceptability [67].

Identification of key stakeholders and their expectations is crucial for the successful implementation of the road pricing scheme. Different stakeholders seem to have various expectations that may not in line with the objectives of the road pricing scheme and thus create a negative voice. But there were many proposals failed in the past with perfect schemes due to poor presentation, unclear objectives or improper timing [65] [68]. As a result, it can be concluded that clear objectives, effective communication and re-investment of revenue are primary to win the majority of the stakeholders. As per the prediction of the approach-avoidance theory, public attitudes were mostly adverse just before the introduction of road charges but this dynamic attitude could be changed with test runs [59].

As explained in cognitive dissonance theory [59] [69] people initially tend to refuse the road pricing scheme if the implementation is doubtful, however, more people react positively if the proposal is known to be implemented anyhow. Therefore, uncertainty among policymakers and politicians about the scheme, usually based on its promised results, may lead to more anxiety when trying to win public acceptance [70]. At the same time, the non-availability of a similar system for comparison or maybe a success story to rely on could reduce the acceptance of such schemes at an introductory level [65]. In summary clearly stated objectives, the process and type of data collection and associated technology, and how the revenue raised from the charge is disbursed and detailed information about implementation costs are critical factors that need to be given careful consideration before announcing a road pricing scheme.

Equity issues make the policy proposal politically weak. Horizontal equity and vertical equity are two dimensions that policymakers work on when distributing revenues generated from road pricing. However, who deserves the benefits is yet a debate according to [71]. It could be those who paid the toll or those who change their trip pattern in response to toll providing better conditions for toll payers. In the early nineties, [62] argued that the “political response will depend on the relative numbers of losers and gainers and the amount of utility lost by those motorists who are priced off”. But this has grown to a far more extent over the years where equity is concerned.

Apart from whether the public accepts this proposal or not, it was found that politicians have concerns about technology requirements, costs of technology and administration sometimes. Furthermore, enforcement issues and uncertainty with data collected (privacy issues) makes them more worried before presenting to the public [65].

Supporting technology for effective collection of RUC in terms of tolls is a must to the successful implementation of any RUC scheme. Starting from manual toll collection systems in early days, where toll booths are maintained with human support to collect tolls or with automatic coin machines [72] have transferred to more technical based systems such as video or CCTV cameras based systems with the aid of license plate recognition software’s and so on. For example, infrared cameras with ANPR technology have been deployed on routes into the Stockholm city centre [73]. Initially, the tolls levels were kept flat due to charging (technical) and administrative difficulties [74] but with the advent of new technology such as automatic vehicle identification (AVI) systems make it feasible to invent sophisticate pricing systems in a user-friendly manner [58].

The present era of this technology is Electronic Road Pricing (ETC) which uses gantries and beyond that satellite-based ETC system [73] are now in place in Germany to charge lorry road users. A mobile telephone or GSM based technology has been discussed for some time, but the idea was rejected majorly due to privacy issues. The next generation of technology for RUC could be GPS (satellite) based charging mechanism where data about many parameters such as time of the day, location (geographical location and corresponding road type in use), speed profile, distance travelled, time held up in congestion can be collected promptly. Once data is ready it’s just a matter of time to calculate user charge for each vehicle both in terms of direct and externalities using the available schemes, given suitable schemes are available and imposing them on users.

It is worth mentioning that technology used for road pricing charges has led to several problems in the past at the implementation stage and some schemes were even abandoned as a result. Gaining political and public acceptance in the technology used is yet another hurdle in road pricing schemes. When Cambridge tried to introduce congestion metering, the complexity and acceptability of the technology were perceived as a major disadvantage [65]. A device was introduced to be fitted on each vehicle and there were many negative concerns regarding such installation. Respondents expressed that a scheme with less-advanced technology would be more acceptable in the first instance [65]. Similarly, in Hong Kong, the public was doubt about electronic road pricing technology, which was new, untested in a real situation and more likely to fail [75]. Further, the public was suspicious that the government would use this data to tax drivers callously in the future and thus only a few agreed to have their cars fitted with electronic number plates [75]. Also, some groups in Hong-Kong were opposed to electronic road pricing considering it as being technically unrealistic, ethically suspect, absurdly expensive, and socially divisive [75]. Invasion of privacy is also challenged by the public when this nature of advanced technology is used on roads [25] [75] [76]. Experience from Taiwan shows that initial acceptance of ETC was lower than expected [77]. On the other hand, studies from the UK have shown that this is one of the least concerned items by respondents in the UK [68]. Therefore, it can be concluded that technology and data privacy is decisive to introduce and promote a road pricing scheme and needs to be handled carefully.

[78] studied the acceptability of technology by freight vehicles, especially acceptance of the ETC system. From the study, it was found that freight vehicles do not accept electronic tags compared to private cars due to various reasons. Unawareness was found to be the greatest challenge faced by freight vehicles hindering the usage of ITS features. Business attributes and attitudes towards technology were found to be highly correlated with freight carrier willingness to accept technology.

The traditional weigh limit enforcement is by static weighing. Weighbridges, wheel and axle scales are used to measure gross vehicle weight and wheel or axle loads [79]. Fixed or portable, static weighing systems suffer from several limitations. Time taken, staff requirements and practical difficulties to measure each truck on a highway are major issues. With the development of technology high-speed weigh-in-motion (WIM) technology has resolved key issues. A more sophisticated version of WIM sensors is available today in the market which can be fixed to the vehicle and communicate with the central control centre.

Usually, the cost of operation is a major consideration in the implementation of a more equipped, technically advanced system and it was found to be approximately 20% of the revenue collection based on the literature. However, this was found to not be the case with the Swiss system, being 4% - 6% of the revenue which is comparatively very low. Thus, it can be considered as the cost is not a constraint anymore for this nature of the application. Moreover, it’s a well-accepted fact that technology is advancing every day and more cost-effective devices are coming into the market.

Crash costs (Accident costs) are calculated as an exposure function considering past data in Victoria, Australia. Thus, it’s a function of many vehicles in each class on different road types as mathematically constructed below.

A C a m = f ( x a m )         ∀ a ∈ A ∪ A ¯ ,     m ∈ M (1)

The flowchart in Figure 1 depicts the crash cost calculation process and Table 3 depicts the final figures applicable for Australia. Since crash data, vehicle kilometers traveled and injury cost data are available in Victoria, Australia the method was successfully adopted to calculate crash costs.

Therefore, the total crash cost can be calculated as:

∑ a ∑ m A C a m x a m d a         ∀ a ∈ A ,     m ∈ M (2)

Congestion costs (A$/km) are a function of the v/c ratio and class-specific vehicle volumes, x a m .

C C a m = f ( x a m , x a / C a )         ∀ a ∈ A ∪ A ¯ ,     m ∈ M (3)

The total congestion costs can be calculated as:

∑ a ∑ m C C a m x a m d a         ∀ a ∈ A ,     m ∈ M (4)

The relevant cost figures are extracted from literature and given in Appendix A.

Infrastructure costs (IC) can be broadly categorized into three elements. Namely, fixed costs, variable costs, and profits from infrastructure development (e.g. PPP Projects). Fixed costs (capital investment) can be shown as a cost per kilometer based on the number of lanes of the road which affect the capacity. Variable costs can be subdivided into two based on its nature of expenditure. First, as non-usage based fixed annual costs and secondly as, usage-based (variable) element. Finally, profits are determined based on market rates and the terms and conditions of the agreement made with the road authority. Since some road links are built under PPPs and other links are mainly funded through government budgets using the taxes of residents, capital cost investments, expected profit are quite subjective. Therefore, the IC calculated in this study is mainly looking at maintenance cost only for various infrastructure types caused by multi-class traffic.

Therefore, the cost of infrastructure (A$/km/veh. type) can be considered simply as a function of a number of vehicles on that link for a unit of time (per year).

I C a m = f ( M a m , x a m )         ∀ a ∈ A ∪ A ¯ ,     m ∈ M (5)

M a m is taken from the literature (see Appendix B). The numbers given in Appendix B are cross-checked with the Australian (Victorian) real numbers and matched well. This cost can be calculated as:

∑ a ∑ m I C a m x a m d a         ∀ a ∈ A ,     m ∈ M (6)

For this study emission costs generated by different vehicle types are calculated considering several harmful pollutants. Existing models and data were used to arrive at the final numbers and the process is briefly explained here. Fuel consumption for multi-class vehicles under different speeds and various emission components from fuel-burning were extracted from the [95].

Since the type of emission not only varies with vehicle type but also with vehicle age, an assumption was made to simplify the calculation complexity by considering all heavy vehicles were manufactured in years between 2003-2008, and for all light commercial vehicles, from years 2004 to 2005. In other words, a representative age was selected for each vehicle type. For simplicity, only the most significant and harmful emission types were considered, namely, carbon dioxide (CO₂), nitrogen oxide (NOx), sulfur oxide (SO_x) and particulate matter (PM2.5).

To calculate the damage cost of each pollutant type, a damage cost model developed by the European Commission was used [96] [97]. In that model, the authors considered the health impacts on humans by toxic pollutants from road transportation and quantified the impacts in monetary terms. Impacts of GHG’s were considered separately in terms of damage to buildings and materials, crop losses and biodiversity and ecosystems. The costs were converted to A$ (Australian Dollars) using the GDP ratio as instructed in the report. Therefore, emission costs can be written as a function of:

E a m = f ( e i m , e c i , v a , f m ( v ) , x a m ) (7)

where

e i m —Different emission types (i) given by weight (g or kg) per vehicle type per liter of fuel consumed (g or kg/one liter of fuel);

e c i —Cost by weight (g or kg) for different emission type (i) (A$/g or kg)

v a —Average speed on the link a (km/hr);

f m ( v ) —Fuel consumption per kilometer travel by each vehicle class under given speed (liters);

i = 1 for CO₂, i = 2 for PM2.5, i = 3 , for NO_x and i = 4 for SO_x.

Since the speed is classified into two sections, namely, free-flow conditions and stop-go conditions, there are two main scenarios in the emission cost calculation. After plotting all numbers in a chart, the following equation was extracted as a representative emission cost formula under different driving conditions. The coefficients and other relevant statistics are presented in Appendix C.

The total emission cost can be calculated as:

∑ a ∑ m [ ( a 2 v a 2 + a 1 v a + a 0 ) γ + ( c 0 + c 1 v a ) β ] x a m d a         ∀ a ∈ A ∪ A ¯ ,     m ∈ M (8)

if   100 ≥ v a ≥ 60 ,     γ = 1 ,     0     otherwise (9)

if   60 > v a ≥ 10 ,     β = 1 ,     0   otherwise (10)

where,

a 2 , a 1 , a 0 , c 0 , c 1 coefficients (values are given in Appendix C).

Equation (7) denotes the emission costs as a function of its variables and Equation (8) denotes the final equations used to calculate emission costs under both free-flow and stop-go traffic conditions. The last two Equations (9) & (10) denote the applicability of γ and β based on respective average link speed.

Noise on roads refers to unwanted sounds and vibration from motor vehicles. Noise cost is a preliminary function of vehicle type, the number of vehicles, traffic flow condition (dense or thin), time of day and surrounding population density. To determine the traffic flow condition, the volume/capacity has been used in this study. For simplicity purposes, surrounding conditions are considered as urban only. The noise cost can be mathematically denoted as:

N C a m = f ( x a m , x a / C a , L t )         ∀ a ∈ A ∪ A ¯ ,     m ∈ M (11)

where,

L t —time of day, t = 1 for day and t = 2 for night.

The total noise cost can be calculated as:

∑ a ∑ m N C a m x a m d a         ∀ a ∈ A ,     m ∈ M (12)

Equation (11) denotes the noise cost as a function of vehicle type, number of vehicles, v/c ratio and time of the day. Equation (12) represents the total noise cost calculated for given conditions. The coefficients required to calculate noise costs are given in Appendix D.

Vehicle operation cost estimates been published by authors elsewhere [87] and the following equation is used:

∑ a ∑ m ( x a m d a V C a m + x a m t a V T m )           ∀ a ∈ A ,     m ∈ M (13)

Equation (13) denotes the vehicle operating costs formula that has been developed in previous studies. The VOC is not seeming to be so relevant here, but it is required when calculating appropriate toll charges when managing congestion and other traffic management related activities. Therefore, its presented in this paper for completeness.