Decarbonizing intraregional freight systems with a focus on modal shift

Despite the uncertainty across the studies in figure 1, we find that the range of road freight energy intensities varies between regions. This has to do with economic, geographic and infrastructure constraints that determine which vehicles are in use. For a detailed review on heavy-duty vehicle technology and utilization we refer the reader to [3, 36, 37].

The US National Academies of Science, Engineering and Medicine (NASEM) [36, 38] has performed several assessments regarding the medium and heavy-duty sector in the US. In its 2010 study, the Academies suggest that a series of vehicle technologies—ranging from improvements in aerodynamics to the use of low rolling resistance tires—could reduce fuel consumption for US trucks by up to 15%. Improvements to diesel engines could reduce fuel consumption by up to 20%, and a shift to hybrid drive trains could halve it. The actual savings and the cost of achieving them would depend on the size and function of the vehicle. A recent report by the IEA [3] concludes that the energy intensity of trucks could be reduced by 34% by 2050, relative to a business-as-usual reference scenario.

Another strategy for decarbonizing road freight is by electrification using batteries, hydrogen fuel cells, or electrified roadways. However, due to the relatively low energy density of current battery technologies, electrification with batteries is only feasible for freight vehicles that have short ranges [3]. Indeed, Sripad and Viswanathan [39] find that the cost of a lithium-ion battery pack that gives a large truck (i.e. with a payload of up to 16 tons) a range of 300 miles on a single charge would cost more than US$160 000. By comparison, a diesel-powered semi-truck costs US$120 000. Guttenberg et al find that, if seven trucks—each one eight feet behind the other—were to travel in a platoon, the range would be extended to 800 miles on a single charge [40].

Electrified roadways include overhead catenary lines, tracks in the road, and inductive transfer of power, which can be combined with other electric powertrain technologies [3]. All systems require additional receiving technology in the vehicle. In partnership with Siemens and several automobile companies, Germany, Sweden and the US are currently testing this technology [41], and a number of other projects are underway in Germany and Sweden [42, 43]. This strategy would only likely be applicable in short corridors with high traffic density and requires considerable investments. The electrification of a one-mile stretch of a road near the Port of Los Angeles cost US$13 million and costs for electrifying a Swedish highway are estimated to around one million euro per kilometer [44].

In the United States and Europe, the transport work done by a truck falls substantially after four to six years [45]. However, old trucks remain in service, and may even enter heavy service in secondary markets if they are sold abroad. As such, in the absence of a policy intervention, new technology may diffuse across the fleet much more slowly than it is introduced.

There are also several important barriers to the adoption of efficient technologies, such as many users wanting a payback of less than two years, lack of access to high upfront capital, lack of information; and doubt regarding the reliability and safety of new technologies [3, 45, 46]. For example, the uncertainty in exploring new technology options has been a market barrier for small fleet shippers (fleet size <20) [46], which make up over 97% of all US trucking fleets [47].

Much of the world's rail system is already electrified [5] and thus its potential for decarbonization is closely tied to the decarbonization of the electricity sector. A big exception is the US, where diesel-electric locomotives are used on virtually all freight routes. Hoffrichter et al estimate that the well-to-wheel emissions from the diesel electric locomotives currently in use in the US are 1.02 kg CO₂ per kWh of tractive energy delivered, while emissions from electrified locomotives powered by the US grid in 2008 would have been 0.9 kg per kWh [48]. The current US grid is around 20% cleaner than in 2008 [49], and the benefit of a shift to electric traction would be correspondingly greater. However, in countries like the US with very long haul routes, the economics of electrification are not attractive as long as there are no limits on GHG emissions [50].

The Norwegian research organization SINTEF has conducted an assessment of hydrogen and battery-operated alternatives to conventional electrification and concluded that hydrogen or hydrogen-hybrid propulsion would provide a factor of two lower cost solution than installing catenaries on the remaining un-electrified lines in Norway [51]. While somewhat more expensive, in this case, large battery storage units (with reroute replacement/recharge) also dominated electrification in that application.

While the average carbon intensity for water shipments is lower than that for road shipments, small vessels carrying less than 250 tons of non-bulk cargo have GHG intensities that are similar to, or higher than, that of trucks [52, 5]. Much of the developing world relies on small vessels for freight transportation [53, 54].

Some zero-cost operational measures may be able to reduce fuel use by up to 33% [55]. Geertsma et al have reviewed a number of mechanical-electric hybrid propulsion technologies, and conclude that they could reduce fuel use by 10%–35% [56]. Vessels that use electrochemical storage in the form of electric batteries have been used for operations with limited ranges [57, 58]. Belgium and the Netherlands plan to bring fully electric and autonomous container barges into operation in 2018 [59].

Air freight only handles a small portion of the total freight activity in tonne-km, focusing on high-value and time-sensitive goods [12]. Strategies to reduce emissions from air freight overlap with those for passenger air transport, since a large part of air freight is transported by passenger aircraft [60, 61] and most freight aircraft are minor variants of passenger aircraft.

Schäfer et al [4] find that, for narrow body aircraft, cost-effective measures can reduce emissions by about 20%. Several of these measures such as winglets are already being deployed. Since the dawn of the jet age, aircraft engines have reduced specific fuel consumption by 70%. A recent NASEM study concluded that the overall efficiency of gas turbine engines could be increased by up to 30% relative to the best engines in service today [62]. However, further efficiency gains will likely require a significant change in aircraft configuration [62, 63]. Both US and European agencies aim to produce practical designs that reduce fuel burn by 70% compared to current aircraft by mid-century. Few of the designs developed in response to these goals are likely to achieve them, even after assuming significant breakthroughs in technology, regulatory posture, and customer preferences. Concepts include combinations of laminar flow, open rotors, unswept wings, wing-body hybrids, aeroelastic control surfaces, lower cruise speeds, and hybrid turbo-electric propulsion [64]. It is unlikely that industry will be able to produce motors and generators with sufficient power density (measured in kilowatts per kg) to propel large electric transport aircraft for at least the next two decades, nor is it likely that batteries with sufficient energy density (in kilowatt-hours per kg) will be available [62]. Given the extraordinarily high levels of safety demanded by the flying public, industry and regulators are cautious about introducing new technologies [65, 66].

Low-weight and high-value goods are frequently shipped by road or air, while heavy, low-value bulk goods, such as coal, grain and gravel, are often transported by rail or water [12, 102]. The decisions on shipment modes are influenced by the limit that a country and operators impose on truck and rail car weights. Shipment volume is constrained by vehicle or container size, and many shipments cube out before they weigh out [3, 12].

Rail dominates long-distance freight, which is for example defined as more than 500 miles in the US [12]. Shifting freight from short-haul trucking to rail is difficult because local rail lines are not in place and, with the exception of some heavy industries and particular bulk commodities such as coal, transfer centers or warehouses are mostly not designed for railcar-size deliveries [12].

Some commodity groups, such as live animals or chemicals, require specialized equipment. Shelf life, temperature and humidity requirements, sensitivity to acceleration forces and the risk of accidents can limit the modal choice [12, 104]. Protection of the shipment may lead to shippers only trusting a few carriers to transport high-value goods such as luxury cars [12, 102].

Transit time and reliability are critical components of mode choice [12, 103, 105]. Reliability determines whether requirements such as just-in-time, quick response, port deadlines, and hub-and-spoke operations can be met [103]. It is affected by congestion, maintenance, accidents, natural disasters, extreme weather, and mismanagement [12, 105]. Shippers consider trucks the most reliable mode, because of, for example the ability to avoid congestion by taking alternative routes [12]. For just-in-time operations, high-value or high-demand goods, the shipment frequency is also important, as it affects inventory costs [12], and shippers place similar importance on service frequency as on reliability [106, 107].

Some studies argue that shipping costs are less important for mode choice than quality of service attributes [108], while others suggest that shipping costs, especially when including inventory cost or financial risks, are decisive [12, 107]. Clearly the factors discussed above are not independent: for example in-transit carrying costs vary by shipment time, or inventory costs are affected by frequency and reliability.

De Jong et al [26] review elasticities and cross-elasticities in the freight sector. They find that transport price changes are more likely to drive changes in modal shares than in absolute freight transport demand. Furthermore, they find considerable variability in the elasticities and cross-elasticities, which are highly depend on shipment characteristics described above and on the market condition of the modes. Similarly, Christidis and Leduc [109] have reviewed a number of studies and found cross elasticities of rail and road that range from 0.3 to 2. This means that a 10% increase in the cost of road freight will increase rail freight by 3%–20%. Other authors present cross-elasticity estimates as large as 3 [110]. For intermodal transport, Arencibia et al find cross-elasticities of greater than 2 with respect to costs, which are smaller than the ones they find with respect to transit time [107]. Elasticities often cannot be transferred from one country to another [26], and those values likely do not hold for low-income countries.

Although not a primary concern for most shippers [111], there is increasing pressure on firms to reduce the environmental footprint, including CO₂ emissions, of their logistics supply chains. Programs such as SmartWay by the US Environmental Protection Agency and other international green freight programs [112] aim to make environmental performance more transparent and to ease the integration of environmental impacts into firms' decision-making [46].

Rail intermodal shipments have a significant disadvantage compared to trucks when speed and predictable time of delivery are priorities [104, 113, 114]. For example, freight train delays due to extreme weather can induce a shift back to trucks [115]. Winebrake et al [87] compared costs, transit time and GHG emissions of route options, and the lowest-emissions choice, predominantly rail, was sometimes cost competitive but always had a longer shipping time than other mode options. However, there are also cases where rail intermodal is faster than the conventional lowest cost-option, as for example the Europe-Asia road-rail link that is faster than ship [104]. To speed up intermodal, significant investments in freight capacity, modal connectivity, and efficiency of operation are required [35, 94, 95, 116]. Intermodal would also be more attractive if clients would adjust their strategic and operational decisions to move away from practices like just-in-time manufacturing [35, 117]. In just-in-time production, which originated in Japan, outside suppliers are required to deliver their parts in smaller batches in time for immediate use, thus reducing the need for warehousing [21]. It requires frequent, fast and reliable delivery times, that also resulted in a disadvantage for modes such as rail and inland water [35]. New regulation that incentivizes collaborative logistics, higher inventories and larger shipments could slow down supply chains and enable modal shift [113].

The external costs of road freight transportation are not adequately represented in transportation costs, and a stronger price signal could promote modal shift. One problem is the lack of granular data, which prevents the accurate assessment of environmental performance and obstructs policy analysis and decision making. A fragmented and informal road freight industry, in particular in the developing world, is hard to regulate [118], and competition makes carriers reluctant to share proprietary information.

Infrastructure investments in new or dual rail tracks, waterways, intermodal terminals, and inland ports can increase rail and water freight capacity, enhance modal connectivity, consolidate loads, improve quality of service and extend the rail network to new locations. However, such investments are often lacking. The ITF found that in 2014 the average public and private investment in inland transport infrastructure in OECD countries was 0.75% of GDP. The rail share of this investment was less than 30% [97]. As emerging economies and developing countries are experiencing rapid growth in transportation demand, there is a pronounced need for new transportation infrastructure. The IEA states that by investing in a low-carbon infrastructure and inducing modal shift, these countries could meet demand, boost their economies, and reduce costs [94]. In Africa, for example, the little existing railway infrastructure has been poorly maintained, resulting in a drastic reverse modal shift from rail to road [119, 120]. Operational, financial and contractual standards for intermodal terminals, such as those that exist for ports in the World Bank Port Reform Toolkit, are necessary [121].

In industrialized countries, there is a particular need to develop intermodal terminals and dry ports [122–124]. Dry ports are inland terminals, close to demand, which receive goods directly from the seaport, preferably by rail or inland waterways [123]. They allow containers to be handled close to demand centers (the concept of extended gateway) and increase the share of low carbon modes in the goods' journey [125, 126]. Large cargo-handling facilities may set themselves modal split goals that have an impact on port infrastructure designs [96, 127]. For example, the port of Rotterdam aims to ship 65% of incoming sea freight into the European hinterland by rail or inland waterways by 2035 [127]. Besides environmental considerations, those targets may also be in place to increase capacity or expand the port's reach by making longer-distance shipments available [127].

Operations research, which is essential for efficient infrastructure design, supply chain logistics and terminal operations, can reduce the carbon intensity of intermodal transport [128] and foster modal shift [129]. As intermodal transport research has emerged as an independent field, a number of studies have surveyed the state of the literature [130, 131, 132]. Although a number of freight transport models have been developed [133, 134], and tactical and operational issues have been studied, the lack of realistic physical topologies handicaps modeling freight flows. For example, SteadieSeifi et al [85] note that most analyses model hub-and-spoke freight networks, while many goods are in fact transported along dedicated corridors. In addition, most models optimize for lowest cost: there is a need for models that include other (e.g. environmental) objectives [85, 134].

Besides the vertical integration of dry ports, Logistics Service Providers (LSP) can facilitate co-ordination between different shippers to reduce cost, travel time and GHG emissions horizontally [82, 135, 136]. The EU emphasizes efficiency gains through this cooperative approach and the provision and exchange of reliable information on sustainability metrics [137, 138]. LSPs may allow shippers to choose the desired cost, speed, and level of environmental impact of a shipment, and devise a combination of modes and carriers to meet the shipper's requirements [139, 140]. LSPs can also offer more frequent services to all customers.

One of the key notions regarding the integration of services between modes is synchromodality, which refers to a situation where modal combinations and operational schedules can be changed after the shipment is on the way, in response to new information [85, 136, 137, 141, 142]. This ensures reliability and prevents data lock-in with a carrier. Pfoser et al [140] note that successful synchromodal transportation needs close cooperation between all stakeholders of the logistics chain, a willingness to change existing practices, and harmonized transport regulations, including for data sharing. Contract structures must be developed, which provide legal security for all market participants for the liability for delay, loss or damage. They do not see the development of infrastructure and information technology as significant barriers, although more research and modeling work is needed [140, 142].

Consolidation of smaller shipments that do not fill out a container or truck is offered through the Less-Than-Truckload (LTL) market, which provides door-to-door services where customers can select the transit times and levels of reliability [128].

Information and communication technology (ICT) can be used to improve and automate terminal operations, track shipment locations, improve security and quality control, administer data and aid with routing decisions when optimizing and pricing intermodal shipments [143]. ICT can also be used to automatically assess or collect tolls or constrain heavy vehicles to certain corridors [3, 144]. All of these functions could contribute to decarbonizing the sector [143, 145].

Terminal and port ICTs can address everything from customs information to fully automated freight shuttles between ports and dry ports [143]. Tracking devises include Radio Frequency Identification systems, GPS systems, and fieldbus communications networks, the last of which allows the distributed control of widely dispersed sensors [146]. These technologies can track freight trains, railcars [147], or individual intermodal containers [148]. ICTs can also facilitate the use of intermodal for the market of perishable goods by allowing continuous temperature monitoring in refrigerated containers [148]. Limited GPS coverage, signal blockage, dependence on batteries, and maintenance are problems, but technology is evolving fast [148]. Furthermore, the lack of standards for technologies and for data exchange retard the adoption of these technologies [146].

Consolidation in logistics, as for example in the LTL sector, is a complex optimization problem, which benefits greatly from ICT [146, 149]. ICTs could also provide firms involved in logistics—some of whom might be mutual competitors—secure and anonymous platforms to coordinate activities. For example, an electronic logistics marketplace can permit different companies to anonymously pool their shipping requirements without sharing confidential information or entrusting it to a third party [143, 150]. Shipping companies and banks are beginning to explore the blockchain technology for transactions with customers and tracking ocean and road shipments [151–153].

ICT is also essential for improving the cross-border compatibility of rail systems. Since the 1990s, the EU has worked with the private sector to create a standardized rail communication and signaling system with investments of over €770 million [154]. The European Railway Traffic Management System (ERTMS) is one of the most advanced train control systems [147] and it is employed also in many countries outside Europe [155]. Similarly, the EU fostered the development of an ICT system for inland waterways, the River Information System [156]. In the future, these intelligent transport systems could be combined into an efficient multimodal digital freight system [156, 143]. Many Asian nations have also recently introduced ICT systems to make border crossings faster [157].

While the ICT strategies described above would yield some carbon reduction potential, we are not aware of a systematic assessment of what that potential could be. This could be an important contribution to the literature that could support policy decision making for climate mitigation in the freight sector.

Third-party rail track access can be publicly regulated. The EU requires infrastructure managers to give access to third-party railway operators, while this is done voluntarily in the US [158]. The lack of track access may make hinder growth in rail traffic by allowing incumbent railroad operators (who in the US also own the tracks) to block new entrants by charging high track access charges [159].

Governments may choose to directly subsidize modal shift because it supports the goal of decarbonization. The Marco Polo Programme of the EU [160] compensated projects with €1–2 per 500 tkm shifted (the equivalent of €25–50 per tCO₂ emissions avoided). Although an early assessment has found that the program has underachieved targets for modal shift [161, 162], studies suggest that direct subsidies can be successful [129]. Belgium subsidizes the transfer of containers from sea ports to inland waterways [163].

In transportation, external costs include air pollution, noise, accidents, congestion, infrastructure damage and climate change. While this paper focuses on climate change costs, policies often address multiple, region-specific externalities. If revenues from Pigouvian taxes on road transport are applied to improve other modes, policies can be more effective in promoting modal shift and might gain more public support [164].

Assuming a carbon price of €90 per tonne CO₂ and a discount rate of 3%, the average marginal climate change costs for heavy freight vehicles are estimated at 2.5 to 10.4 euro cents per vehicle-km [165], which is around 0.3 to 0.4 euro cent per tonne-km but can also be much higher depending on load factors and emission standards. Diesel-powered rail is estimated to cost 0.26 euro cent per tonne-km for a load of 500t, and electrified rail is likely lower. These values are larger than the estimates of accident costs but substantially smaller than congestion costs. While a Belgian study showed that internalizing all those external costs can induce modal shift to lower-carbon freight modes [166], very high taxes might be needed induce modal shift [12]. This is also demonstrated by the anemic response to the Marco Polo program discussed above. Combining taxes with other decarbonization policies ('policy packaging') might be more effective [164].

External costs of rail freight and inland waterways are typically lower than for road transportation, but the case of intermodal transportation is less clear. Mostert and Limbourg (2016) [167] find that internalizing external costs can make intermodal shipments more cost-competitive but highlight that this effect diminishes for longer drayage distances. Similarly, Santos et al [129] show that internalizing external costs can also hurt intermodal transport competitiveness, depending on the length of the road haul.

Van Essen et al [164] find that fuel and vehicle taxes, vehicle charges and in some cases emission trading are the most common policies to internalize the external costs of transport. They recommend a strong differentiation for charges and taxes and a clearly labeled CO₂ tax or emission trading for climate change costs. Emission trading schemes such as the EU Emissions Trading System still do not include the transportation sector, with the exception of aviation [168]. We focus on fuel taxes and vehicle charges the following sections.

Harding [169] finds that in almost all OECD countries, diesel fuel taxes are lower than gasoline taxes. She concludes that the externalities associated with the fuels do not justify this tax differential and advises its gradual removal. In particular countries with large distances between cities, or few alternatives to road freight transportation tend to have lower fuel taxes [3]. While some countries still have diesel fuel subsidies in place, a few recently moved to eliminate them [170].

While much scholarship evaluates the effect of fuel prices on the demand for transport and fuel, there is less empirical work on the effect of fuel taxes. An analysis that sought to delineate the consumer response to gasoline taxes from the response to gasoline prices [171] found that the tax-exclusive price elasticity of demand for gasoline was −0.03, only half as large as the tax elasticity of −0.069. This may be because taxes are publicly debated and are therefore more salient or because consumers see tax rises as more permanent than transient price rises. In fact, the tax elasticity is lower in states where the taxes change frequently [171]. We have not encountered studies that separate the effects of prices and taxes for the trucking sector, and an empirical study would be instructive. One could speculate that the contractual arrangements that allow carriers to pass on increases in fuel prices [24] were designed to protect carriers' slim margins from market volatility. That does not apply to predictable taxes. However, some studies suggest that tax increases are rapidly and fully reflected in shipping prices [172]. The majority of the studies focus on the United States, Canada, and Europe, and empirical evidence from large developing countries remains scarce.

As of 2017, most EU countries have implemented tolls for heavy-duty vehicles [173], and this has proved to be effective in inducing a shift to rail or water in Switzerland, Germany and Austria [174]. The US Department of Energy concluded that direct user charges in the form of tolls for trucks could have the largest impact and it is the policy most likely to be implemented in the US [12]. Apart from aiding modal shift, road user charges may generate revenue for infrastructure maintenance, reduce congestion, increase logistics efficiency, and charge foreign-registered vehicles for the use of national infrastructure and the externalities they impose [144].

Technology can make road pricing more efficient and increase compliance rates. Germany has implemented, and other countries are studying, a satellite-based system to collect heavy-vehicle tolls [144]. This system also includes charges on secondary roads that have previously been used to avoid tolls. Australia's satellite-based Intelligent Access Program monitors where, when and how large and heavy vehicles are operated, and could be used to restrict those vehicles to corridors where more sustainable alternatives such as rail or water are unavailable [3].

As labor costs make up a significant portion of trucking costs, reducing the maximum hours of service can induce a shift of 2%–3% of the US tonnage from road to rail [12]. The entry barrier to the road freight industry is very low in most developing countries, resulting in high competition between numerous single-driver companies with low prices and low labor standards [22, 175]. National minimum wage rules also influence which countries bear the largest share of road freight traffic in a region, as for example in Europe [176]. While labor rules have in the past played a significant role in port and terminal operations, this is no longer as important [21]. However, strikes at container terminals and poor human resource management can impact reliability and productivity [177].

Restrictions on truck sizes are in place in most countries, mainly to protect the road infrastructure and to address safety [3]. Road damage increases as the ∼2nd to ≥ 5th power of vehicle weight [178]. In some parts of the world, however, regulations are poorly enforced, resulting in high profit margins for truck operators due to overloading, speeding and lax vehicle requirements [179]. Tightening and enforcing regulations, by cracking down on corruption, would reduce the cost-competitiveness of the road freight sector and promote a shift to rail [22].

In many developed countries, the discussion instead focuses on allowing larger and heavier vehicles (LHVs), which are already permitted in Canada, Australia, Brazil and Scandinavia [3]. Increasing the maximum permissible truck size and weight to more than 70 tons can decrease the carbon intensity [110], but could make road freight cheaper and induce a shift away from rail and water [3, 12]. The effect on the relative share of road and rail freight of allowing LHVs varies by country, depending on the associated cost reductions and the fraction of rail freight that is containerized [110, 180]. Predictions of the modal shift potential of LHV policies are difficult, since modal cross-elasticity values are uncertain and many current studies were conducted by groups with an interest in promoting one or the other mode [109, 110].

A study in Sweden found that allowing longer vehicles capable of carrying two containers for intermodal drayage can reduce total intermodal transportation costs by 5%–10% [181]. There is a need for more research on role of LHVs for low-carbon intermodal freight transportation [180].

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Since 1995, China has pursued the development of a high-capacity freight network with intermodal corridors and hubs, connecting economic and industrial centers [157]. The Chinese government is the main investor in infrastructure, particularly in the railway sector, and project developments are fast [182]. China invested 5.4% of its GDP in inland freight infrastructure in 2015, the highest investment rate reported by the ITF. However, more than three quarters of this investment was in road infrastructure [187]. With the Belt and Road Initiative, China is also rapidly expanding its transportation infrastructure outside of its own borders [188]. The inefficiency of the inland freight system is one of the main challenges for China's logistics system and is in stark contrast to some of the largest and most modern container ports in the world located in China [175]. One of the main barriers to modal shift is modal connectivity and ports are mainly serviced by trucks [182, 186]. For the line-haul movement of containers, on certain corridors the use of rail is prioritized and other modes are restricted [157].

Blancas et al [186] propose reforming the regulatory environment to allow Chinese railways to take a more customer-focused approach, similar to the deregulation in the US. They believe that as new high-speed rail has freed up capacity, such developments become more likely.

Luo et al [183] suggest a regional approach where wealthier eastern regions could experiment with measures such as fuel taxes while western regions focus on infrastructure development. However, different fuel prices in different regions might encourage smuggling.

Jun and Bensman [175] recommend requiring the trucking industry to improve working conditions, reduce overloading, eliminate falsified paperwork, and retire overaged fleets. Well-designed incentives might induce investment in more efficient modern fleets. We expect that these efforts would both lower the carbon intensity and increase the costs of road transportation. The latter could induce a shift to rail- or waterways.

Intermodal policies have been part of the North American transport strategy to improve the freight system since the 1990s [15]. For example, a detailed analysis of potential policy approaches was undertaken in 1998 [199]. Policies have mostly targeted corridor development and infrastructure financing [160, 189, 200], as well as the development and employment of intelligent transport systems [15, 160]. The US invested about 0.6% of its GDP in transport infrastructure in 2015. Less than 10% of it was in the rail sector [187], one of the lowest shares among OECD countries. For historical reasons, the US Federal Government favored investments in highways over railroads [201]. As a consequence, road freight, which travels over publicly-built and maintained roadways, enjoys an implicit subsidy relative to rail transport, which is privately owned and where freight rates must reflect the private cost of owning, building, paying tax on, and maintaining the infrastructure [202].

For the Canadian government, reducing the carbon emissions from the transportation sector is key to reaching GHG emission targets [189], in particular as some provinces have high shares of renewable energy generation [203], which limits the scope for reducing emissions from the electricity sector.

North American freight rail companies are privately owned and vertically integrated as they own both infrastructure and railway operations [189, 192]. In the US, federal deregulation occurred in the 1980s and 90s. This removed constraints on rail companies such as maximum rate regulations, and allowed them to enter into confidential rates and services contracts and to abandon lines [102]. The US and Canada have significantly reduced the extent of their rail track network [94]. There is no track access regulation, and track sharing is voluntary [158, 192].

The US and Canada are among the countries with the lowest motor fuel taxes in the world, equivalent to a carbon price for diesel of US$2.50 or less per tonne of CO₂ [169]. In the US, diesel taxes are lower in states with a larger proportion of employment in the trucking industry [204]. Only four US states (Kentucky, New York, New Mexico and Oregon) have tolls specifically targeting heavy trucks. There are few reports of interest in extending tolls in the US [205] as well as Canada [206].

Canada allows long and heavy combination vehicles with two semi trailers [189]. In the US, heavy vehicles are largely restricted to areas near the US-Canada border to facilitate the smooth movement of goods between the two [102]. The US and Canada both have heavy-duty vehicle GHG emission standards [3, 189].

Although the share of rail freight is high in North America, the competition from the road sector has induced a shift from rail to road. North American investments and policy-making lag Europe, China, and India in efforts to promote a modal shift from road to low-carbon modes. There is a clear opportunity to use motor fuel taxes and road charges to provide such incentives. Heavy-vehicle tolls have been identified by the DOE as an impactful policy for modal shift [12].

Due to the presence of regional rail monopolies in North America, connectivity between different rail segments of long-distance rail shipments is a barrier to modal shift. More terminals that facilitate efficient transfers between different rail lines could accommodate fragmented ownership and relieve congestion in North America [193]. Upgrading intermodal terminals to facilitate the transfer of more commodity groups may also help to reduce GHG emissions [207].

The decline of coal and the surge in oil and natural gas drilling activity in North America are currently changing the rail business. For example, a single new well pad can require up to 40 railcar loads of equipment and raw materials [208]. In regions, where a pipeline infrastructure is not in place, liquefied natural gas is also transported by rail, as for example in the state of Pennsylvania [208]. At the same time, the volume of coal shipped has declined sharply as demand in the electricity sector has fallen [209]. The transition from coal to intermodal freight as the largest source of revenue [195] is changing geographical patterns of rail freight activity, impacting investment needs, and posing a business challenge for railway companies. This challenge has already been recognized by the Canadian government [189].

In a 2011 white paper on transportation [83], the EU set the goal of shifting 30% of long-distance road freight (over 300 km) to rail or water by 2030, and 50% by 2050. This is to be achieved by improving the rail network, modal connectivity, and the quality of service to make multimodal freight economically attractive. The white paper mentions several EU-funded information systems to improve logistics and traffic management. As discussed in section 4.3.4, EU-initiated projects are at the forefront of developing intelligent transport systems.

The EU Combined Transport Directive (Council Directive 92/106/EEC as of 1992) regulates intermodal transport of goods between EU member states and ensures for example that vehicle tax reductions for drayage vehicles are granted through member countries. The EU Commission recently proposed amendments to the Directive, reinforcing sustainability goals [84].

The Marco Polo Programme, which offered subsidies to modal shift projects, ran from 2003 to 2013 [162]. As of 2010, 30 billion tkm of freight have been shifted but the performance lagged behind expectations [211]. However, monitoring of the program's results has continued after 2010.

The EU is proactive in internalizing external costs of freight transportation [164, 165, 167]. European countries have higher diesel taxes than most other countries, equivalent to between 120 and 260 euros per tonne of CO₂ [169]. As of 2017, most EU countries have implemented some form of road charges for heavy vehicles [144, 173], which are harmonized by the Directive 1999/62/EC or Eurovignette Directive [212]. The tolls are intended both to recover infrastructure costs and to provide a means to induce modal shift [164]. Germany applies a heavy-vehicle toll to any truck with a gross weight of at least 7.5 tonnes using a GPS-based system [213]. Since 2005, this toll has been collected on all major highways (Bundesautobahnen) and since 2015 also on a number of secondary roads (Bundesstraßen), which were previously used to avoid tolls. The charge depends on the length of travel, the size of the vehicle and its emissions class, and it is between 8.1 and 21.8 euro cents per km, as of 2015. Belgium, Denmark, Luxembourg, the Netherlands and Sweden have a common truck tolling system, the Eurovignette system, which is based on time of use instead of distance [173]. Austria and Switzerland have particularly high road charges of up to US$0.50 per km [214, 215].

In addition, the EU is integrating modal shift strategies into the development of the Trans-European Transport Network of core corridors [159, 211, 216]. EU-level and national efforts have led to a large number of new intermodal terminals [122]. The European railway system is primarily geared towards passenger service and rail infrastructure is mostly publicly owned [191]. The EU mandates non-discriminatory access to railway tracks [158]. To create a Single European Railway Area (SERA), the EU has created a standardized ERTMS, replacing existing national automatic train protection systems [217]. The EU's Horizon 2020 research and innovation program funds the joint undertaking Shift2Rail to promote the development of rail technology and modal shift strategies [218]. Through the COMCIS project, the EU is identifying and addressing problems of missing collaboration between modes by standardizing data systems and sharing. This includes adopting such concepts as synchromodality and the use of dry ports [219, 220].

Modal shift policies on the national level have been less successful than on the EU level, with the exception of policies targeting intermodal ports [159]. Many European ports only permit sustainable port development and include modal share targets, sometimes even mandated, into their infrastructure planning and terminal concessions [127].

The EU has extensively promoted modal shift to decarbonize freight, but as the average share of road freight activity has increased, it is not apparent that these policies have been successful [82]. It is unclear whether the modal shift initiatives have been without effect or have potentially prevented a larger shift to road transportation. More research is needed in this area. The EU recently proposed further measures expanding information transparency and new investments in terminals [84].

To achieve its stated target of a rail share of 50% by 2031–32, the GoI aims to modernize the freight system and expand capacity [157, 221]. As much of the infrastructure is yet to be built, long-term planning that accounts for the need to minimize GHG emissions could avoid the construction of an inefficient and carbon-intensive system. The GoI currently invests around 4% of GDP in infrastructure but only a tenth of that in the rail sector [221]. The GoI has initiated large multimodal infrastructure projects [221] and is building Dedicated Freight Corridors (DFC) [221, 223, 231, 232] in an attempt to expand rail capacity. Some of these projects are built with international financial assistance [223, 225, 231, 232] and the use of public-private partnerships (PPPs) [233]. India has more than 50 dry ports for containerized transport [225]. Containerized freight is handled by Container Corporation of India, which is part of Indian Railways and has invested in modernizing its operations with ICT; for example to track containers [225].

In addition to the initiatives for improving rail and road networks, the GoI plans to subsidize water shipments [157] and to develop India's first modern inland waterway on the Ganges river, which flows through the densely populated Indo-Gangetic plain [234]. Approximately 40% of India's traded goods make their way through this region at some point in their journey [234], many of which are trucked to ports on the country's west coast. The World Bank has financed the project with a loan of $375 million [235].

Fuel taxes are collected by central and state governments with considerable variation between states [236]. Efforts to reduce air pollution and congestion by heavy-vehicle driving restrictions (e.g. a nationwide ban on trucks older than 15 years), fuel emission standards, and green taxes on vehicles that enter urban areas have been adopted in parts of India [236]. These may promote modal shift as well as help to decarbonize the last mile.

As level of service factors [227] and capacity bottlenecks [157] are the most important barriers to modal shift, the share of infrastructure investments in rail needs to be increased [221]. The United Nations Environment Programme estimates that the planned DFCs can avoid large amounts of GHG emissions, mainly by preventing a shift from rail to road in the growing freight sector. These effects could be increased by a carbon pricing strategy [223]. We found little evidence of the use of cost incentives to promote modal shift and adjusted price signals that are important for optimal infrastructure planning [237]. The GoI acknowledges that Indian subsidies, tariffs and taxation policies send distorted price signals [221]. Heavy vehicle tolls have been recommended as a policy to promote modal shift and could be collected electronically [221, 227].

The goods and services tax (GST), a national value-added tax introduced in 2017, eliminates state taxes in some sectors. The new regime is expected to lower the cost of freight, although it is unclear whether it will benefit one mode more than others [238]. Early reports suggest that by eliminating the need to stop at tax checkpoints every time a truck crossed state boundaries, GST has cut the travel time on road by about 25% [239], potentially making road more attractive. A retrospective analysis of the impact of GST on modal choices would be interesting once the tax regime has stabilized and taken root.

Coal is the commodity with the largest freight activity in India, and almost 80% of it is transported by rail [240], which is already contributing to congestion of the rail network. As demand for coal and for the transport of it increases, it may crowd out other goods if infrastructure does not expand rapidly enough [241], and may be increasingly transported by road [242].