The effects of road pricing on transportation and health equity: A scoping review

Abstract

Road pricing is increasingly considered as an option to support transportation infrastructure costs, manage demand, and reduce emissions. However, the extent to which implementation of such approaches may impact transportation and health equity is unclear. In this scoping review, we examine the differential transportation and health effects of road pricing policies across population groups and geographic region. We conducted a systematic database search of Transport Research International Documentation, GEOBASE, Scopus, and Ovid Medline, supplemented by bibliographic review and internet searches. Fifteen studies were included in the review. The studies evaluated area and cordon road pricing systems in Singapore, London, Stockholm, Milan, and Gothenburg, and had a median follow-up period of 12 months. Outcomes evaluated include car commuting, mode shift to public transit, accessibility to destinations, affordability, welfare, social interactions, air pollution, traffic injuries and deaths, acute asthma attacks, and life expectancy. While more studies across diverse urban contexts and policy settings will be needed to strengthen the evidence base, the existing evidence suggests road pricing has mostly net positive effects related to a reduction in car trips, air pollution, asthma attacks, and road traffic collisions, and increases in life expectancy. Frequency and ease of social interactions were found to be negatively impacted, with fewer visits to family and friends. The population groups that generally fared better across transportation and health outcomes were those with higher incomes, men, and people between the ages of 35-55. Across space, there are benefits for both the areas inside and outside the cordon boundary, but to a greater degree for the area inside. Overall, the evidence base is limited by a narrow set of health-related outcomes and a lack of longer-term studies. We did not come across any studies assessing distributional effects of noise pollution, mode shifts to walking or cycling, or other morbidities in the general population that are not listed above. In addition, there are few evaluations that include non-work trips, therefore potentially missing effects for unemployed populations or women who are more likely to make non-work trips. We find that the limited body of evidence on area and cordon pricing policies suggests these policies are beneficial for a number of transportation and health outcomes, particularly for populations inside the cordon area, but that there may be some degree of inequities in the distribution of the benefits and burdens.

Introduction

Many cities are looking to road pricing as a potential strategy to manage demand for the road, and to achieve transport and environmental goals. Road pricing is any charge applied to use the road, including congestion charges, distance-based fees, and highway and bridge tolls. In addition to reducing traffic congestion, these systems can raise revenue for transportation investments which may be directed towards road infrastructure or targeted towards other modes such as public transit, walking, and cycling infrastructure. Large-scale road pricing policies have been implemented in Singapore, London, Stockholm, Milan, and Gothenburg in the form of area and cordon-based road pricing systems. In such systems drivers have to pay to cross a boundary, often situated around the city centre. The road pricing policies in these cities have been successful in reducing traffic congestion (Börjesson & Kristoffersson, 2015; Croci, 2016; Eliasson et al., 2009).

Road pricing not only presents a promising strategy to address congestion, but may also result in population health impacts (Figure 1). The limited evidence to date suggests mostly positive health effects from road pricing overall, with the exception of social interactions. Road pricing in London, Stockholm, and Milan reduced emissions (CO₂, PM₁₀, and NO_x) by 13-18% in the priced area of the cities (Croci, 2016). As a result of reduced traffic pollutants, it is estimated that the implementation of London’s congestion charge translated to 183 years of life gained per 100,000 people over 10-years, and in Stockholm to 206 years of life gained per 100,000 people (Johansson et al., 2009; Tonne et al., 2008). In terms of road safety, studies in London and Milan found reductions in traffic incidents (Green et al., 2016; Li et al., 2012; Percoco, 2015), with no evidence of traffic incidents being pushed outside of the priced area (Green et al., 2016). The evidence on physical activity benefits is less clear. A 2015 review did not find conclusive evidence for increases in physical activity or mode shifts to walking and cycling, but also noted the quality of evidence to be low (Brown et al., 2015). One health determinant that appears to be negatively impacted is social interactions. In London, congestion pricing was found to reduce the number of visits to family and friends (Munford, 2017; Transport for London [TfL], 2004), which could lead to social isolation and negative health consequences downstream.

Figure 1. Conceptual framework of the pathways from a road pricing policy to health.

While the aggregate effects of road pricing on health may be positive, these effects may not be distributed equitably. Understanding how the health benefits and burdens are distributed is important for determining whether road pricing will reduce or exacerbate health inequities. For example, while road pricing may reduce air pollution inside the priced area, it could lead to increased pollution in surrounding areas. There are also concerns that people with lower incomes will be disproportionately burdened by tolls, especially those with limited public transit access (Eby et al., 2020). However, some argue that road pricing has the potential to be progressive in nature if the revenue is redistributed to improve public transit or other ways that benefit people with lower incomes (Eby et al., 2020; Eliasson & Mattsson, 2006). Another argument is that road pricing is more equitable than the status quo (i.e., “unpriced roads”), which essentially subsidizes travel for wealthier people (Manville & Goldman, 2018; Taylor, 2010). Roads are typically funded through gas taxes, vehicle registration fees, and income, sales, and property taxes. Aside from gas taxes and vehicle registration fees, all residents pay for the roads regardless of use. Yet lower income households use roads less, as they typically own fewer vehicles and travel less than higher income households (Blumenberg & Pierce, 2014; Manville & Goldman, 2018).

The equity of road pricing can be considered in terms of procedural and distributional equity (Levinson, 2010). Procedural equity refers to fair access to the planning and decisionmaking process. Distributional equity refers to how benefits and burdens are distributed across populations and space (Karner et al., 2020; Levinson, 2010), including transportation outcomes, such as costs, mode shifts, travel times, and accessibility; and health-related outcomes, such as social interactions, air pollution exposure, traffic incidents, morbidity, and mortality. We focus on distributional equity in this scoping review.

There are two previous narrative reviews on the equity of road pricing which summarize theoretical considerations and evidence prior to 2010 (Ecola & Light, 2009; Levinson, 2010). These reviews provided overviews of the ways equity could be considered and some discussion of how particular population groups could be affected, but did not focus on health considerations and the distributional impacts across population groups. Since these reviews were published, new road pricing policies have been introduced.

For these reasons, we believe it is timely to re-examine the evidence base, with a focus on updating reviews by: (1) examining the effects of road pricing on health and the factors along the pathway to health in addition to transportation outcomes, and (2) evaluating whether effects differ spatially or across population groups. We focus only on ex-post studies, as they are based on actual policy implementations and require fewer assumptions. Thus, our aim is to conduct a scoping review to summarize evidence on the equity implications of large-scale road pricing policies, with a focus on outcomes related to health and transportation. As the literature on health is limited, our scope includes transportation outcomes which ultimately have downstream effects on health. This review will provide practitioners and researchers with an overview of the effects of road pricing policies for different regions and population groups, and identify areas for further research.

Conceptual framework

The conceptual framework guiding this review was adapted from those developed by the San Francisco Department of Public Health [SFDPH] (2011) and Nieuwenhuijsen (2016) (Figure 1). SFDPH provides a comprehensive framework of the potential pathways from a road pricing policy to health. Nieuwenhuijsen’s (2016) framework more broadly links urban and transport planning to health, and highlights the importance of context and socio-demographic characteristics in modifying the relationship.

In our framework, road pricing affects transportation outcomes, which are associated with a range of health behaviours and environmental exposures, and in turn are associated with health outcomes. We are interested in both health determinants and outcomes in this review. Health determinants are factors that influence outcomes (e.g., air quality, noise pollution), while health outcomes are endpoints related to specific morbidities or mortality (e.g., asthma, injury). The effects from road pricing can be both positive and negative. For example, reductions in congestion can improve air quality, and fewer cars on the road can improve air and noise pollution. On the other hand, less congestion may increase travel speeds, which could lead to more severe traffic incidents, and an additional charge to drive may deter travel for certain populations.

Importantly, the aim of this review is to understand how different population groups and regions are affected by road pricing. People of different incomes, gender, age, and race, and intersections between these identities, are likely to be impacted differently. Across income groups, people with lower incomes who drive may be more likely to switch modes as a result of an additional charge, whereas people with higher incomes may not be as affected by an additional charge. On the other hand, as people with lower incomes and racialized populations are more likely to live along high traffic roads (Rowangould, 2013), a reduction in driving could result in greater improvements in air quality for these populations. In terms of gender, women are more reliant on public transit for transportation (Ng & Acker, 2018). If revenues from road pricing are redistributed to improve public transit, women may benefit more from reduced travel time. Younger and middle-aged populations may be more affected by road pricing as they make more trips relative to older adults. However, there could be negative effects for older adults if an additional charge deters family or friends from making less visits or caregiving trips. Finally, there could be differential effects by race and ethnicity if there are pre-existing differences in travel patterns or differences in where racial and ethnic groups tend to live and work. In Canada and the US, there is a long history of racial segregation and racialized populations on average face higher exposure to air pollution (Bauder & Sharpe, 2002; Clark et al., 2017; Giang & Castellani, 2020; Pinault et al., 2016; Tessum et al., 2019). Geography plays an important role: where people live in relation to the priced area and their regular destinations will have great bearing on how they are impacted by the policy. This is particularly relevant in cordon and area-based pricing policies, where impacts may be quite different for those who live within versus outside the boundary (Levinson, 2010). People who cannot afford to live inside the priced area now have to pay a fee to drive into this area and may feel constrained into choosing a less efficient mode of transportation.

Methods

Scoping review framework

We used a scoping review methodology. This approach was appropriate as the literature on road pricing spans across many disciplines and there are limited studies on the distributional effects of road pricing policies. We followed a pre-defined 5-stage process: 1) identifying the research question; 2) identifying relevant studies; 3) selecting studies; 4) charting data; and 5) reporting results (Arksey & O’Malley, 2005).

Key definitions and positioning of research

Road pricing is any charge applied to use the road. Pricing systems focused on reducing congestion often have time-differentiated charges with higher charges at peak travel times. There are three main types of road pricing: facility-based (High-occupancy toll (HOT) lanes,bridge tolls), area-based (cordon and area pricing), and network-wide (Levinson, 2010). HOT lanes are traffic lanes that can be used for free by high occupancy vehicles, and for a fee by single-occupancy vehicles. Cordon and area pricing systems impose a charge to enter into a defined area, often around the city centre. The difference between the two is that cordon systems charge drivers per crossing, whereas area schemes charge a flat daily fee to drive in the priced area. Network-wide systems can be implemented as distance- or time-based charges, or congestion-point charges on major links throughout a region. Although distance-based charges are being considered in a number of cities, the only example to date is Germany’s distance-based toll for freight traffic. In this review, we focused on the equity implications of large-scale road pricing systems that are area-based or network-wide. Further, we limited the scope to include road pricing policies that are aimed at reducing congestion. This excludes policies solely aimed at reducing pollution such as Low Emission Zones (LEZs), which restrict access to polluting vehicles. The majority of LEZs only apply to larger vehicles, such as trucks and buses, and will therefore have different effects than congestion-focused policies that apply to a wider range of road users (Urban Access Regulations, 2020).

Stage 1: Identifying the research question

This scoping review aimed to provide an overview of the literature on the effects of large-scale road pricing on transportation and health equity. To address this aim, we identified three specific research questions:

• (1)What methods are used to evaluate the distributional effects of road pricing policies on transportation and health?
• (2)How are the transportation and health effects of road pricing policies distributed across population groups and geographic regions?
• (3)What are the research gaps?

Stage 2: Identifying relevant studies

From March to May 2020, KH conducted searches in four databases: Transport Research International Documentation (TRID), GEOBASE, Scopus, and Ovid MEDLINE. In consultation with a university librarian, these databases were selected to cover articles in the transportation, geography, and health disciplines. We adapted the search strategy developed by Hosking et al. (2019) that was specifically designed for investigating whether the effects of transportation interventions differ by socioeconomic status and ethnicity. The search used keyword combinations of terms for “road pricing”, “transportation and health”, and “equity” (Appendix A.1). Searches were limited to publications between 2000 and 2020 in the English language. We searched references of relevant articles and google scholar to identify additional articles.

Stage 3: Study selection

All studies were uploaded to Rayyan (Ouzzani et al., 2016), where duplicates were automatically identified. Our inclusion and exclusion criteria were adapted from Hosking et al. (2019). For inclusion, studies had to 1) conduct an ex-post analysis of a road pricing policy that had already been implemented, 2) assess the impact of the policy on a transportation outcome, health determinant, or health outcome, and 3) present data on differing effects across socio-demographic or geographic strata (Appendix A.2). For impacts across geographic region, we were interested in studies that assessed impacts for the area inside the cordon compared to areas just outside the cordon or the outer areas of the city. KH conducted title and abstract screening. Abstracts that were identified as potentially relevant were double screened by KH and CF. Full-text reviewing was carried out by KH and CF with conflicts resolved by MW.

Stage 4: Charting the data

The data were charted by KH in MS Excel. Where available, this included: 1) study characteristics, 2) intervention context, 3) population group/geographic region, and 4) key findings. To assess whether studies assessed shorter- or longer-term impacts, we recorded the timeline of when outcomes were evaluated in relation to when the policy was implemented

Context

The equity of road pricing policies is no doubt influenced by the intervention context. To describe the broader intervention context of road pricing policies considered in our review, we considered context across five domains: (1) geographical context; (2) exemptions and discounts; (3) other major transportation interventions implemented alongside road pricing; (4) how revenues are spent; and (5) overall transportation and health-related impacts. We first retrieved contextual information available directly in studies, and then used internet searches to fill in missing information.

Stage 5: Collating, summarizing, and reporting results

Given the diversity of study designs and outcomes included, it was not possible to provide a quantitative synthesis of results. Instead, we provide a narrative summary of the evidence base. Key findings from each study, organized by population group and geographic region are provided in Appendix B. Some studies evaluated more than one outcome. In these cases, we included results for each outcome that was within our conceptual model (Figure 1) and was stratified along socio-demographic or geographic strata. There were limited findings across specific transportation and health outcomes, therefore, we synthesized evidence across population groups and geographic region.

Results

Search results

The search produced 2,316 articles and we identified 8 additional articles through other sources. Following removal of duplicates and screening, 15 articles were included (Figure 2). The main reasons for exclusion were because articles did not focus on road pricing, had a methodological focus, or were not empirical in nature.

Figure 2. Flow chart for study selection.

Study characteristics

The 15 eligible studies were published between 2004 and 2020 (Table 1). Thirteen were peer-reviewed studies and two were reports prepared by Transport for London as part of their congestion pricing monitoring program. The studies evaluated road pricing systems in London (n=7), Stockholm (n=4), Gothenburg (n=3), Singapore (n=1), and Milan (n=1). Only half stated an intention to investigate impacts along socio-demographic or geographic strata in their objective. All studies evaluated impacts in adult populations, except for Simeonova et al. (2019)’s study on acute asthma attacks in young children. Seven evaluated a transportation outcome, including impacts of road pricing on commuting by car, mode shift to public transit, accessibility, affordability, and welfare (quantified in terms of transportation costs and benefits). Four studies evaluated a health determinant, including social interactions and air pollution concentrations. Five studies evaluated a health outcome; three on traffic injuries and deaths (measured), one on asthma attacks (measured), and one on mortality (modelled). Nine studies assessed impacts by income or socio-economic status, five by gender, four by age, and seven by geography.

Table 1. Characteristics of studies included in the review (n=15).

Author (Year); Location	Study objective	Methods (Time horizon for intervention effect)	Relevant Outcomes	Socio-demographic/geographic strata
Agarwal & Koo (2016); Singapore	To analyze the effect of congestion toll rate adjustment on the change of commuters’ transport modal choice in Singapore.	Pretest-posttest; Ecological; Difference-in-differences (2 months)	Mode shift to public transit	Income
Transport for London (2004); London, UK	To evaluate the social impacts of the London congestion charge.	Pretest-posttest; Descriptive statistics (1 year)	Car trips; Travel time; Accessibility; Affordability; Social interactions	Income, Gender; Age; Geographic
Beevers & Carslaw (2005); London, UK	To estimate the impact of congestion charging on vehicle emissions in London.	Pretest-posttest; Emissions modelling (1 year)	Air pollution	Geographic
Transport for London (2008); London, UK	To describe the impacts of congestion charging in and around central London, with a focus on the impacts of the Western Extension.	Pretest-posttest; Descriptive statistics (1 year after Western extension)	Car trips; Affordability	Income
Tonne et al. (2008); London, UK	To model the impacts of the London Congestion Charging Scheme on levels of traffic-pollutants, life expectancy, and socioeconomic-inequalities.	Pretest-posttest; Ecological; Emissions modelling; Life-table analysis (<1 year)	Air pollution; Mortality (years of life gained)	Area-level deprivation; Geographic
Noland et al. (2008); London, UK	To investigate the effect of the London congestion charge on traffic casualties for all motorists, pedestrians, cyclists, and motorcyclists, both within the charging zone and in areas of London outside the zone.	Pretest-posttest; Intervention time-series analysis (22 months)	Traffic incidents	Geographic
Green et al. (2016); London, UK	To examine monthly traffic incident counts in central London before and after the congestion charge.	Pretest-posttest; Difference-in-differences (6 years)	Traffic incidents	Geographic
Munford (2017); London, UK (Western extension)	To examine the impact of the London Congestion Charge on social capital.	Pretest-posttest; Difference-indifferences (9 months)	Social interactions	Geographic
Karlstrom & Franklin (2009); Stockholm, Sweden (Trial)	To assess the equity effects of the Stockholm trial on mode choice and departure time.	Pretest-posttest; Propensity score matching; Welfare distribution analysis (2 months)	Mode shift to public transit; Welfarea	Income; Gender
Franklin (2012); Stockholm, Sweden (Trial)	To investigate the effect of demographic factors and travel responses to congestion pricing and the indirect role that context might play in mediating how different types of households adjust their behaviours.	Pretest-posttest; Structural equation modelling (2 months)	Car trips	Income; Gender; Age
Simeonova et al. (2019); Stockholm, Sweden	To examine the effects of a congestion tax in Central Stockholm on ambient air pollution and health of local children aged 0 to 5 years.	Pretest-posttest; Difference-indifferences (3.5 years)	Asthma	Age (0-5 years)
Percoco (2016); Milan, Italy	To evaluate the effects of a road pricing scheme in Milan on accidents.	Pretest-posttest; Interrupted time-series (3 years)	Traffic incidents	Geographic
Andersson & Nassen (2016); Gothenburg, Sweden	To analyse the effects and perceptions of the Gothenburg congestion charge on commuting habits, attitudes, and satisfaction with travel.	Pretest-posttest; Logistic Regression (11 months)	Car trips	Income; Gender
West & Borjesson (2020); Gothenburg, Sweden	To explore how the costs and benefits of the Gothenburg congestion charge are distributed across different segments of the population.	Pretest-postest; Welfare distribution analysis (11 months)	Welfareb	Income; Gender; Age
Eliasson (2016); Stockholm and Gothenburg, Sweden	To explore the fairness of congestion pricing from the consumer and citizen perspective.	Posttest; Descriptive statistics (Stockholm: 5 years; Gothenburg: 11 months)	Cost	Income

^aWelfare effects quantified in monetary units, based on mode choice before and after the congestion pricing toll, amount of toll paid, and estimated travel times.

^bWelfare effects quantified in monetary units, based on the travel time per trip and the value of time, the frequency of charged trips, and access to a company car for private trips.

Methodological approaches

Data

Most studies used data collected before and after policy implementation, either through travel surveys or routinely collected administrative data. Pre-post travel surveys were conducted in London, Gothenburg, and Stockholm as part of larger evaluation programs. Several studies included in this review drew data from these pre-post travel surveys (Andersson & Nässén, 2016; Eliasson, 2016; Franklin, 2012; Karlström & Franklin, 2009; Munford, 2017; TfL, 2004; TfL, 2008; West & Börjesson, 2020). To quantify changes in air pollutants, Beevers & Carslaw (2005) and Tonne et al. (2008) used traffic flow and speed data. Studies assessing traffic incidents used local traffic incident data. The majority of studies assessed shorter-term impacts of road pricing policies; eleven of fifteen assessed impacts at 1 year or less.

Target population

Adults were the target population for most studies. Of the studies assessing behaviour change, many examined impacts more narrowly for car drivers either by focusing on car trips as the outcome (Andersson & Nässén, 2016; Franklin, 2012; TfL, 2008), considering the treated group to be those who drove across the congestion pricing boundary (Munford, 2017), or drawing a sample of only car owners (Andersson & Nässén, 2016; West & Börjesson, 2020).

Analytical approaches

Methods included simple descriptive statistics (n=3), difference-in-differences (n=4), time-series analysis (n=2), structural equation modelling (n=1), logistic regression (n=1), and emissions modelling (n=2). Two studies conducted welfare distribution analyses (n=2) by taking into account travel costs and benefits such as toll payments, travel time gains, travel mode shifts, and in the case of Gothenburg, access to company cars (Karlström & Franklin, 2009; West & Börjesson, 2020). In these studies, “welfare” is represented in monetary units, and is an aggregate measure of the calculated gains (e.g., decreases in travel time) and losses (e.g., cost).

Distributional effects of road pricing

To assess the distributional impacts of road pricing across socio-demographic or geographic strata, studies generally examined the effects separately for each population group or geographic region under consideration. The seven studies that assessed impacts by geography either split the region into two areas (area inside and outside the cordon) or three areas (cordoned area, area just outside the cordon, and outer zone).

Intervention context

A brief context for cities’ road pricing policies is provided in Table 2. Reducing traffic congestion was a primary objective for all. For the systems in Stockholm, Milan, and Gothenburg, reducing air pollution was also an objective. Singapore’s system is both a cordon-based and congestion-point system, with a cordoned area around the central business area, and charge-points located along expressways and arterial roads leading into the central area. London’s system is an area-based charge around the city centre. The Western extension, implemented in 2007, doubled the area but was later removed in 2011. Stockholm, Milan, and Gothenburg all have cordon-based systems implemented around the city centre. At the time of implementation, all cities increased bus service in anticipation of a mode shift to public transit. Other interventions included improvements to rapid transit and rail lines (Singapore, Stockholm), increased parking rates (Singapore), implementation of a park-and-ride (Stockholm), and new bus and bicycle lanes (London, Gothenburg). Overall, net transportation and health impacts from the congestion charge in each of the cities are mostly positive and in the expected direction.

Table 2. Intervention context for road pricing systems considered in this review.

	Singapore	London	Stockholm	Milan	Gothenburg
Date introduced	Electronic road pricing, September 1998; replaced Area Licensing Scheme, June 1975	February 2003; Western extension, February 2007-January 2011	January 2006 (trial), August 2007 (permanent)	Area C charge, January 2012; replaced Ecopass, January 2008	January 2013
Road pricing objectivesa	To reduce congestion and improve travel speeds.	To reduce traffic congestion in central London.	To reduce congestion, improve travel speeds, and reduce traffic pollutants.	To reduce congestion and air pollution.	To reduce congestion, improve the environment, and raise revenue for transportation infrastructure.
Design	Cordon pricing and congestion-point charges; charges vary by time and location; weekday travel	Area pricing; fixed charge; weekday and weekend travel	Cordon pricing; charges vary by time of day; weekday travel	Cordon pricing; fixed charge; weekday travel	Cordon pricing; charges vary by time of day; weekday travel
Geographical context	Priced area: 93 charge points around city centre and major expressways; Metropolitan population: 6 million; 67% of peak-period trips made with public transitc	Priced area: 21km2; Western Extension: 18km2; Metropolitan population: 9 million; 83% of trips into Central London made with public transit prior to implementation of the charged	Priced area: 36 km2; Metropolitan population: 2.2 million; 77% of commuting trips that cross the cordon were made with public transit prior to implementation of the chargee	Priced area: 8 km2; Metropolitan population: 3 million; 34% of trips that cross the cordon were made with public transit prior to implementation of the chargef	Priced area: 36 charge points around city centre; Metropolitan population: 1 million; 26% of commuting trips that cross the cordon were made with public transit prior to implementation of the chargee
Exemptions & discountsb	Exemptions for buses and emergency vehicles	90% discount for people living in the congestion zone; Exemptions for buses, emergency vehicles, drivers with disabilities, taxis, electric vehicles, and motorcycles	Exemptions for buses, emergency vehicles, people with disabilities, electric vehicles, and motorcycles	Exemptions for buses, emergency vehicles, people with disabilities, taxis, electric vehicles, and motorcycles	Exemptions for buses, emergency vehicles, people with disabilities, and motorcycles
Other interventions	Increased bus services, and expanded rapid transit and light rail; Increased downtown parking rates	Increased bus service; Installation of bus lanes; Low Emission Zone (2008); Ultra Low Emission Zone (2019)	Increased bus services and extra rail capacity; Implementation of a park-and-ride scheme	Increased bus service	Increased bus service, Implemented bus lanes and more bicycle lanes
Revenue distribution	Not specified.	Revenues re-invested in London’s transportation system, mainly public transport, pedestrian, and cycling infrastructure.	Revenues used for road improvements.	Revenues reinvested in public transport and bicycle sharing.	Revenue re-invested in transportation, including improvements to public transport, bus lanes, and a new road tunnel.
Transportation & health impacts	↓ Congestiong ↑ Mode shift to transith	↓ Congestioni ↓ Traffic volumei ↓ Social interactionsj ↓ Air pollutionl,k ↓ Traffic collisionsi ↑ Life expectancyk	↓ Traffic volumei ↑ Mode shift to transiti - Physical activitym ↓ Air pollutioni ↑ Life expectancyn	↓ Traffic volumei ↓ Air pollutioni ↓ Traffic collisionsi	↓ Traffic volumeo ↑ Mode shift to transito

↑ Increase;↓ Decrease; -- neutral; no change or null results

^a Van Amelsfort & Brundell-Freij (2018)

^b Urban access regulations (2020)

^c Land Transport Authority Singapore (2019)

^d Transport for London (2003)

^e West & Borjesson (2020)

^f Martino (2012)

^g Santos (2005)

^h Agarwal & Koo (2016)

ⁱ Croci (2016)

^j Munford (2017)

^k Tonne et al. (2008)

^l Karlstrom & Franklin (2009)

^m Bergman et al. (2010)

ⁿ Johansson et al. (2009)

^o Borjesson & Kristoffersson (2015)

Overall trends, by outcomes

This section presents overall trends and findings for studies included in our review. We do not include results from the broader literature (i.e., those that did not meet the inclusion criteria). There were limited studies on each outcome, so results should be interpreted with some caution.

Transportation outcomes

Studies found decreases in car travel after policy implementation (Andersson & Nässén, 2016; Karlström & Franklin, 2009; TfL, 2004; TfL, 2008). For instance, the number of vehicle kilometers travelled within the Western Extension of London’s congestion charge decreased by 11% the year after implementation (TfL, 2004). Mode shifts to public transport were observed in Stockholm and Gothenburg among those who initially commuted by car into the cordoned area (Andersson & Nässén, 2016; Karlström & Franklin, 2009). In Singapore, areas with increases in toll prices saw greater increases in public transport ridership (Agarwal & Koo, 2016). No studies in our review focused on mode shifts to walking and cycling. This is likely because mode shifts to walking and cycling are small and thus require large sample sizes to assess the distributional impacts across socio-demographic characteristics.

Health determinants

Outcomes included impacts on social interactions and air pollution concentrations. Congestion pricing was found to have a negative impact on social interactions, leading to fewer visits to family and friends and to act as a caregiver (Munford, 2017; TfL, 2004). In terms of traffic-related air pollution, congestion pricing decreased concentrations of NO₂, NO_x, and PM₁₀ (Beevers & Carslaw, 2005; Tonne et al., 2008). There were no studies that assessed the distributional impacts on noise pollution or physical activity.

Health outcomes

The health outcomes examined included traffic incidents (measured), asthma (measured), and life expectancy (modelled). Generally, road pricing was found to reduce traffic incidents. Two studies reported reductions in traffic incidents inside the priced area after implementation (Green et al., 2016; Percoco, 2016). Although, one study assessing impacts in the first two years of London’s policy did find evidence of increased motorcyclist casualties in the area outside the priced area (Noland et al., 2008). Through the pathway of reduced air pollution, congestion pricing was found to reduce acute asthma attacks in young children in Stockholm and lead to modest gains in years of life in London (Simeonova et al., 2019; Tonne et al., 2008).

Distributional effects across population groups and regions

Our goal was to assess the distributional impacts across population groups and geographic areas. Studies examined the effects of congestion pricing by income, gender, age, and geographic region (refer to Appendix B for results in table format). We did not find peer-reviewed literature that examined the distributional effects of road pricing by race or ethnicity.

Income

Nine studies examined effects by income level. Outcomes included commuting by car, mode shift to public transit, affordability, welfare, social interactions, air pollution exposure, and life-years gained. There were differential impacts across income groups for all outcomes, with the exception of car commuting in Stockholm and Gothenburg, and mode shift to public transit in Stockholm. From a transportation perspective, trends across studies seem to suggest lower income groups are more likely to be change travel behaviour and be negatively impacted by the cost. There are too few health studies to make a similar claim for health, but one ecological study did find reductions in health inequities in terms air pollution exposure and life expectancy gains (Tonne et al., 2008).

The evidence is not entirely consistent across cities, but overall findings suggest that congestion pricing is more disruptive to people with lower incomes. After the implementation of the Western Extension in London, lower income households were more likely to change their car use as compared to higher income households (TfL, 2008). Interestingly, lower income respondents were equally likely to have increased their travel as decreased, with ‘making the most’ of paying the charge being cited as the most common reason for increasing car travel. Bus ridership in Singapore was more likely to increase in lower income areas after increases in toll prices (Agarwal & Koo, 2016). In contrast, three pre-post studies on the Stockholm and Gothenburg systems did not find differences in car commuting or mode shift across income. (Andersson & Nässén, 2016; Franklin, 2012; Karlström & Franklin, 2009). In terms of affordability, lower income households were more likely to report difficulty affording the charge (TfL, 2004), and in Stockholm and Gothenburg toll payments made up a larger share of total income for lower income households than for higher income households (Eliasson, 2016). The two studies that compared the average welfare effect (in terms of travel costs and benefits) arrived at different conclusions for the people in the highest income group (relative the middle income group), but both found that lower income groups were worst-off (Karlström & Franklin, 2009; West & Börjesson, 2020). In terms of social connectedness, lower income households were more likely to find it difficult to visit family and friends after the implementation (TfL, 2004). There was one study in London that found greater benefits for more deprived areas with respect to decreases in air pollution and mortality (Tonne et al., 2008).

Gender

Five studies examined effects by gender. Outcomes included commuting by car, mode shift to public transit, welfare, affordability, and social interactions. They were differential impacts by gender for commuting by car, welfare, affordability, and social interactions. Across studies, it appears that men fare the effects of road pricing slightly better than women.

For impacts on commuting by car, the evidence was not consistent between Gothenburg and Stockholm. In Gothenburg women were two times more likely than men to reduce car commuting (Andersson & Nässén, 2016). In contrast, women in Stockholm were less likely to reduce their car trips (Franklin, 2012). This finding from Stockholm was explained by the tendency for women to have lower access to a car and greater possession of a transit pass in the first place. Another study in Stockholm on mode shift to public transit did not find any differences between men and women, but found that for those that did switch, travel times increased more for women than for men (Karlström & Franklin, 2009). In the welfare analyses, one study found no significant differences between men and women (Karlström & Franklin, 2009), and the other concluded that women suffered larger net losses because they were less likely than men to have access to a company car (6% of women compared to 28% of men), which means they are less likely to have charges for private trips covered by their employer (West & Börjesson, 2020). Women in London were two times as likely as men to find the charge difficult to afford (TfL, 2004). In terms of social interactions, women were more likely than men to find it more difficult to visit with family and friends in the congestion charging zone (half of women compared to 36% of men) (TfL, 2004).

Age

Four studies considered effects by age. Outcomes included commuting by car, travel time, welfare, accessibility to shops and services, social interactions, and asthma. Three studies examined the distributional effects of congestion pricing in adult populations and one focused on children. There were differential impacts by age group in terms of travel time, welfare, accessibility, and social interactions. It is difficult to make overall conclusions as to whether transportation and health inequities are being reduced or exacerbated without more local context about existing inequities across age groups.

Across three outcomes - travel time, accessibility, and welfare – people aged 35-55 benefited most. Amongst those who live inside London’s congestion charging zone, people aged 35-54 were most likely to say they spend less time travelling as a result of the congestion charge and have better access to shops, facilities, and services (TfL, 2004). The welfare analysis of Gothenburg’s congestion charge also found that people between the ages of 36-55 benefitted most, as this age group made the most trips, and therefore profited most from travel time savings (West & Börjesson, 2020). In terms of social interactions, a fifth of people aged 25-34 found it more difficult to visit family and friends after the implementation of the London congestion charge compared to 12-13% in other age groups (TfL, 2004). Only one study examined the impact of congestion pricing on morbidity, and this was specific to children aged 0-5. Simeonova et al. (2019) found that the congestion tax in Stockholm significantly reduced the number of hospital visits for acute asthma by 9.6 cases per 10,000 children/month over two years.

Geographic region

Seven studies considered effects by geographic region. Outcomes included car travel, accessibility, affordability, social interactions, air pollution, life expectancy, and traffic collisions. Six were based in London and one in Milan. There were differential impacts for the area inside and outside the cordon across all outcomes, with the area inside typically benefitting to a greater extent. Unfortunately, no context was provided in the studies about the income distribution of residents living inside the cordon relative to those outside, so the overall impacts on transportation and health equity are unclear.

In London, the proportion of people who drive remained stable for respondents living inside the priced area (eligible for 90% discount), but decreased for respondents living in the area just outside congestion charging zone, particularly for commuting and business trips (TfL, 2004). Perceptions of accessibility to shops, services, facilities, and places remained the same for the majority of residents after the implementation of the charge. Slightly more respondents living inside the congestion charging zone indicated better access as a result of the charge, compared to respondents living outside (20% vs. 14%) (TfL, 2004). The majority of respondents in London found the charge easy to afford, however, a higher proportion living in Outer London found the charge difficult to afford (TfL, 2004). In terms of social interactions, the charge was found to have a negative effect on social interactions for both those living inside and outside the zone (Munford, 2017). There appears to be benefits in terms of decreases in vehicle emissions for both the areas inside and outside the zone, but to a greater extent for the area inside. However, Beevers & Carslaw (2005) did estimate a slight increase in NOx emissions (+1.5%) along the Inner Ring Road, which encircles the charged zone in London. Through the pathway of reduced air pollution, Tonne et al. (2008) estimated that gains in mortality would be greater for the area inside the zone (183 years of life gained per 100,000 people) compared to the area outside (18 years of life gained per 100,000 people).

Three studies evaluated the effects on traffic collisions. One assessed impacts in the first two years of London’s congestion charge, and found evidence of a small decrease in car casualties (Noland et al., 2008). A subsequent study reported more conclusive evidence: traffic collisions inside the priced area decreased by 35% per month, compared to a 10-12% per month decrease in the two spillover regions just outside the priced area (Green et al., 2016). There was also evidence of increased injuries and deaths for motorcyclists just outside the congestion charging zone. This was attributed to an increase in the number of motorcyclists, which are exempt from charges (Noland et al., 2008). The study in Milan reported a 18.8% decrease in traffic incidents inside the boundary but no decrease outside (Percoco, 2016).

Discussion

This scoping review mapped the effects of road pricing on transportation and health equity, covering a diverse range of transportation outcomes, health determinants, and health outcomes, and synthesizing evidence across income levels, gender, age, and geographic region. The existing evidence suggests that area and cordon pricing is beneficial from a transportation and health perspective, but that the benefits are not always distributed equitably across populations and space. Here we summarize the main takeaways and identify key research gaps and opportunities for future research.

Road pricing had a neutral or positive effect for most transportation and health outcomes. People travelling by car adapted to road pricing by reducing the frequency of car trips or switching to public transit. Via the pathways of decreased air pollution, congestion charging related to gains in mortality in London and to decreases in acute asthma attacks in young children in Stockholm (Simeonova et al., 2019; Tonne et al., 2008). Traffic incidents decreased overall (Green et al., 2016; Noland et al., 2008; Percoco, 2016), although one study did show an increase in motorcyclist casualties (Noland et al., 2008). On the other hand, the frequency and ease of social interactions was negatively impacted, with the implementation of congestion charging leading to fewer visits with family and friends (Munford, 2017; TfL, 2004; TfL, 2008). This was also observed in qualitative interviews with people with disabilities; interviewees felt that visits from family and friends had reduced on weekdays and they were more reluctant to ask friends and family for help because of the charge (TfL, 2008). This is an important consideration for current and future road pricing policies, given established links between social connectedness and health (Umberson & Karas Montez, 2010).

Across populations and space, the groups that generally fared better were people with higher incomes, men, people aged 35-55, and residents living inside the priced area. People with lower incomes were more likely to change their travel behaviour in response to road pricing and report difficulty affording the charge (Agarwal & Koo, 2016; TfL, 2008). Notably though, the studies reviewed here did not take into account revenue redistribution or improvements to public transportation, which could positively affect travel for lower income households. In terms of gender, men fared slightly better than women: women’s travel times increased more, women were less likely to have access to a company car, and were more likely to indicate difficulty affording the charge and visiting with friends and family. People aged 35-55 benefited most in terms of travel time savings, accessibility, and welfare effects (TfL, 2004; West & Börjesson, 2020). Across space, there are benefits for both the areas inside and outside the priced area, but to a greater degree for the area inside. This reinforces that, in area and cordon pricing systems, the location of the boundary is a critical determinant of the transportation and health consequences for the region. The studies we reviewed did not provide context of geographic income distributions. Since the area inside the cordon experiences a greater share of benefits, assessing the geographic distributions of household income and other socio-demographic variables relative to the cordon boundary should be a focus.

To make more definitive conclusions about whether road pricing reduces or exacerbates transportation and health inequities, studies would ideally a) provide data on baseline differences pre-implementation and whether there are differential impacts postimplementation, and b) discuss the transportation or health consequences of an inequitable distribution of the outcome under consideration. For example, Transport for London (2004) reports that 30% of women (versus 16% of men) found it more difficult to meet up with friends and family. However, baseline data of the relative differences in the ease or frequency of social connections by gender is not provided. Several studies focused on the impacts on the frequency of car travel, but did not take that next step to discuss the transportation or health consequences of a shift away from car travel. Positive consequences of a shift away from car travel may include less sedentary time and decreases in air pollution, but on the other hand, it could also lead to increased travel time and lower accessibility. One example of a study that provided a clear description as to how road pricing impacted inequities was Tonne et al. (2008). They found more deprived neighbourhoods in London had higher concentrations of air pollutants than less deprived neighbourhoods to begin with, but that congestion pricing reduced health inequities as it led to greater reductions in more deprived neighbourhoods. More details like this - describing the equity implications of differential impacts across population groups - is needed to clearly understand how road pricing is affecting transportation and health equity.

In addition, the evidence base is limited by insufficient data in three areas: 1) a broad set of health determinants and outcomes; 2) longer-term impacts; and 3) evaluation focused on: non-work-related trips, people who use sustainable modes of transportation, or impacts across different racial groups. First, we originally set out to understand the impacts of large-scale road pricing policies on health inequities, however, there are so few studies on health – even before looking at differential impacts or inequities. For select health outcomes there is some additional literature on aggregate impacts: physical activity (Bergman et al., 2010; Brown et al., 2015; Kaida & Kaida, 2015), traffic incidents (Li et al., 2012), air pollution (Cavallaro et al., 2018; Johansson et al., 2009), and mortality (Johansson et al., 2009; Yu et al., 2019). We did not find ex-post studies on the impacts on noise pollution or morbidities, with the exception of Simeonova et al. (2018)’s study on acute asthma attacks in young children. Second, few studies provide evidence on the longer-term impacts: eleven of the fifteen studies assessed impacts at 1 year or less after implementation. Third, there is insufficient evaluation in a few key areas. Studies focused on commute trips, limiting insights for non-work travel that account for over half of trips (TransLink, 2019). As non-work trips are more likely to be made by women (Ng & Acker, 2018), this is particularly important for understanding the gendered impacts or road pricing. There was also a tendency toward drivers’ perspectives. A fulsome evaluation should consider impacts for all road users, as improvements to public transit are often touted as a way road pricing can support transportation equity. Finally, we found no ex-post studies that considered impacts by race. Transportation planning can create conditions that further perpetuate or reduce racial inequities in transportation and health (Pitter, 2020), and no doubt road pricing has a role.

Our scoping review focused on ex-post rather than ex-ante studies of proposed policies. Evidence from ex-post evaluations reflect ‘real-world’ implementation, while ex-ante studies require more assumptions about how road pricing will be implemented and how people will adapt to the new policy. Ex-ante studies are useful for comparing potential effects of different road pricing scenarios, but results are heavily influenced by the assumptions of the researcher. Our focus on ex-post evaluations limited our review to the small number of cities which already have area- and cordon-based road pricing systems. There are different types of road pricing schemes not included here. Distance-based systems are not widely implemented yet, but are being considered in a number of places including Vancouver and Washington State (TransLink, 2018; WSTC, 2020). Modelling studies suggest these systems have potential to advance equity (Mitchell, 2005; Sana et al., 2010). As these systems are implemented, equity for this type of road pricing can be a focus for future reviews. Finally, we did not include LEZs. Over 250 European cities have implemented LEZs (Müller & Le Petit, 2019), including the cities in this review. Future research may consider how effects from a congestion pricing policy interact with effects from LEZs.

Limitations

Since large-scale road pricing systems are implemented in a handful of cities, our scoping review only includes systems in Europe and Singapore where public transit ridership is relatively high. Findings may not be transferable to cities with more car-dependant populations. In addition, there were limited studies across specific transportation and health outcomes. In some cases - such as for social interactions and air pollution - conclusions were based a single system. More studies across a wider range of settings are needed to build a stronger evidence base. Since the purpose of this scoping review was to map the literature, we did not appraise the quality of evidence. Lastly, we examined the distributional effects across socio-demographic factors (e.g., income, age, gender) but did not explore the interactions between these factors. Most studies included in our review did not take an intersectional lens. We encourage future research on the equity implications of congestion pricing to consider the different social identities, power relations, and experiences that shape inequities in road pricing.

Conclusion

We present a synthesis of the effects of road pricing on transportation and health across populations and space, highlighting key areas where gaps exist. Despite some degree of inequity, we find that area and cordon pricing are beneficial for numerous transportation and health outcomes including a reduction in car trips, air pollution, asthma attacks, and road traffic collisions. More evidence in diverse urban settings on a wider range of health determinants and outcomes is needed to strengthen the evidence base. Future work on area and cordon pricings should report details about the geographic distribution of socio-demographic factors relative to the cordon boundary, as benefits are more pronounced for those inside the cordon. Further, studies should investigate the impacts of road pricing on non-work-related travel, for non-car travel, and across racial and ethnic groups. These considerations can help advance our understanding of the effect of road pricing policies on transportation and health equity.