It is increasingly popular for cities to exempt clean cars from congestion pricing tolls. This column uses data from Stockholm to show that congestion pricing with exemptions for alternative vehicles can reduce traffic while incentivising green vehicle adoption, However, these exemptions are best employed when there are few existing electric vehicles on the road.
Road pricing is an old economic idea that has been the topic of much debate in 2024. While New York recently shelved its planned congestion pricing programme, many other cities are still considering similar policies. Congestion pricing is a topic of perennial debate in San Francisco and Boston, and many cities in the UK and Spain have recently implemented ‘Clean Air Zones’ – a related policy already widely adopted in Europe.
The world’s first congestion pricing zones in Singapore and London focused on traffic externalities, not environmental outcomes. Increasingly, however, cities are attempting to fold environmental goals into these transportation policies by exempting or discounting green vehicles. In Milan, for example, all electric, hybrid, and biofuel vehicles are exempted from the city’s congestion charge. London introduced an electric vehicle (EV) exemption to their congestion zone in 2021, and Oslo’s urban toll ring features discounts for clean cars.
There is a considerable body of empirical evidence suggesting that congestion pricing policies can reduce local air pollution (Gibson and Carnovale 2015, Green et al. 2020) and improve health (Simeonova et al. 2018, 2019). At the same time, the economics literature provides examples of well-intentioned urban transportation policies that were ineffective or counter-productive because of unanticipated responses by commuters. Mexico City’s license plate-based driving restrictions were undetermined by second vehicle purchases (Davis 2008) and corruption (Oliva 2015). Green et al. (2020) show that although London’s Congestion Zone reduced many pollutants, NO2 pollution increased because of exemptions for certain types of vehicles (like taxis) which were more often diesel powered. Nishitateno et al. (2024) suggest that alongside implementing traffic pricing measures, promoting the adoption of clean vehicles is essential for reducing pollution from the road sector.
The argument on the side of policymakers is that green vehicle exemptions distill the best parts of congestion pricing schemes by reducing traffic and tilting the scales toward clean transportation (Santos 2007). But do these exemptions encourage the use and adoption of clean cars, or are they simply weaker congestion pricing policies? And if green vehicles are exempted, how should policymakers set the prices on the remaining non-exempt vehicles?
We answer these questions in a new paper (Nilsson et al. 2024). First, we describe how a policymaker should set road prices on non-exempt vehicles, and how this toll depends on the different ways that drivers respond to road pricing. We then use data from Stockholm – which has a congestion zone that temporarily exempted clean vehicles – to understand how drivers react to these policies. Finally, we use these estimates to calculate the optimal toll on the remaining priced vehicle trips in Stockholm and to discuss whether green exemptions pass the cost-benefit test.
What responses matter?
In the simplest model of a Pigouvian tax (like a congestion price or a carbon tax), the optimal tax does not depend on how people or firms respond to the tax; it only depends on the externalities generated by some activity. In practice, however, Pigouvian taxes often influence related behaviours that come with social costs.
This is the case with road pricing: when drivers face higher tolls, they may take fewer trips. They may also respond by changing where they live, taking detours around the congestion zone, or changing the vehicle that they drive. These responses may alleviate or exacerbate existing externalities, which complicates the calculation of an optimal congestion price.
In our paper, we describe how the optimal price on non-exempted vehicles depends on three types of responses: (1) driving behaviour (i.e. the number of trips), (2) vehicle type, and (3) commuting distance (i.e. where people live and work).
All else equal, when responses to the congestion price exacerbate other externalities, policymakers should lower congestion prices; when responses decrease related externalities, policymakers should increase congestion prices. For example, if a higher congestion price with a clean car exemption leads to increased adoption of clean vehicles, policymakers may want to charge conventional vehicles a toll that exceeds their direct social cost. The same is true for moving: if a congestion zone encourages individuals to ‘sort’ and live closer to their workplace (and thus drive fewer miles on their average commute), this is another incentive for policymakers to increase tolls on the margin. Alternatively, if the main response to a policy like Stockholm’s or Milan’s is that people drive longer trips to circumvent the congestion zone, then policymakers may want to lower tolls.
The interplay between these different responses is not trivial. In the second part of our paper, we turn to a natural experiment in Stockholm to understand how drivers respond to congestion pricing in general, and to green vehicle exemptions in particular.
Is there a trade-off between environmental and transportation goals?
In addition to calculating the optimal toll for non-exempt vehicles, we also ask a simple but policy-relevant question: do exemptions for green vehicles pass the cost-benefit test?
Exempting green vehicles incentivises commuters to adopt electric vehicles, or switch from using conventional vehicles to using clean vehicles if their household has a mixed-type fleet. But this incentive comes at a cost. Regardless of a vehicle’s emissions intensity, driving that car downtown still causes congestion, meaning that any adoption of green vehicles induced by an exemption comes with some cost in foregone traffic improvements.
So, for every additional green vehicle added to the fleet because of the exemption, what are the foregone externality benefits? This depends on the size of the existing fleet and how responsive EV purchases are to an exemption. If the existing fleet is large or people don’t readily switch to EVs in response to the policy, the opportunity cost of these exceptions could be high. Alternatively, if the exemption only applies to a small share of vehicles, and the policy encourages many prospective car buyers to switch to clean cars, then these exceptions could be worth their cost. In our paper, we derive a simple expression that demonstrates this trade-off, and again turn to Stockholm’s policy to evaluate this expression using real data.
Takeaways from a natural experiment in Stockholm
After a brief trial period, Stockholm fully implemented a congestion pricing zone in 2007. Existing studies have shown that this policy reduced traffic (Börjesson et al. 2012) and air pollution (Simeonova et al. 2018) in Stockholm.
One under-studied component of the congestion pricing zone is its exemption for alternative fuel vehicles: for 18 months from 2007 to 2009, newly registered alternative fuel vehicles were exempt from the congestion tolls. This policy ended for new registrations in 2009, and originally exempted vehicles retained their free trips through August of 2012.
Figure 1
We revisit how drivers responded to Stockholm’s congestion charge with an eye on changes in the adoption and use of alternative fuel vehicles. Our approach uses the fact that two motorways in Stockholm – the Essinge bypass and the Lidingö route – were exempted from the city’s congestion charge. We then rely on comprehensive administrative data to compare the driving behaviour, vehicle purchase decisions, and changes in work/home locations of ‘treated’ commuters (who faced tolls) and non-treated commuters (who did not face tolls because of where they lived and worked).
By comparing individuals who were and were not exposed to congestion tolls (and the accompanying green car exemption), we find that commuters responded by (a) taking fewer trips, (b) adopting alternative vehicles in place of conventional vehicles, (c) substituting miles in conventional vehicles for miles in alternative fuel vehicles, and (d) moving where they live and work.
Both driving and vehicle purchases tell a similar story: people substitute from conventional vehicles to alternative fuel vehicles, and reduce reliance on vehicles overall. Relative to those who aren’t exposed to tolls, tolled commuters were 0.6 percentage points more likely to own an alternative fuel vehicle, and 0.8 percentage points less likely to own a conventional vehicle. Similarly, vehicle kilometres travelled by the average alternative fuel vehicle increased by 103, and vehicle kilometres travelled by conventional vehicles fell by 206 on average.
Regarding moving, we find that treated commuters are 0.2 percentage points more likely to change where they live, and 1.6 percentage points more likely to change their work location relative to non-treated commuters. Sorting is a fundamental part of many models in urban economics; we believe that this is the first evidence of sorting in response to congestion pricing. As discussed above, this sorting has important implications for setting congestion prices. We find that because commuters sort, the distance between work and home decreases, leading to lower per-trip externalities.
Finally, we combine all of these empirical estimates with our model to estimate the optimal price on non-exempt trips in Stockholm, and to discuss the cost-benefit of exempting green vehicles. We calculate an optimal price of €9.46 per trip – which is roughly twice as large as current peak-hour congestion pricing in Stockholm. We also find that the total cost of each alternative fuel vehicle that was adopted because of the policy was €3,795 in avoided congestion benefits. This cost is similar to, but slightly larger than, the avoided emissions costs.
A second point of comparison is the cost of inducing adoption through conventional subsidies. At the time of the exemption, conventional alternative fuel subsidies were roughly €1,057. Later, green vehicle subsidies were €7,400 in Sweden. An apples-to-apples comparison of these approaches requires knowing the share of inframarginal adopters in conventional subsidy programs. Assuming that fewer than 70% of purchasers were inframarginal, Stockholm’s green car exemptions were likely a more costly method of encouraging adoption than the contemporaneous conventional subsidies (because €1,057/0.30 < €3,795). At the same time, our estimates suggest that Stockholm’s exemptions were more cost-effective than later green vehicle subsidy programmes.
Implications for congestion pricing and urban policy
As the expansion of congestion pricing policy coincides with a period of intense focus on environmental outcomes, many cities are attempting to combine environmental and transportation goals in road pricing policies. We find that exempting clean cars can lead drivers to use these cars more often, and to adopt these cars over conventional vehicles. But these exemptions can be costly, and are more likely to pass the cost-benefit test when the existing clean fleet is small. For this reason, exemptions may not be appropriate in regions that are already far along the adoption curve. Similarly, cities considering exemptions or aggressive discounts for green cars might consider temporary exemptions, following the example set by London.
Source : VOXeu
