Should Sweden invest in electric road systems (ERS), or abandon the concept in favor of more fast charging stations and bigger batteries in vehicles?

Last week, the Swedish Transport Administration ( Kenneth Natanaelsson et al.) published the report Planeringsunderlag elvägar, or Planning Basis for Electric Roads, which contains an analysis of the socioeconomic impact of electric road systems (ERS) in Sweden. The report's cost-benefit analysis is negative, indicating that the socio-economic benefits of a public investment in ERS construction would not outweigh its costs. The findings presented in the new report are similar to those in another analysis by the Transport Administration published in 2021.

The report is in Swedish and I have been asked by international colleagues to explain the underlying analysis. This article presents:

• A chapter-by-chapter summary of the 75 page report, focusing on assumptions, methodology and results.
• Contextualization of the report in relation to other research from Sweden and abroad.
• My own subjective assessment of the report.
• My recommendations for the future of ERS in Sweden.

Summary of the Swedish Transport Administration's report

[Disclaimer: I did not write this report and while I have tried to accurately represent the contents, I do not take any such responsibility. I include a few personal inline comments in square brackets.]

All costs in the report are in SEK (2019 monetary value), which I have converted to euro with a conversion rate of 10 SEK = 1 €.

1. Introduction

Electric road systems enable dynamic charging via overhead lines, conductive tracks, or inductive systems, requiring new infrastructure and vehicle adaptations. Additionally, payment systems, monitoring, and traffic control infrastructure are necessary. Four technology demonstrations have been conducted in Sweden and tests are ongoing in several countries around the world. No international agreements exist yet for what charging interface to use or how to pay.

The Swedish Transport Administration has been tasked to deliver a basis for a plan for construction of ERS in Sweden, including a cost-benefit analysis, to support decision making.

The analysis considers only heavy-duty vehicles as ERS users, primarily trucks in long-haul operation. Light-duty vehicles are excluded on the basis that static charging is deemed sufficient. The context for the plan is an assumption that Sweden achieves 100% zero-tailpipe-emission heavy-duty truck sales by 2040, mainly battery-electric, without ERS. Infrastructure for public static charging and hydrogen refueling has good geographic coverage today in Sweden, but the capacity must increase. Vehicle OEMs produce BEV trucks, plan for FCEV trucks by 2030, but have no public plans for ERS truck production.

2. ERS expansion options

Four road networks are discussed. Key parameters for international readers include the single-lane bidirectional road lengths and traffic densities of each network. Investment cost scales with road network length and revenue potential scales with traffic.

1. Small (Malmö - Gothenburg - Stockholm, 987 km, 4940 heavy AADT today of which 2740 with trailers).
2. Medium (Most main roads in southern Sweden, 1602 km, 4170 heavy AADT today of which 2250 with trailers).
3. Large (Expansion to the less populated north, 3010 km, 3060 heavy AADT today of which 1610 with trailers).
4. Passenger cars (3800 km, approximately 1800 long-distance trips per day today).

ERS infrastructure is assumed to be installed on 40-60% of the road distance within each road network. Gaps are traversed on battery power. Driving within the ERS network should yield a net gain in battery charge so that the full trip energy [I believe - it says start to end] can be covered by ERS charging.

It would be very risky for Sweden to start ERS construction before other countries in Europe have committed to do so. A Swedish financing decision can be taken no earlier than 2030, followed by planning 2030-2033; procurement 2032-2033; and construction 2033-2038 (i.e., 100 km per year). The years are for road network Small. Construction of each network step is assumed to continue sequentially, with planning and procurement overlapping with construction of the previous step. Completion of the Large network is deemed feasible by 2048.

3. Compared alternatives

A base alternative without ERS is compared to alternatives 1 and 2 with ERS. ERS construction is not expected to reduce GHG emissions, as 100% ZEV sales will already have been achieved by the time the infrastructure becomes operational. In alternative 1, ERS is deployed on road network Small. In alternative 2, ERS is deployed on road network Medium.

GHG emissions are not expected to decrease as a result of ERS construction, as new ZEV sales have already reached 100% when the infrastructure is fully operational.

Chapter 1 noted that ERS may contribute to e-retrofits, which motivates that the cost-benefit analysis assumes the final rate of utilization is reached from the date the network is completed.

4. Assumptions for the cost-benefit analysis

ERS infrastructure costs

Costs include investment, operations and maintenance. Costs are uncertain as all estimates are based on small-scale trials, experiences from similar infrastructure, anticipated future market developments and forecasts from experts and manufacturers. The full installation cost per bidirectional single-lane infrastructure km is estimated at €3.5-4.5 million/km. Annual operations and maintenance costs are estimated at 2-4% of that.

The authors note that costs could be reduced if ERS installation takes place during regular road maintenance, in particular for the in-road technologies, but it is not specified by how much.

Battery costs

€140/kWh (pack) in 2025, average €51/kWh during 2040-2070 (listed under vehicle costs).

Truck batteries are assumed to have 600 kWh capacity, 650-875 km range, covering 77-97% of daily operating ranges for long-haul trucks. Cases with 405 and 870 kWh range will also be analyzed.

The report mentions that academic studies that analyzed single-route trips (including my own) have concluded ERS can enable up to 70% reductions in battery capacity for affected vehicles. Scania expects 30-50% capacity reduction. The AFIR-mandated distance between public static charging points prevents more than 70% capacity reduction.

The authors assume that vehicles that drive more than 50% of their annual distance outside of the ERS network will have the same battery capacity as without ERS. A vehicle that drives 90% on the ERS network is assumed to get 70% capacity reduction. A weighted average of 30% reduction for long-haul trucks is used in the analysis, and a "smaller reduction" for distribution and construction trucks.

Costs of static charging

Electricity is assumed at €0.11/kWh. Static charging infrastructure is assumed to cost €0.05, €0.175 and €0.3 per kWh for private (100 kW), semi-public (350 kW) and public (1000 kW) charging.

No difference in energy consumption is assumed for ERS and non-ERS BEVs.

ERS is assumed to contribute 39-43% of total energy to ERS-adapted vehicles. Static charging infrastructure costs and percentages of delivered energy are reported on page 37. ERS reduces demand for static charging, at different percentages.

ERS is estimated to yield a reduction of static charging costs at approximately €0.05/kWh. [I don't follow this part.]

Vehicle costs

Long-haul, 2025: €550,000 for BEV, and €500,000 for ERS-BEV (15% reduction). I believe the cost includes the trailer.

Long-haul, 2040-2070: €323,000 for BEV, and €315,000 for ERS-BEV (2% reduction).

Regional and distribution, 2040-2070: €183,000 for BEV, and €180,000 for ERS-BEV (2% reduction).

[I believe only the cost differences affect the analysis, not the actual costs.]

Value of time

It is qualitatively discussed that ERS could yield time savings, but I don't believe value of time is included in the cost-benefit analysis.

Value of weight

Based on reasoning about cargo types and other statistics, very few ERS-BEVs are expected to be weight-limited. If I understand correctly, the conclusion is that 30% of ERS users will gain a 2-3% increase in cargo capacity.

[If battery weight savings do not contribute to increased cargo loads, they should contribute to energy savings. Assuming a rule of thumb that a 10% reduction in vehicle mass yields a 5% saving in energy consumption, 600 kg weight reduction in a 44 ton loaded/16 ton unloaded/60% fill rate truck should yield 1% average reduction in energy consumption.]

Total operating costs

Long haul: €2.22/km for BEV, 2.20/km for ERS-BEV.

Distribution: €2.09/km for BEV, 2.07/km for ERS-BEV.

Land use

Large differences in local land use for vehicle charging with and without ERS, but small differences in total. Static charging is assumed to be almost fully built out by 2040, thus ERS is not assumed to reduce the land use of static charging.

[If 100% new sales is reached by 2040 and 20 years vehicle lifespan, approximately 55% of vehicles and a bit less of traffic should still be ICEVs. This should mean at least ~45% of future static charging capacity has not been installed by 2040. Since national ERS decisions are assumed to be taken in 2030, charging infrastructure investors should have plenty of time to align their expansion plans with the new context.]

ERS utilization rate

[This part is tricky to follow and the risk for misinterpretation is therefore greater.]

The report cites a 2021 analysis by the Swedish Transport Administration, which presented data from the Swedish truck OEMs, showing that in the population of trucks that visit the Small/Medium Swedish road networks at least five days per year, 4%/15% of the trucks spend more than 50% of their annual distance within each of the two Swedish networks. Higher resolution histograms are presented in the earlier report on page 45. [Then I don't completely follow the report's explanation on page 50.] Based on these data, 12.5% AADT heavy with trailer on the ERS networks are expected to come from trucks that drive at least 50% on the Medium network. Including other vehicles, the utilization rate of a Swedish road network is assumed at 20% of heavy AADT.

[As far as I can tell, no unit conversion is made from percent of vehicles in the population that visits the network to percent of traffic (AADT) within the network. To illustrate, if 10 vehicles drive on a road and 2 of the vehicles (20% of the population) contribute 800 of 1000 vehicle passages, then the share of traffic from frequent visitors is 80%, not 20%. This is further discussed in Concern no. 1, below.

Also as far as I can tell, the data from the vehicle OEMs specify the share of total annual distance (not of distance in Sweden) on three different Swedish road networks. Long-haul vehicles, especially foreign vehicles, drive both inside and outside of Sweden. This means that distance outside of the Swedish ERS network does not equal distance without access to ERS. The report assumes ERS in Sweden will only exist as part of an international ERS network, which, as far as I can tell, is not what is actually being calculated.]

5. Cost-benefit analysis

Alternatives 1 and 2, introduced in the beginning of the report, are not used. Instead, the calculation is based on a 60 km average road.

Parameters:

1. Discounting year: 2028. Start of ERS operation: 2040. 20% tax financing cost.
2. Investment cost is €4 million per ERS km. Annual cost of operations and maintenance is 3% of the investment cost. 30 year amortization period.
3. ERS revenue is set equal to the cost of operations and maintenance, canceling it out.
4. 60 km road network. ERS distance is 50% of the road distance.
5. 6000 AADT, 20% utilization rate. There is some separate handling of different vehicle classes, but I think it has little impact.
6. ~€0.05/kWh saved investments in static charging. ~€0.025/km in reduced "other" costs.

This yields a negative net present value, i.e., investment results in a net socio-economic loss.

[Total vehicle distance driven on this 60 km stretch over 30 years is ~4 billion km. Total non-discounted infrastructure costs are ~€200 million. At 20% utilization, that translates to €0.25/km, and at 100% utilization to €0.05/km. Divide by distance-averaged power draw (~2.2 kWh/km including charging) to get €/kWh. Then add costs of energy, operational grid fees and taxes to arrive at an estimated total cost of charging.]

6. Sensitivity analysis

None of the sensitivity analysis scenarios yield a significantly positive net present value. One of the scenarios assumes a utilization rate of 50%, which yields a net-present value near zero.

Alternatives 1 and 2 are discussed qualitatively. The report notes that ERS may influence road wear, traffic accidents, noise and particle emissions, which are briefly discussed. These effects are assumed to be minor. The cost-benefit analysis is best for the Small and worse for the Medium and Large networks, as expansion adds cost but traffic density is lower. [Network expansion may increase utilization on already built sections, which is not considered. The effect may be minor if a Swedish ERS network is an expansion of an EU network.]

There is a qualitative discussion about visual and wildlife impact (some for overhead catenary) and electromagnetic emissions (take care around water bodies).

ERS is not considered to have potential to reduce Sweden's climate impact, based on that the infrastructure will not increase BEV uptake when built by 2040, that >90% of battery resources will be in light-duty vehicles, and that only a minority of heavy-duty vehicles are assumed to use the infrastructure.

BEV and ERS-BEV trucks are presented as having half the GHG impact of FCEVs. The market share for FCEVs is only assumed to be significant if their total cost of ownership becomes similar to that of BEVs, at which point their inclusion in the economic calculations would have minor impact.

7. Uncertainties

Uncertainties are discussed with regards to autonomous vehicles, hydrogen, EU CO2 regulation, AFIR and geopolitics.

7. The report's recommendations

Based on the risks in relation to the expected benefits, it is deemed too risky for as a small country like Sweden to be a pioneer in ERS rollout.

There is no climate benefit, if the network comes into operation in 2040.

The sensitivity analysis is considered robust, as all scenarios have negative or zero benefit.

The recommendation is that Sweden should not plan for a larger national deployment of ERS.

Sweden should continue following developments at EU level and should consider investigating if there are market-driven ways to introduce ERS.

My opinions on the report

First of all, I think the report's main conclusion is sound: there needs to be a clear commitment at EU level for Sweden to proceed with construction of a nation-wide ERS network. Most of the assumptions and methods in the report are also sound. The qualitative discussion is both clear and informative.

However, I have three concerns with this report, of which no. 1 is the most important.

Concern no. 1: ERS utilization is underestimated

The report uses an assumption that 20% of heavy trucks within the ERS network will be users of the infrastructure. I have not been able to fully understand the data processing steps that lead to this assumption.

However, I have previously attempted to replicate the analysis in the 2021 study from which the figure originates, starting from the same input data, then converting from share of trucks to share of traffic, and accounting for ERS use abroad. My analysis, documented in this flowchart and Jupyter Notebook, arrived at a significantly higher >90% estimated utilization. I encourage further discussion to clarify the correct interpretation of these data. The Swedish Transport Administration has maintained in personal communication, both following publication of the 2021 report and the report from last week, that their analysis is correct. My manager agrees with my interpretation of the data and has approved my public questioning of this analysis.

In a recent simulation study that also only considered heavy-duty traffic but used different data and methods, I found that around 70% of heavy-duty traffic on the main Swedish road network would have economic incentives to use dynamic charging if the utilization fee is set to maximize operator revenue. Utilization would be higher if socio-economic benefits are prioritized than if revenue is maximized, a finding that agrees with prior work (I think that study also concludes around 70-80% utilization rate, but I don't find an explicit estimate). A study of the cost-competitiveness of different powertrains found ERS-BEVs to be attractive for a majority of heavy trucks. Since the share of ERS-using traffic within the ERS network should always be higher than the share of ERS-adapted vehicles in the total population, 70% utilization seems a conservative estimate also based on that study.

I would like to have more studies to compare with, but I have not been able to find others where the utilization rate as a share of traffic is determined experimentally from population data rather than through a series of high-level assumptions. I am particularly hesitant to rely on studies that assume from the onset that one form of charging is always preferred and that utilization of other methods is capped by what remains. If you know other forecasts, please share them in the comments.

As far as I can tell, if a higher utilization rate had been assumed in last week's report (e.g., 70% of trucks or 50% of trucks and 20% of light-duty traffic), the cost-benefit assessment of the main scenario would turn positive. I believe most, but not all, of the scenarios in the sensitivity analysis would also have ended up positive.

Concern no. 2: The timeline for ERS deployment lacks ambition

The marginal value of adding an ERS network to the base scenario increases the earlier the infrastructure can become operational. Part of the reason why the cost-benefit analysis is negative is because of a late ERS introduction. In a recent publication, I concluded that every year of delayed ERS deployment in Sweden past 2035 has an opportunity cost of €300 million, plus approximately €100 million worth of CO2 emissions (lifecycle, not tailpipe, €300 per ton SCC). Combined, that is almost 10% of the 25-30 year lifecycle infrastructure cost of a 1500 km road network, per year.

The timeline presented in figure 8 is in my view too sequential. Planning activities for the Small and Medium networks are expected to take a total of five years. Planning does not cost €4 million per km and it should be possible to complete planning by 2030. By 2030, a European decision should have been reached on the joint future of ERS on the TEN-T network. If a positive decision is reached, procurement for the Small network needs a year, followed by five years of construction. It should be possible to proceed with procurement of the expansion to the medium network immediately when the first procurement is completed.

ERS construction is likely least disruptive if it takes place during regular road maintenance. On roads with heavy traffic, maintenance is required every 5-10 years. If only 50% of the total road distance needs ERS infrastructure, construction should not need to take more than five years. This means the Medium network could be completed in 2037, six years earlier than indicated in Figure 8.

I am not aware of any reason why construction of different parts of the network could not proceed in parallel. ERS segments should become operational soon after they have been installed, not when the entire network is completed. Both costs and revenue should increase as the network grows.

Concern no. 3: Light vehicles are excluded from the cost-benefit analysis

This one is tricky. I am not aware of a single study anywhere in the world that has managed to estimate the incentives for light-duty traffic to use an ERS network. ERS charging is forecast to be cheaper than public static charging for cars. However, vehicle prices are likely unaffected, as the value of any battery savings in passenger cars are likely cancelled out by additional costs for ERS-specific in-vehicle equipment.

It is probably safe to say that not every passenger car would be ERS-equipped. At-home charging likely meets the needs of many. But what about commuters that drive mainly on motorways anyway? What about the approximately 20% of cars that lack private parking and lag far behind in BEV uptake, despite nearby access to public charging? What about taxis and other public transport? And though less relevant for Sweden, what about fitting public chargers into dense and old European cities like Paris and Rome? Can we safely assume that static charging will be the most attractive solution for all charging or all cars and vans while only 6% of cars in use are fully electric?

Since the outcome of the report's cost-benefit analysis for a Swedish ERS network so strongly depends on total sold energy, and so weakly on in-vehicle savings, assuming zero energy sales to all traffic other than heavy trucks appears overly conservative. The combined energy consumption of Swedish light-duty traffic is higher than for heavy-duty traffic on motorways near cities, but lower on the long motorways in between. Including all roads, light-duty traffic consumes approximately three times as much energy as heavy-duty traffic. A fraction of this would make a big difference in the calculations.

Furthermore, 90% of battery resources for road traffic are forecast to end up in light-duty vehicles. Europe has both geopolitical and environmental incentives to reduce total battery demand. ERS should enable (but not necessarily incentivize) battery capacity reductions of similar magnitude in cars as in trucks. Range limitations due to restricted maximum battery weight is also one of the main barriers to e-retrofits for light-duty vehicles.

My recommendations for the future of ERS in Sweden

To conclude, I agree with the report's recommendation that Sweden should not commit today to an ERS investment. I also believe the decision to pause the procurement of the planned 21 km bidirectional permanent installation between Örebro and Hallsberg is logical. While I have not seen any assessment of the economic viability of the project, I personally doubt that such a short stretch offers sufficient utility for sufficiently many that enough revenue can be collected to cover the project costs. It's also unclear to me what can be learnt from the project that was not already learnt from Sweden's four prior physical ERS trials, and that cannot instead be learnt more quickly by following ongoing projects abroad.

As explained in Concern no. 1, above, I believe a higher utilization rate in line with other studies should have been assumed in the report's analysis, and as far as I can tell, adjusting this single assumption would change the cost-benefit assessment from negative to positive.

This is important, as it would indicate both that it is in Sweden's best interest to actively push for a European ERS network, and that even if Sweden pursues a delayed low-risk low-reward national strategy for ERS, this would still bring economic benefits to Sweden. It also appears that rewards can be increased without increasing risk by optimizing the timing of planning, procurement and construction, so that operation of a Swedish ERS network can start gradually from 2033 to 2037, instead of beyond 2040. An earlier start of operation is associated with higher willingness to pay among users, which leads to quicker pay-off time.

While Sweden should not yet commit to invest, I believe Sweden should be among the global leaders in ERS planning and research, to ready both itself and others for rapid procurement and construction if Europe moves ahead with dynamic charging. This is important both for the Swedish economy, which would most likely benefit from lower transport costs with ERS, and for the Swedish vehicle and charging infrastructure industries, which risk falling behind if Sweden does not stay ahead in areas of possible international growth.

Even though many factors related to transport electrification remain uncertain, academic studies (pre-print one, pre-print two) that estimate socio-economic benefits of ERS under uncertainty appear to conclude that ERS infrastructure is a no-regret investments for Europe's larger economies, where marginal impacts range from non-negative to very positive under different possible developments. The lower traffic density in peripheral countries like Sweden means that some developments are still possible that would bring net negative outcomes, which is why these countries should not be leaders in adoption.

A pan-European ERS network costs no more to build than a single year of European fossil fuels imports for road transport: €100-150 billion per year / (€4 million per km * 50% coverage) = 50 000-75 000 km, which is more than the length of the TEN-T Core network. Assuming that 1) main roads are resurfaced every 10-15 years, 2) ERS construction is scheduled together with periodic road resurfacing, 3) infrastructure is installed on 50% of the road network, and 4) countries deploy ERS in parallel, construction should only take 5-10 years. Such a network would bring both short-term and long-term transport system cost reductions, and make electrification of late-adopter segments easier. Sweden should encourage, facilitate and be part of this European future.

As research and planning continues, I believe the focus should shift from only enabling long-distance truck travel between remote cities, to an integrated infrastructure solution with great economies of scale that is used to power traffic where traffic is most dense - in and around cities. Long-distance trucks can top up while passing cities, rather than in remote areas far from other traffic. This perspective is both under-researched and under-funded.

Finally, in a study published earlier this year, my co-authors Mats Alaküla, Patrick Plötz, Francisco J. (Fran) Márquez-Fernández, Lina Nordin and I concluded that e-retrofits are the single most impactful pathway to speed up the reduction of greenhouse gas emissions from European road transport. The same analysis indicated that a pan-European ERS network would increase the probability that such e-retrofits gain mass-market appeal. While last week's report acknowledges this potential, the potential for ERS to facilitate e-retrofits needs far more attention than it has been given so far.

What do other researchers say?

Readers interested in what studies by other researchers and from other countries have concluded may want to check the various links throughout this article, or take a look at this ongoing crowdsourced inventory of ERS research.

Thank you for reading. Please share your own thoughts and comments below.