Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.


Infrastructure killed the electric car

When prices are adjusted for quality, electric vehicles stood their ground to petrol cars in the early twentieth century United States. If the electricity grid had developed twenty years earlier, they might have reached a 68–79% market share and CO2 emissions per car could have declined by 60%, a new study finds.

Within the diffusion of innovations, a long-running debate has been what comes first, the product or the system supporting it: ocean-going vessels or modern harbours, appliances or the grid, cars or hardened roads, planes or airports. This timing concern is particularly salient for technologies that require new infrastructure to achieve mass-scale adoption, such as electric vehicles (EVs) that need a network of charging stations. Now, writing in Nature Energy, Josef Taalbi and Hana Nielsen address the issue by measuring the role of infrastructure in resolving the combustion engine versus EV automobile technology competition1 at the beginning of twentieth century, and find a significant contribution of delayed grid development in favouring the uptake of combustion engine vehicles.

Credit: incamerastock / Alamy Stock Photo

In 1900, cars based on steam, electricity and petrol technologies accounted for 40%, 38% and 22% of the 4,200 cars produced1. In 1905, over 50% of commercial vehicles were electric2. Conventional wisdom has it that the gasoline car won because it was more efficient, and thus lower-priced. Carefully analysing the historical data, Taalbi and Nielsen cast doubt over this explanation and compel us to look deeper.

The researchers gathered and analysed large datasets, including details of over 37,000 different passenger car models introduced between 1895 and 1942, and linked data at the county level on electricity and gasoline infrastructure availability, roads and household incomes.

The study investigates whether EVs were outcompeted due to the better performance and lower price of petrol cars or, if not, what could explain the demise of EVs by 1920. It finds that average petrol car prices were never lower than 85% of EV prices. When adjusting for quality, EVs were not more expensive than petrol cars until 1910, but were aimed more at urban consumers.

The question remains what then explains the downfall of passenger EVs. An important factor was the availability of electricity infrastructure. Taalbi and Nielsen first studied the impact of infrastructure development on technology selection for automobile development. They then built a counterfactual that shows that 64–79% of models introduced by the 1920s would be electric if the 1922 diffusion rate of the US grid would have been reached by 1902. Though electricity might not have won all-out, EVs probably would have been the top choice for urban trips, while petrol cars would still be used for long-range driving3.

Taalbi and Nielsen’s findings are consistent with present-day studies that show that infrastructure development is important for EV adoption and retention decisions4. Increasing the number of charging points, for example, is an essential part of the UK government’s recent £2.8 billion push for EVs5.

Taalbi and Nielsen also tentatively estimated how pollution would have been affected if EVs won. They suggest CO2 emissions per car would have been reduced by 0.4 kilograms per mile in 1920 (to 42% of a petrol car’s emissions) and total emissions would have been reduced by 20 million tons.

Of course, it is not clear if, at the time, better decisions on infrastructure development could realistically have been made. Social scientists sometimes distinguish three degrees of this ‘path dependence’ problem: the decision does not matter or if it does, an optimal decision is made; the decision leads to regrettable outcomes but these were unknowable initially; the decision leads to regrettable outcomes that were known and remediable6,7. Considering pollution, the electricity infrastructure was probably somewhere between the second and third degree of path dependence. CO2 emissions and other exhaust gases were likely to be seen as less of a problem at the time, especially given the enormous city-centre pollution by horse manure. Yet, at the system level, some foresight might have been possible. A modern petrol car generates about the same CO2 emissions per mile as an EV in 1920, yet it took us a lost century to get to that level8.

One does wonder, however, how feasible a faster spread of electricity grids could have been. Electric utilities had the second-fastest productivity growth of all US industries between 1899 and 1941, at 5.0% per annum for 42 years, followed by cars with 4%. Electric machinery achieved 2.0% (ref. 9). These rates at which outputs grew faster than inputs enabled similar price falls. They are stellar compared to present-day productivity growth rates and suggest that these were highly dynamic industries that could have responded well to incentives, such as earlier stimulation of infrastructure diffusion, or the removal of impeding factors.

One such impeding factor was the famous ‘war of the currents’ (DC versus AC) between Edison and Westinghouse in the late nineteenth century, and the subsequent patent pool. This returns us to the question of what came first — the product or the system? AC won, but not totally: the last DC power station in New York City, for example, was only closed in 2007. Edison and Westinghouse might have come to an agreement much earlier, if they had realized how important fast adoption was for electricity applications that faced competition from other technologies.

Future studies could explore whether the same results hold for other countries historically, and for other competing technologies, such as zeppelins versus passenger planes, steam versus diesel trains, steam versus diesel ships, bicycles versus scooters, or ferries versus hovercrafts.

There are many challenges encountered in this kind of historical research: the farther back in time we go, the sparser the data become and the more inventive we need to be to test hypotheses. Taalbi and Nielsen’s study shows ways of doing this and highlights how relevant infrastructure development can be for EV adoption, how far-reaching and long-lived the consequences of decisions 120 years ago can be for the present, and how impactful our present decisions can be for the future.


  1. 1.

    Taalbi, J. & Nielsen, H. Nat. Energy (2021).

  2. 2.

    Mom, G. The Electric Vehicle: Technology and Expectations in the Automobile Age (Johns Hopkins Univ. Press, 2004).

  3. 3.

    Kirsch, D. A. The Electric Vehicle and the Burden of History (New Brunswick, 2000).

  4. 4.

    Tran, M., Banister, D., Bishop, J. D. K. & McCulloch, M. D. Nature Clim. Change 2, 328–333 (2012).

    Article  Google Scholar 

  5. 5.

    Johnson, B. Now is the time to plan our green recovery. Financial Times (2020).

  6. 6.

    David, P. A. Am. Econ. Rev. 75, 332–337 (1985).

    Google Scholar 

  7. 7.

    Liebowitz, S. J. & Margolis, S. E. J. Law Econ. Organ. 11, 205–226 (1995).

    Google Scholar 

  8. 8.

    US Environmental Protection Agency, Office of Transportation and Air Quality. Greenhouse Gas Emissions from a Typical Passenger Vehicle (EPA, 2018);

  9. 9.

    Bakker, G., Crafts, N. & Woltjer, P. Econ. J. 129, 2267–2294 (2019).

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Gerben Bakker.

Ethics declarations

Competing interests

The author declares no competing interests.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Bakker, G. Infrastructure killed the electric car. Nat Energy 6, 947–948 (2021).

Download citation

  • Published:

  • Issue Date:

  • DOI:


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing