Last week, we took a look at the role incentives can play in encouraging people to buy electric vehicles (EVs). Today, we bring you a paper from the National Bureau of Economic Research that attempts to calculate the environmental benefits of EVs versus conventional vehicles in light of those subsidies. Is it as desirable to encourage EV use in a state where the electricity comes from burning coal as it is in a state where that electricity comes from natural gas or nuclear power?
The authors, four economists from the University of North Carolina (UNC) Greensboro, Dartmouth College, Middlebury College, and UNC Chapel Hill have created what they describe as “a powerful and unprecedented modeling framework for analyzing electric vehicle policy.” They do this with three different components. First, a model of consumer choice between EVs and gasoline-powered cars. Next, they incorporate the effect of EV charging on air pollution from individual power stations. Finally their model compares the emissions from these power stations with the emissions internal combustion vehicles would produce at the same location.
The analysis uses some quite complicated formulae to calculate the damages that result from emissions per mile from 11 different battery EVs on sale in 2014, compared to the closest internal combustion engine-powered equivalent, independent of price. Where possible they’ve compared like models, so the EV Ford Focus vs a regular Focus, a Fiat 500e vs a regular Fiat 500, and so on. For cars where there isn’t a conventional model (Nissan Leaf, Mitsubishi i-MiEV, Tesla’s Model Ss) the authors picked cars they believed were equivalent in features (Toyota Prius, Chevrolet Spark, BMW 7-series). Then they compared the EVs’ kWh/mile rating with the gasoline cars’ fuel economy, as well as pollution from nitrogen oxides, sulfur dioxide, small particulates, and volatile organic compounds.
EPA city and highway mileage figures are used to calculate the effects of gasoline vehicles in urban and rural areas (their model goes down to the county level). For EVs, the authors start with EPA’s MPGe figures and then adjust this for the temperature profile for each county. Then they factor in the amount of each of the pollutants listed above at each of 1,486 power stations across the country per kWh of electricity (the data is from 2010 to 2012). Those pollution estimates then get modified again by an assumed daily charging profile to calculate the emissions per mile of each power plant for any given county in the US.
The US electricity grid consists of three main regions (East, West, and Texas), which the authors further split into nine smaller regions as defined by the North American Electricity Reliability Corporation (NERC). Apparently there isn’t much transmission of electricity between the main regions, but the authors assume that within the nine NERC subregions, the pollution from an EV charging will be the same regardless of the county. Finally, monetary values for the social costs of pollution come from the EPA for carbon dioxide and the AP2 model for local pollutants. And in case that wasn’t enough, they weighed the statistics for vehicle miles traveled in each county to get a sense of how important driving distances are.
The result of all this complex mathematics is that, outside of a number of Californian and Texan cities, driving an EV may result in more damage from pollution than driving an equivalent conventional car. In Los Angeles, which has a lot of traffic and which benefits from relatively clean electricity, an EV is the right choice, they argue. Alternatively, the rural midwest is the opposite story, since the low population density means comparatively little air pollution from traffic, but electricity comes from lots of coal power stations. But even Chicago and New York fare badly under their model, despite both cities paying a hefty price from conventional traffic pollution.
The paper also suggests that EVs export pollution across a much larger geographical area than a gasoline vehicle. This is something that local governments don’t take into account with subsidies in their view, and that the hefty EV subsidies in states like Georgia ($5,000 per EV) are based on an incomplete picture of the problem.
The conclusions are sure to be welcome news to EV skeptics; if recent discussion threads for EVs here at Ars are anything to go by, there are plenty of people out there who want to see battery-powered vehicles fail. However, there are quite a lot of assumptions made in the paper. The estimates are based on traveling 15,000 miles a year (24,000km), which is fifty percent greater than the current average for cars. And as they note, the data for power stations is now several years old, and electricity generation is becoming ever-cleaner in the US. They also haven’t calculated the various impacts from mining or extracting the fuel for power stations and cars or the materials for batteries, nor the benefits to encouraging adoption of EVs now to spur the market to develop and refine powertrains and battery technology, something we’re definitely seeing happen.