WHETHER they want to or not, carmakers are having to rush out all manner of zero-emission vehicles (ZEVs)—ie, plug-in electric vehicles that use either rechargeable batteries or hydrogen fuel-cells to drive the wheels. Three years ago only two plug-in electric vehicles—the Nissan Leaf and the Chevrolet Volt—were available in America, and even then in only a handful of states. Today, no fewer than 20 different models are on sale across the country.

The sudden flurry of activity stems from the tighter emission standards expected shortly from the Environmental Protection Agency (EPA). In the past, states could chose whether to accept EPA’s own emission requirements or adopt tougher ones promulgated by the California Air Resources Board (CARB). The latter were subsequently adopted by a dozen or more other states. Carmakers wanting to do business in any of these CARB states—which account for over a third of new car sales in America—opted reluctantly but sensibly to meet the stricter Californian standard.

However, from the 2016 model year onward, the EPA will adopt CARB’s latest, even tougher, requirements for the rest of the country. When the new rule kicks in, carmakers across the country will then have to sell at least one ZEV for every six petrol or diesel vehicles they move off the lot. Failure to do so will mean having to acquire pollution credits from others that have already met the target, or pay a hefty fine.

Between them, the 14 CARB states plus the District of Columbia have offered financial incentives to encourage motorists to trade up to ZEVs. California has long provided a tax credit worth $2,500, on top of an additional $7,500 from the federal government. With the coming stricter national requirement, officials expect to see 3.3m ZEVs on the road by 2025, up from less than quarter of a million in the handful of states where they are available today. Others think there will be many more ZEVs around long before then.

But will they make any difference to the air quality? Difficult to tell. While ZEVs have no greenhouse gases puffing out of their tailpipes (if they have one), there are still emissions associated with the extraction, production and distribution of the fuel that generates the electricity in the case of battery vehicles, or the hydrogen for fuel-cell vehicles. In short, ZEVs trade pollution at the tailpipe for pollution at the power station. The question comes down to whether ZEVs in general—and battery electrics in particular—produce less carbon dioxide overall than petrol or diesel cars? As they say, it all depends.

For instance, it depends on what type of car is being replaced by an electric vehicle. If it is an eight-cylinder clunker that is about to be scrapped anyway, then indubitably so. But if the buyer is choosing between an electric vehicle and a new petrol or diesel car of comparable size and weight, then which will produce less greenhouse-gas emissions is moot.

Which is the cleaner also depends on where in the country the comparison is made. The electricity used to charge a battery vehicle can come from a clean source—such as hydro-electricity, nuclear power or even renewables like wind and solar energy. Or it can come from bituminous coal—still the leading fuel for power generation in America, accounting for 39% of the country’s electricity. However, thanks to the recent abundance of cheap natural gas from fracking operations, king coal is fast losing its crown as the fossil fuel of choice for generating electricity.

The mix of energy sources used to generate electricity varies considerably across the country. Of the three grid interconnections (self-contained markets, within which electricity is traded), the eastern one has some of the filthiest power stations, thanks to the preponderance of coal mining in Wyoming, West Virginia, Kentucky and Pennsylvania. The eastern interconnection, stretching from the Rockies to the Atlantic, also generates four times more electricity than the western interconnection, and ten times more than the Texas interconnection. The cleanest electricity in America comes from the west, where major rivers are harnessed and renewable wind- and solar-farms dot the deserts. California is well on its way to producing a third of its electricity from clean, renewable sources by 2020.

Finally, the cleanliness, or otherwise, of an electric vehicle depends on what time of the day it is recharged. Plugging into the power supply during the peak hours of the early evening causes proportionately less carbon dioxide to be emitted. That is because every scrap of generating capacity—from squeaky clean nuclear, hydro and wind power to dirtier gas turbines and filthier still coal-burning power stations—is pressed into service to meet peak demand. Adding clean renewable power to the mix reduces the overall amount of carbon dioxide emitted per kilowatt-hour (kWh) of electricity generated.

Peak electricity may be cleaner overall, but it is also a good deal more expensive than off-peak power. Besides, plugging into the grid at peak hours is just about the last thing utilities encourage when the public is arriving home from work, turning on the heating or air-conditioning, starting to cook supper and watching television. Indeed, utilities, as well as makers of electric vehicles, urge owners not to plug their vehicles into the grid until after midnight, when the base-load power stations alone can produce ample electricity at rock-bottom rates.

Unfortunately, base-load electricity produced throughout the day and during the night tends to come mainly from coal. The cleaner load-following power plants that operate gas turbines are usually switched off after peak demand for electricity has subsided in the evening. Thus, the most popular time to recharge a battery vehicle is when electricity is at its dirtiest.

How, then, to determine which is the cleaner: an electric vehicle recharged with electricity generated (most likely) from coal or (if lucky) from natural gas or (better still) from nuclear power; or a conventional vehicle with an internal combustion engine that burns hydrocarbon fuel made from oil? And how to assess the difference in a way that captures all the emissions and energy losses between the coal mine and the electric vehicle on the one hand, and the oil well and the hydrocarbon vehicle on the other?

There have been numerous attempts to do just that, none wholly satisfactory. In most instances, researchers have used the comprehensive GREET methodology developed over the past two decades by the Argonne National Laboratory near Chicago. This can analyse the well-to-wheels energy losses and greenhouse-gas emissions for more than 100 different fuel pathways. As comprehensive as it is, though, the framework is designed primarily to handle fuels for transport, not power generation. And although it includes coal among its spectrum of energy sources, the pathways involved focus on the production of such fuels as hydrogen, diesel and methanol for the transport sector.

The biggest pollution problem with coal, even before burning it, is the amount of methane released during mining. As a greenhouse gas, methane has 23 times the global-warming potential of carbon dioxide. This is rarely, if ever, taken into account when computing the environmental impact of running coal-fired power stations around the clock. Much the same goes for comparable emissions produced by freight trains hauling coal across the country, or pulverising it at power stations. True, such effects are dwarfed by the sheer quantities of carbon dioxide released during combustion, which is responsible for a quarter of all greenhouse-gas emissions in America.

The Union of Concerned Scientists (UCS), a non-profit advocacy group based in Cambridge, Massachusetts, has recently updated its previous “State of Charge” study that compared emissions from electric vehicles with those from their hydrocarbon-fuelled counterparts. According to Don Anair, UCS’s research director for clean vehicles, battery cars—no matter where in the country they are recharged—produce lower greenhouse gas emissions overall than the average new car.

As with all analyses, however, the devil is inevitably in the details. And the trouble with such claims is that the details are masked by the use of averages. An analysis that avoids this trap has recently surfaced following painstaking work done by three economists affiliated with the National Bureau of Economic Research. Published last March, the study by Joshua Graff Zivin of the University of California at San Diego, Matthew Kotchen of Yale University, and Erin Mansur of Dartmouth College focuses on emissions caused by marginal increases in electricity demand, in different parts of the country, at different times of the day.

Unlike an averaging approach, the marginal analysis seeks to determine in each case the additional amount of carbon dioxide produced when one further electric car is plugged into the grid. Here, the change in emission depends on which individual power station provides the additional electricity—ie, the plant on the margin. That plant can be anywhere within the interconnection market where electricity is traded.

While the study shows that the marginal effect of charging an electric car, averaged across all regions of the country and hours of the day, is 1.21 pounds (550 grams) of carbon dioxide per kilowatt-hour of electricity consumed, the researchers found substantial differences between locations and times of the day. For instance, the marginal effect in the upper Midwest was almost three times greater than that for the western part of the United States. And for some hours of the day, the differences were even greater.

Applying this methodology to the real world, your correspondent reckons that recharging a Honda Fit EV’s 20 kWh battery pack from scratch in the early hours of the morning produces 16.6 pounds of carbon dioxide west of the Rockies, 21.8 lb in Texas, and 50.6 lb in the upper Midwest. Thus, with a range of 82 miles, the Fit causes the emission of an additional 0.202 lb/mile in the west, 0.266 lb/mile in Texas, and 0.617 lb/mile in the upper Midwest.

A regular Honda Fit equipped with a 1.5-litre petrol engine and a continuously variable transmission produces 0.534 lb/mile of carbon dioxide from its tailpipe—making it dirtier than its battery equivalent in the renewable-energy west and even the gas-fired south, but cleaner by a considerable margin in the coal-burning north. Other electric vehicles and their petrol or diesel counterparts will have a different pattern of pollution. As Dr Graff Zivin and his colleagues point out, though, plug-in electric cars generate more emissions per mile in the upper Midwest than the average car currently on the road. In most other places—and if charged during the day rather than the night—they can be cleaner all round.

Readers wishing to see how clean, or otherwise, is an electric car in their neck of the woods can enter their zip code and vehicle of choice into the EPA’s online calculator. Most will be pleasantly surprised. But forget it if you live in Minneapolis or many other parts of the Midwest.