Some time back, the Wall Street Journal ran a Bjorn Lomborg opinion piece about electric cars. Lomborg is an academic and activist who founded the Copenhagen Consensus, an organization dedicated to improving global welfare and the use of cost-benefit analysis to determine the most important problems to address, and the best methods to use. Lomborg has been attacked by many climate scientists for his position that global warming is happening and is man-caused, but is not among the most critical issues the planet faces. In the WSJ piece, Lomborg argues that because battery-electric vehicles (BEVs) may not be as low-carbon as many people believe, government subsidies for them are bad policy.
Timothy Taylor at the Conversable Economist points us to the original research piece (PDF) on which Lomborg bases his argument. One of the big caveats in the article is that the source of electricity has a very large effect on how much carbon is emitted to power a BEV. Electricity generated from coal is, rather obviously, among the worst sources in terms of carbon. An electric vehicle operating in Ohio is largely coal-powered; one running in Idaho, OTOH, runs largely on low-carbon hydro power. Tim also makes a point that had occurred to me while reading Lomborg's piece: there are considerations other than just how much carbon is emitted over a vehicle's total life cycle.
As it turns out, I agree with Lomborg in general that there are things we should be spending money on first with respect to energy use patterns. In the transportation arena, electrified light rail in metro areas and moving long-haul freight out of diesel trucks onto much more efficient diesel trains are two of them (and electric freight trains might be even better). These are areas where the federal government should have a significant role in undoing the enormous shift of transport to roads since World War II, just as it had a significant role in supporting that shift to roads in the first place . Those are changes that should be applied broadly. There can be, however, local and regional reasons for supporting BEVs that make sense now. They're just not necessarily reasons having to do with CO2 .
CO2 is a long-lived pollutant and the atmosphere does an excellent job of mixing it uniformly across the planet. But CO2 isn't the only pollutant. A BEV powered by coal-fired electricity still has the potential to accomplish two important local goals: space- and time-shifting of the non-CO2 pollution. The space-shifting is pretty obvious: a gasoline-powered car being driven downtown emits its ozone precursors downtown; a BEV's emissions occur at a coal-fired plant well away from the city center. That can have an enormous effect on the quality of the air in the crowded parts of the region. It may also make it easier to reduce emissions of ozone precursors such as nitrous oxides, because it's easier to install and maintain pollution controls on the single power plant than on 100,000 cars.
The benefits of time-shifting might not be quite so obvious. It seems a safe assumption that most BEV charging, at least in the near future, will occur overnight at the owner's home. Electricity is very much an on-demand thing, so the pollution created by that charging also occurs overnight. Stretching the pollution emission out over a longer period of time may be beneficial. Some areas may experience an additional benefit. I live in the Denver metro area. While the region has made enormous strides in cleaning up the infamous Brown Cloud, early-morning winter temperature inversions (coinciding with the morning rush hour) can trap pollutants close to the ground and create ozone and other problems. Shifting the pollution creation temporally out of the rush hour offers a benefit in addition to shifting it spatially out of downtown.
 Anecdotes are not data, but... I drive across parts of Colorado, Wyoming, and Nebraska on the Interstate Highway System at least once a year. I-80 in western Nebraska and eastern Wyoming is, in practice, an expensive and poorly run railroad with big trucks acting as inefficient single-car trains.
 And yes, I know that this entire post takes a very parochial view. From a pure global cost-benefit perspective, the US (along with Japan and Western Europe) ought to spend enormous amounts of money solving problems in poorer parts of the world. Adding clean electric generating capacity to the grid in Pakistan, for example, would produce much larger global benefits than cleaning up a power plant in the US. But that sort of trade-off is unlikely to be made on any meaningful scale, because most of the people who live in developed countries aren't that interested in solving global problems.