As pretty much everyone who has dared to peek from beneath the covers understands, climate change is a clear and present danger. And one of the few bright spots for those following the race to save the planet from climate disaster is the promise of rapid electrification — in particular, the electrification of transportation using renewable sources of energy.

The transportation sector does, indeed, seem to be in the midst of a transformational shift from the fossil-fuel-dominant technology that has powered mobility for the past century. And since the sector accounts for 29 percent of U.S. carbon emissions (and 24 percent worldwide), it’s only natural that electrification of the vehicle fleet, paired with the rapid greening of electricity production, is widely viewed as key to containing climate risk.

Accordingly, many countries have set ambitious targets for electric vehicle adoption. Germany, for example, aims to have six million electric vehicles on the road by 2030. Others have pledged to phase out the sale of internal combustion engine (ICE) vehicles entirely — among them, France and the UK (by 2040), China (2035), India (2030), Norway (2025) and, closer to home, the state of California (2035).

The commitment of these governments to speed this transition is reflected in a host of initiatives to encourage consumers to “go electric” — and, less frequently, to discourage consumers from purchasing ICE vehicles. As one example (on the carrot side of the equation), President Biden’s American Jobs Plan envisioned an investment of nearly $175 billion dollars to speed the adoption of electric vehicles through a combination of consumer subsidies, grants to build charging stations, incentives for automakers and government EV purchases.

If you’ve been keeping up on green news, all this is probably familiar. But what follows may not be: there has been little careful analysis of the long-term consequences of these plans — in particular, whether current policies to electrify surface transportation at warp speed will get us where we need to go in a cost-effective manner, and whether the haste will close off alternative scenarios for climate change mitigation.

The Private Economics of Electric Vehicles

Start with some basics. Conventional wisdom holds that EVs are more expensive to build and less expensive to operate than gasolinepowered cars. The reality is a bit hazier. Presently, EVs do face a manufacturing cost disadvantage that’s reflected in higher sticker prices. And while battery costs have fallen by over 85 percent over the past decade, the upfront costs of electric vehicles remain higher than ICE alternatives. This gap is closing, although there is considerable disagreement about when EVs might reach cost parity. On the one hand, the National Academies of Sciences, Engineering and Medicine projects cost parity later this decade. On the other, automakers question whether parity (with regard to both cost and range) will be possible within 15 years.

In exchange for higher upfront costs, electric vehicles offer advantages over ICEs arising from radically different power trains. EVs offer superior torque and acceleration, drive more smoothly (no gear changes) and quietly (no combustion to provide power), and offer owners the option to charge at home. Yet dependence on batteries introduces unappealing features as well. The driving range of EVs is typically shorter than the range offered by ICEs, and, even with fast-charging infrastructure, batteries take far longer to charge than it does to fill a gas tank. In fact, the marketing imperative to increase the range of EVs may partially offset declining battery costs, and improvements to charging speeds might require innovation beyond the current generation of battery technologies.

EVs are, as widely billed, cheaper to fuel than ICE vehicles, a reality reflecting both the relative efficiency of the electric power train and the relative prices of gasoline and electricity. But these savings do vary quite a bit across place and time. Retail electricity prices differ substantially, depending in part on the electricity utility’s underlying costs and, in part, on regulators’ appetite for recovering the utility’s fixed costs through charges on sales. For instance, the top marginal retail electricity price for residential customers in PG&E’s service territory in the Sacramento area is a whopping 42 cents per kWh — almost four times higher than the 11.4 cents rate for households buying their power from Sacramento Municipal Utility District.

Similarly, gasoline prices, which are linked to the global price of crude oil, are also volatile. Since 2005, the average gasoline price in the U.S. has fluctuated between $2 and $4 a gallon. Across states, gasoline taxes vary from roughly 35 cents per gallon (in Alaska, Missouri and the Gulf Coast states) to roughly 80 cents per gallon in California.

All this translates into substantial variation in the operational cost savings offered by electric vehicles. Using data from 2016 as a snapshot, the figure on page 17 highlights the state-level variation in operational cost savings from driving an electric Nissan Leaf rather than a similarly sized ICE-powered Nissan Versa.

Unsurprisingly, EVs reduce operational costs by over 50 percent in the Pacific Northwest, where (usually) abundant hydropower provides low-price residential electricity. On a per-mile basis in the Pacific Northwest, a Leaf costs 4 to 5 cents per mile less to drive than the Versa, adding up to a savings of roughly $500 per year for a typical household. But savings are much lower in other parts of the country — most notably, in California and New England — where residential electricity rates are substantially higher.

Finally, because electric vehicles have fewer moving parts, they generally enjoy lower maintenance costs. The electric power train obviates the need for the maintenance required by the stresses associated with internal combustion. A recent study examining the operations costs of the federal vehicle fleet concluded that the ongoing maintenance costs of a fully electric power train are roughly 60 percent of the ongoing maintenance costs of a vehicle that either fully or partially relies on burning fuel.

But less is known about how maintenance costs stack up over the life of an electric vehicle. Compared to conventional vehicles, for which we have a century of experience, the relatively short history of electric vehicle ownership creates more uncertainty around maintenance costs arising from the eventual need for battery replacement. As a point of reference, the average EV battery costs $137 per kWh of capacity. At this price, a 30-kWh battery (roughly the size of the relatively small Nissan Leaf battery) would cost $4,000, while a 75-kWh battery for the Tesla Model 3 would cost more than $10,000 — considerable expenses even if battery costs continue to decline.

Failures in the Market for EVs

As you may remember from Econ 101, when firms and consumers do not pay the full societal costs or reap the full societal benefits of their decisions, markets tend to produce lessthan- optimal outcomes in terms of efficiency. So when considering how government can engage productively in the market for EVs, it helps to understand what factors are in fact “external” to private decision makers — and thus whether and how government intervention can increase economic efficiency. For electric vehicles, these external benefits and costs arise from two sources: usage of an electric vehicle rather than an ICE vehicle (the “intensive” margin) and the number of EVs on the road (the “extensive” margin).

Usage-Based Externalities

The externality that comes most readily to mind when thinking about the use of an electric vehicle is air pollution. Global greenhouse gas (GHG) emissions, local pollution and traffic congestion are typically present to some degree for both internal combustion and electric vehicles. To the extent EVs produce fewer negative externalities (or more positive ones) than ICE vehicles, government may want to deploy policies that encourage EV adoption and use rather than use of ICE vehicles.

Emissions from ICE vehicles are easy to understand. Combustion of fossil fuels creates GHGs and local pollution. But the damages a given amount will do varies with timing and location, depending on the pollutant. GHGs are widely diffused in the atmosphere, and the most important GHG (CO₂) takes millennia to be reabsorbed. Local pollutants, on the other hand, typically contribute to adverse health outcomes only to people directly exposed to them, with concentrations dissipating in days or hours.

The emissions created by electric vehicles are more elusive to measure. The electricity generation sector, of course, depends on a variety of technologies: wind, solar, hydro, nuclear and various fossil-fuel technologies. The pollution created by charging an electric vehicle is the pollution created by the marginal source of electricity — the generator that “turns on” or ramps up production when demand increases.

A host of studies bearing this in mind offer estimates of the relative emissions damages associated with driving an EV instead of an ICE — and, happily, the conclusions of these studies are qualitatively similar. EVs tend to provide some GHG savings relative to ICEs in areas where natural gas is on the electricity generation margin, but tend to be more GHG-intensive if coal is on the margin and temperatures are cold.

One 2015 study estimates that a Nissan Leaf EV generates roughly $425 worth of lifecycle GHG savings when driven in California instead of its ICE-powered near-twin, the Nissan Versa. That translates into a 20 percent life-cycle GHG savings in California, compared to 5 percent nationwide and -10 percent (that’s right, negative 10 percent) in the Midwest, where fossil fuels dominate electricity generation at the margin. As the electricity sector moves away from coal and natural gas and toward renewables (potentially with storage that allows round-the-clock use), the GHG profile of electricity supply will improve.

Most damage from local pollution comes in the form of impaired respiratory health, but it has also been linked to decreased labor productivity and impaired cognitive performance. Another study compares the local health implications of EVs relative to ICEs. Not surprisingly, large benefits accrue to driving EVs in cities where the ambient air quality is poor and where the electricity grid is powered by relatively clean electricity (e.g., Los Angeles). But, where coal is likely to be on the electricity generation margin and is upwind of population centers, driving EVs can actually increase net local pollution. Encouragingly, recent estimates suggest that local pollution from electricity production is declining. And as the electricity grid continues to become cleaner, the local air pollution impacts of EVs will decline.

A bit of perspective here: congestion externalities and accidents are the largest market failures related to driving, far exceeding damage from GHGs and local pollutants. And replacing an EV with an ICE should have little overall impact on these externalities. Perversely, however, some policies designed to encourage EV adoption, such as single-occupancy access to carpool lanes and electric vehicle purchase subsidies, might add to congestion externalities by reducing the utility of these lanes or adding to the total number of cars on the road.

Then there are the proverbial wheels within the actual wheels. There is growing evidence of “portfolio effects” within a household’s set of vehicles. If EV buyers are inclined to purchase larger vehicles in an attempt to diversify the vehicles they own, incentives to encourage EV adoption may increase the dispersion of the vehicle weight distribution and increase accident externalities by mixing more heavy vehicles with light ones on the road. Not only that, but if EV adopters also diversify their portfolio with large vehicles (i.e., low fuel economy vehicles), using the large vehicles might offset some of the air pollution benefits offered by the electric vehicle.

Stock-Based Externalities

The growing size of the EV fleet and the number of charging stations might also generate external benefits. Consider the prospect for “network effects” from charging infrastructure, whereby the increasing availability of charging stations stimulates demand for EVs and vice versa. Network effects don’t automatically justify government intervention, however. To the extent that the benefits are captured by market participants alone, the market will still allocate the efficient amount of capital to building charging stations. And even when benefits are external, there are undesirable consequences to overstimulating charging infrastructure supply too soon. So to understand the degree to which government intervention adds to efficiency, one must first understand the workings of the particular network.

Start with the fact that charging infrastructure takes different forms. Level 2 chargers supply 220 volts and can fill batteries at a rate of up to nearly 20 kW per hour, though most max out far below that pace. The fixed cost associated with installing these stations is modest (on the order of a few thousand dollars), as are the adaptations that power utilities must make to accommodate 220- volt chargers. That explains why Level 2 chargers are already commonplace — some 57,000 of them have been built in California. They are typically found in parking lots, commercial and retail locations and multi-unit apartment buildings.

By contrast, Level 3 stations cost on the order of $100,000 and can provide a nearly full charge in under 30 minutes. These fastcharging stations are more likely to be found along highways to facilitate larger numbers of EVs making long trips.

Now, network effects are notoriously hard to measure. But economics does provide some guidance as to the conditions under which they are likely to justify government intervention in the market for EVs. At the risk of repeating ourselves, the presence of network effects does not necessarily imply a market failure. These effects must be external to firms and consumers — that is, third parties must enjoy some of the benefits from the growth of the system. So, while network effects are undeniably present in the EV setting (EV buyers benefit from easier access to chargers and charging networks benefit from having more EVs on the road to service), these benefits may be fully appropriated by market participants.

Here, a distinction between Tesla and everyone else is warranted. Tesla’s network of fast-charging stations provides a unique benefit to buyers of Tesla vehicles, which increases consumers’ willingness to pay for Teslas. Tesla has every incentive to build out its charging station network optimally, since the company can charge a higher price for their cars as a result. However, network externalities likely exist when the many other EV makers share the same charging technology.

There are currently three “standards” for fast-charging connectors adopted by EV automakers: CHAdeMO, Combined Charging System (CCS) and Tesla. (Asian automakers have generally used CHAdeMO, while U.S. and European automakers have used CCS.) In other contexts — for example, with voltage requirements for appliances — governments have long played a role in pushing incompatible standards toward alignment.

With EVs, makers have incentives to design vehicles that are either compatible with a prevailing standard or to offer adapters to allow their vehicles to be used across multiple charging platforms. The version of the Tesla Model 3 sold in Europe, for example, is built with a CCS connector allowing owners to take advantage of the existing fast-charging network.

There may well be multiple ways in which market forces are at play, which has implications for efficiency. Consider a hypothetical example where a single company profitably placed sufficient chargers along a route to attract and serve, say, 1,000 vehicles a day. An alternative equilibrium may exist with a far higher number of charging stations and 10 times as many EVs using the chargers. A market failure exists if society as a whole would be better off at this alternative equilibrium — and if a hurdle exists that inhibits the free market from reaching that alternative.

We don’t know whether such a high-EV equilibrium exists. Moreover, there is little empirical guidance to inform us as to whether society would be better off with the denser network of EVs and chargers. The environmental benefits would likely be higher, but the cost to build out the charging network and to adapt the electricity grid to much higher demand at peak periods would also be high. The relative magnitude of these costs and benefits is uncertain.

One often-ignored source of uncertainty here comes in the potential to foreclose other technological approaches to climate change. Over-building infrastructure today will reduce the expected benefits of exploring alternative paths, ranging from hydrogen power to ethanol fuel from grasses — or even reenvisioning the role of cars in the urban environment.

Separating the Signal From the Noise

For better and worse, the transportation, environmental and energy policy landscapes are interconnected, and often in ways that yield unanticipated effects. As described above, externalities arise when markets do not internalize the full social costs (or benefits) of the decisions they make. We highlight three themes common across these externalities.

First, an externality only arises when the benefits of a buyer’s or seller’s decisions accrue to others. If the decision makers reap what they sow, there is no externality-based market failure and no efficiency-based justification for a production or sales subsidy.

Second, where the externalities from switching to EVs are uniform across time and space, a one-size-fits-all approach to policy might be appropriate. But many externalities related to EVs have external costs or benefits that vary by location. Local pollution depends on how electricity is generated and how close the facilities are to population centers. By the same token, network externalities depend on the confluence of EVs and charging stations. In both cases, decentralized policy will better address local externalities created by EV fleet expansion.

Finally, some of the externalities discussed above will change, as will optimal policy. Where external benefits are likely to exhibit diminishing marginal returns (as in the case of network effects), it would probably be best to front-load subsidies when they would affect efficiency more. But for externalities related to, say, local or global pollution, where present benefits of EVs might be modest (or negative where electricity is generated with fossil fuels), it might make sense to back-load subsidies to kick in when electricity production is largely or entirely in renewables.