Blog Post

Subsidies for EVs: Good Policy or an Unnecessary Handout?

Jul 25, 2013 | Virginia D. McConnell

In this series of blog posts, RFF researchers Virginia D. McConnell and Joshua Linn take a look at the current state of the electric vehicles (EVs) and the effect of current and future policies on the market. Click to read the first, second, third, and fourth installments.

It is clear from our earlier blog that electric vehicles (EVs) are still expensive, that the EV subsidies are large, and that sales are strongly influenced by them. But are there good economic arguments for such subsidies? Standard economic policy analysis holds that where there are “market failures,” government policy can indeed improve social welfare.

Two market failures are most relevant for EVs: greenhouse gases (GHGs) and spillovers (we leave aside market failures relevant to all vehicles such as traffic congestion and accidents). Regarding the first: there are external costs of owning and using vehicles of all types that are not accounted for in vehicle and fuel markets. GHG emissions from gasoline use and carbon-based electricity production contribute to climate change and its associated damages. Gasoline and electricity policies have not historically accounted for such damages.

In effect, the policy decision has been made to address this externality and reduce GHG emissions through increasingly strict CAFE rules, which will cause vehicles to become more fuel efficient and emit fewer GHGs over time. An alternative policy for addressing this market failure is to increase gasoline taxes, which most economists would prefer, but which has proven to be less politically popular.

The CAFE standards set binding GHG limits that must be met regardless of the number of electric vehicles that are sold. Subsidies to EVs change the makeup of the fleet but—if power sector emissions were properly credited—would not affect the average emissions rate of new vehicles sold (more EVs, but less fuel economy improvements from other vehicles). But, because of the preferred treatment for EVs discussed in the last blog, subsidies to EVs could result in higher GHG emissions from new passenger vehicles. Just how much depends on electricity sector policy. If there were a binding cap on electricity sector GHG emissions, EV subsidies would result in more EVs on the road and the associated higher electricity use, but overall electricity sector emissions actually would not increase. Although regional GHG caps exist, the US as a whole does not have such a policy. Instead, there are many local, state, and federal policies promoting renewables. And EPA is preparing GHG regulations for new and existing power plants. With these policies in place along with CAFE, EV subsidies probably increase average GHG emissions rates.

So, it would appear that in the presence of stricter CAFE standards, EV subsidies over the next few years cannot be justified on the basis of reducing GHG emissions. But, if there are advances in EVs, there is a potential for much lower GHG emissions and less reliance on oil over the long term. Many argue that EVs hold such promise but, because they are at an early stage of development, manufacturing and consumer learning will determine their future success in the market.

“Spillover” effects are the second potential market failure for EVs. On the production side, investments by one manufacturer in R&D on battery innovation, power-train development, or vehicle design are likely to provide information and cost reductions to other manufacturers. Other firms may benefit from the learning or mistakes of first-mover companies, in which case firms will not invest in enough innovation or product development because their efforts spillover to others. On the other hand, if the firm captures all the benefits of its learning, there is no market failure and no reason to subsidize learning. Taking a lesson from the semiconductor industry, when learning is internalized, firms “price to the learning curve” and subsidize their own product to sell more and learn more. If, however, more innovation and product development by one firm informs other firms, then there are spillovers and hence a justification for government intervention.

In the case of EVs, auto manufacturers probably cannot capture all the benefits of their learning. Auto manufacturers try to protect new technology and production design, but as soon as prototype vehicles are available, other firms have access and they learn in ways that improve designs or lower costs or both.

There may also be learning on the consumer side of the market. Consumers are likely to be unfamiliar with the performance of EVs relative to traditional gasoline vehicles. One person’s use and learning about a new vehicle technology such as EVs can provide information to friends and neighbors or to the broader public.

Such spillovers may be an argument for subsidies to EVs, but because subsidies for a particular technology amount to “picking winners,” they carry a good deal of risk. It may be that EVs will remain too expensive, and will not get the improvements in battery development, design, and consumer acceptance that are needed, even with large subsidies. And what is the appropriate amount for society to spend to spark the possible spillover benefits? Subsidies should be aligned to the size of the spillovers, but it is very difficult to assess their magnitude, both on the manufacturing and consumer sides.

The purpose of any subsidy for EVs should be clearly identified, against which progress can be assessed. If the goal is to address manufacturer spillovers, is there evidence of learning across firms?  Are declining costs over time not just due to internal scale economies at a given firm, but due to learning across firms?  Finally, subsidy policies should be targeted as closely as possible to the market failure, and avoid distorting other markets. For example, it makes sense to target purchase subsidies to consumers who would not have purchased EVs otherwise. But what about offering HOV lane access to single-occupant EVs? Bento et al. show that granting single-occupancy access to HOV lanes for EVs can impose large unintended costs on carpoolers.