A Free-Market Energy Blog

U.S. EPA’s Adventures In Arithmetic: A Look at the CO2 Car Standards

By Donald Hertzmark -- April 8, 2010

The U.S. Environmental Protection Agency (EPA) has proposed to effect a reduction in CO2 releases in the U.S. by raising the required fuel economy standards for new cars in 2014 and again in 2016.  The current standard, now at 30.2 mpg for passenger cars (everything here is about passenger cars, the analysis of light trucks will have to wait) will rise to 35.5 mpg in 2016.

EPA claims that they used a carbon price of $21/tonne to establish the appropriate increase in fuel economy.  The EPA also claims that these standards will reduce CO2 releases from the vehicle sector by 21%.  Well, at least they are not using the number 19.  This proposal will have a minute effect on CO2 levels and is unlikely to come in at the very low or “negative” cost per tonne of CO2 claimed by its proponents.

Static Scoring and the Jevons Paradox

One of the oldest professional fights in Washington is about whether the objects of the government’s interest at a given time will simply accept what is put in front of them or react, perhaps in ways not included in the “scoring.”  Decades of poor predictions on cost and performance should have taught us that assuming no response by consumers, investors or others is probably the least likely of all possible scenarios.

In the energy business people react when their real costs of energy change.  If better furnaces, doors, windows and insulation make a home comfortable with fewer BTUs per cubic foot of space, then many consumers will respond by adding space to their dwellings.  It all depends on whether the demand for the fuel is responsive to the change in its real price.  Clearly the Jevons Paradox has been operating in the US housing sector, with average house size nearly doubling since 1970.

Something similar has occurred in the US automobile market.  Since 1970 gasoline demand in the US has risen by 10% (NB: passenger cars only).  At the same time the vehicle population has increased by 52%, vehicle miles traveled are up by 84% and mileage per vehicle is up by 20%.  What makes all of this possible is that gasoline use has become increasingly sustainable for consumers.  In other words, the real cost of traveling by car has gone down for most Americans.

As a consequence of improved technology, fuel use per vehicle has fallen from 760 gal/year in 1970 to 554 in the late 2000s, and average fuel economy is now above 22 miles per gallon.

U.S. drivers have evidently responded to the falling real cost of driving by doing more of it.  Miles per vehicle per year rose from 10,300 in 1970 to 12,400 in 2006.  There is no indication that these trends are slowing down.  In fact, almost all of the increase in miles traveled per vehicle since 1970 came after 1990.

What Does Any of This Have to Do With Emissions, You Ask?

The EPA, like most government agencies, assumes that people will not change their behavior when the price and performance parameters in their economic decision-making are adjusted.  This is unlikely to be the case.  If consumers pay more up front for a car that uses less fuel, then chances are they will drive it more, especially in preference to an older, less efficient sibling in the garage.  On the other hand, people are more likely to defer purchases (or speed them up) to avoid the higher costs associated with the higher mpg vehicles.

In any event, fuel consumption is not very likely to be what the EPA projects for 2030 because their behavioral parameters are off.  And if the miles traveled are off then so will be the claimed savings.  In the next section I have tried to figure under what conditions the EPA standards may reduce carbon releases at the claimed cost per tonne of CO2 and what such reductions might actually cost.

Whatever Happened to the Number 21?

The US Department of Transportation estimates that the additional cost per vehicle to meet the 2016 fuel standard will be about $1,000.  As with all such estimates this is an opening gambit.  And so is the claimed fuel savings.  Through a combination of reduced fuel use, impact of improved efficiency on oil prices (why do they claim market effects only when it benefits them?), drivers are expected to save more than $3,000 in gasoline expenditures over the life of a vehicle.

Such results are actually possible if:

(i) the price of gasoline goes to $5/gallon in real terms and stays there;

(ii) drivers do not increase their motoring to take advantage of the cheap miles;

(iii) the additional costs of fuel economy are limited to $1,000 per vehicle; and

(iv) one uses nominal dollars without discounting. 

If all of the EPA’s assumptions hold, then the cost of reducing CO2 releases is negative – the Nirvana of environmentalists, negative cost emissions reductions.  In fact, if all three of these conditions hold, then the cost of reducing CO2 by one tonne is an astounding minus $275.

A quick spin around Kelly Blue Book indicates that each extra mpg in a given model range costs about $400-600 (e.g., the differentials between gasoline and hybrid vehicles of the same model or diesel v. gasoline engines in the same model).  So the move from 30.2 mpg to 35.5 mpg in 2016 is not likely to come in at anything lower than $1,500 per car; something north of $2,500 per vehicle is far more likely.  We also need to note that $5/gallon gasoline would arise if the real price of crude oil rose to $150/barrel or more (note that is the price without European-style gasoline taxes).  Since the world economy cannot function well with such oil prices, this should be considered to be an outlier case.

Using more realistic figures for additional vehicle costs and gasoline, $2,000 and $4/gallon in real terms (that’s about $110/bbl for crude), respectively, the cost of saving one tonne of CO2 rises to $99.40.  The present value of the gasoline saved is $373 less than the present value of the additional vehicle costs.

If improved fuel economy induces more driving, an eminently reasonable view in light of historical experience, then the cost of reducing CO2 releases will increase further.  For example, if the owner of one of these vehicles drives an additional 1,000 miles annually for the ten year life of the car, then the present value of the fuel savings is $1,121 less than the additional cost of the vehicle and the cost of one less tonne of CO2 rises to $552.

The calculated savings in emissions and money are extremely sensitive to various assumptions about oil prices and costs of mileage improvements, etc.  If the price of crude oil stays below $100/barrel in real terms (i.e., if the world economy remains functioning) then the additional costs per vehicle can be no higher than $1,248.25 to get to a CO2 reduction cost of exactly $21/tonne.


Whenever claimed performance is exquisitely sensitive to changes in the assumptions, one should be skeptical of the supposed benefits.  So it is with the EPA’s backdoor regulation of greenhouse gases through fuel economy standards.  Prior to the Climategate fiasco, it was indeed the case that CO2 reduction certificates were selling for about $21/tonne.  Since then the price has followed a roller coaster course and now stands at about $15/tonne, with great futures market compression (reduced contango for you players out there).

So the EPA’s claimed miracle, a negative price for reducing CO2, cannot be true unless several implausible conditions are also true – that the price of oil is high enough to destroy the airlines, trucking and chemical industries, while leaving consumers with enough money to purchase new cars; that fuel economy becomes a great deal less costly to augment than it is today; and that people would not react to lower real costs of additional mileage by driving more.

What is more likely is that we will pay far more to reduce CO2 released through the tailpipe of new cars than we would in reducing it at power plants or other energy conversion sites.[i] Ignoring the real (lower) cost of additional miles renders the forecasts useless, as does the dependence on very high oil prices to make their numbers work.[ii]

[i] But then again, reducing emissions may not be the real goal here.

[ii] And speaking as an occasional economics professor, I would like to see the government start using present value methods so that we can actually compare costs and benefits in the future with those of today.


  1. Mark Walker  

    Well put – your main argument holds in that even as retail gas prices increased 2 summers ago, higher mileage vehicles, and the fact that the real cost of gasoline/energy expenditures as a proportion of personal income is much lower than in the ’70’s, resulted in virtually no change in miles driven when gas neared $5.00/gal.

    Increasing gas mileage will result in more miles driven. Making cars lighter and smaller to meet mileage targets will result in more highway and driving deaths and accidents as miles driven increase.

    No reduction in CO2 tailpipe emissions (not that it matters), increased gasoline consumption and more traffic related deaths – bureaucracy in action.


  2. Donald Hertzmark  

    Actually, we do not know what would happen if the real price of fuel went to $5/gal and stayed there forever. On the other hand (see fn 2) evidence from a briefer period of such prices in 2008 indicates that it can certainly tank the economy and, by the way, crater new car sales.

    If high gasoline prices do come to pass (by restricting US oil and gas production?) as a permanent feature, the miles driven will go down at some point, reducing the emissions “benefit” of the fuel economy standard. If fuel prices remain in their current range, then better fuel economy will likely lead to an increase in miles driven, thereby blunting the CO2 “emissions” impact of the standard.


  3. Mark Krebs  

    Here is an even more illustrative adventure in arithmetic:
    In their ruling, electric vehicles are given credits for zero-emissions. But at least the number is capped at 200,000 per maker through 2016. After that, additional EVs would be charged “some” emissions responsibility.


  4. Gary Novak  

    Fuel efficiency numbers don’t mean what they used to mean. They used to mean actual improvements in efficiency due to improved technology. That isn’t possible anymore. Now efficiency means getting less work done, with about the same cost per unit work. It means smaller and lighter vehicles with less power. That’s what drove people to SUVs and Pickups–when they needed to get more work done, they needed bigger equipment. Which way are they going to turn now? Driving dump trucks to the grocery store? In many cases, they will make more trips or use more vehicles, which well result in more energy used. This may be part of the reason why people are driving more miles. They can’t get as much done per trip anymore.
    Gary Novak


  5. Cooler Heads Digest 9 April 2010 | GlobalWarming.org  

    […] EPA’s Adventure in Arithmetic Donald Hertzmark, MasterResource.org, 8 April 2010 […]


Leave a Reply