Friday, November 6, 2009

The REAL Final Clash for Clunkers Numbers

Given the doubts I expressed in this earlier post, I downloaded the NHTSA data and calculated the harmonic mean for the old and new vehicles in the 2009 Cash for Clunkers program.

The harmonic mean prevents the MPG Illusion by first converting all car MPG values to gallons per mile (GPM), averaging GPM, and then converting it back to MPG.

Here are the results:

Old vehicles
Reported Average MPG = 15.8
Average GPM = 0.064082517
Actual Average MPG = 15.6049

New vehicles
Reported Average MPG = 24.9
Average GPM = 0.041966565
Actual Average MPG = 23.8285

Well, that final actual MPG figure is a full mile per gallon below the reported figure of 24.9.

That 1 MPG difference translates to 1.2 million tons of CO2 by the time those 677,000 new vehicles drive 100,000 miles.

The supplement to the 2008 Science article discusses an "averaging illusion" example.

Final Cash for Clunkers Numbers (But with Doubts)

The final Cash for Clunkers numbers are in:

Old vehicles: 15.8 MPG on average
New vehicles: 24.9 MPG on average

That saves about 2 gallons of gas every 100 miles of driving, or 2 tons of CO2 every 10,000 miles of driving.

There is an AP story ridiculing the fact that some pick up trucks were traded in for other pick up trucks with essentially the same MPG (a 15 MPG truck for a 16 MPG truck), but that loophole was obvious from the start. It was built in by design to support Detroit. The Feinstein/Schumer/Israel contingent tried to hold out for good size increases in MPG, but had to compromise with the Stabenow contingent. Bare minimum truck trade ins was the disappointing concession. The time to publicize this flaw was in May. (As we noted in June, "The final bill has decent MPG improvements for cars (4 MPG and 10 MPG)--enough to "payback" the carbon released in producing the car--but not for large light duty trucks. We believe a tiered system based on GPM and requiring larger gas savings would have been better for reducing CO2 emissions.")

It's a little late now to lament it.


The numbers on the NHTSA site inspire no confidence, so all of these conclusions, from my perspective, are in doubt. For example, the NHTSA reports an overall new MPG level of 24.9 based on the following data:

59% new vehicles with an average 27.9 mpg
34% new vehicles with an average 21.6 mpg
7% new vehicles with an average 16.2 mpg

Amazingly, they have averaged the mpgs to get their final figure. Of course, they need to take the harmonic mean, which requires converting the mpg figures to gpm before calculating the final average. The harmonic mean is 24.2 mpg. So which is the real final figure: 24.9? or 24.2?

Update: This post examines the data more closely and finds that the actual MPG of the new vehicles is 23.8.

Wednesday, October 28, 2009

Trucks and Low Hanging Fruit

MPG obscures the value of small MPG improvements on inefficient vehicles.

Long haul trucks are a perfect example of the benefits of small MPG improvements. As this Greenbiz article notes, long haul trucks are low hanging fruit. The average truck gets 6 miles to the gallon; according to the article, technology exists that can double this number. A 6 MPG improvement may not sound impressive. A little reflection makes it obvious that this cuts gas consumption in half for a given distance of driving. And a little more math quantifies the gain:

6 MPG = 16 Gallons per 100 miles
12 MPG = 8 Gallons per 100 miles

The improvement saves 8 gallons per 100 miles of driving.

Compare that gas savings to the amount saved under currently planned CAFE improvements. Between now and 2016, CAFE standards will be increased from 25 MPG to 36 MPG. This 11 MPG improvement saves a little over 1 gallon per 100 miles of driving.

Policy should focus on improving the most inefficient vehicles because that's where the big savings are ripe for the plucking.

Here is a short excerpt from the article:

For 50 years, long haul tractor-trailer designs have remained fundamentally unchanged. Basically a giant box hurtling down the highway at 55 miles per hour, most trucks average only six miles to the gallon.

Their ubiquity in America is undeniable. Today the trucking industry transports about 70 percent of all the goods in the county, moving nearly $24 billion in value in 2008, according to the U.S. Department of Transportation.


But the time is ripe for change. According to recent analysis by Rocky Mountain Institute the technology already exists to double the energy efficiency of long-haul trucks in the nation's fleet. Their size, speed and poor aerodynamics mean they are laden with "low-hanging fruit" in terms of cost-effective efficiency and retroļ¬tting opportunities.

Friday, October 16, 2009

Proposed Rulemaking To Establish Light- Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards

The Federal Register has posted a new document from the EPA and NHTSA entitled:

"Proposed Rulemaking To Establish Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards"

The document contains a very useful summary of past conversations about supplementing MPG with a measure of fuel consumption. I have cut and pasted three sections on how consumers think about MPG. Page numbers refer to the Federal Register numbering. The passages appear below.

If you support the idea of making a gas consumption figure more salient to car buyers, you can comment on this proposal. Here is how to comment:
Submit your comments, identified by Docket ID No. EPA–HQ–OAR–2009–0472 and/or NHTSA–2009–0059, by one of the following methods:

• Submit your comments on line at this link.
• E-mail your comments at this email and include one of the Docket IDs above:

Pages 49577-49580

In 2006 EPA redesigned the window stickers to make them more informative for consumers. More particular, the redesigned stickers more prominently feature annual fuel cost information, to provide contemporary and easy-to-use graphics for comparing the fuel economy of different vehicles, to use clearer text, and to include a Web site reference to which provides additional information. In addition, EPA updated how the city and highway fuel economy values were calculated, to reflect typical real-world driving patterns.187 This rulemaking involved significant stakeholder outreach in determining how best to calculate and display this new information. The feedback EPA has received to date on the new label design and values has been generally very positive.

During the 2006 label rulemaking process EPA requested comments on how a fuel consumption metric (such as gallons per 100 miles) could be used and represented to the public, including presentation in the annual Fuel Economy Guide. EPA received a number of comments from both vehicle manufacturers and consumer organizations, suggesting that the MPG measures can be misleading and that a fuel consumption metric might be more meaningful to consumers than the established MPG metric found on fuel economy labels. The reason is that fuel consumption metric, directly measures the amount of fuel used and is thus directly related to cost that consumers incur when filling up.

The problem with the MPG metric is that it is inversely related to fuel consumption and cost. As higher MPG values are reached, the relative impact of these higher values on fuel consumption and fuel costs decreases. For example, a 25 percent increase in gallons per 100 miles will always lead to a 25 percent increase in the fuel cost, but a similar 25 percent increase in MPG will have varying impacts on actual fuel cost depending on whether the percent increase occurs to a low or high MPG value. Many consumers do not understand this nonlinear relationship between MPG and fuel costs. Evidence suggest that people tend to see the MPG as being linear with fuel cost, which will lead to erroneous decisions regarding vehicle purchases. Figure III.E.11–1 below illustrates the issue; one can see that changes in MPG at low MPG levels can result in large changes in the fuel cost, while changes in MPG values at high MPG levels result in small changes in the fuel cost. For example, a change from 10 to 15 MPG will reduce the 10-mile fuel cost from $2.50 to $1.60, but a similar increase in MPG from 20 to 25 MPG will only reduce the 10-mile fuel cost by less than $0.30.

Because of the potential for consumers to misunderstand this MPG/ cost relationship, commenters on the 2006 labeling rule universally agreed that any change to the label metric should involve a significant public education campaign directed toward both dealers and consumers.

In 2006, EPA did not include a consumption-based metric on the redesigned fuel economy label in 2006. It was concerned about potential confusion associated with introducing a second metric on the label (MPG is a required element, as noted above). EPA has developed an interactive feature on which allows consumers, while viewing data on a specific vehicle, to switch units between the MPG and gallons per 100 miles metrics. The tool also displays the cost and the amount of fuel needed to drive 25 miles. As indicated above, however, EPA is alert to the problems with the MPG measure and the importance of providing consumers with a clear sense of the consequences of their purchasing decisions; a gallon-per mile measure would have significant advantages. EPA plans to seek comment and engage in extensive public debate about fuel consumption and other appropriate consumer information metrics as part of a new labeling rule initiative. EPA also welcomes comments on this topic in response to this GHG proposal.

Pages 49602-49603

The central conundrum has been referred to as the Energy Paradox in this setting (and in several others).331 In short, the problem is that consumers appear not to purchase products that are in their economic self-interest. There are strong theoretical reasons why this might be so.332 Consumers might be myopic and hence undervalue the long-term; they might lack information or a full appreciation of information even when it is presented; they might be especially averse to the short-term losses associated with energy efficient products (the behavioral phenomenon of ‘‘loss aversion’’); even if consumers have relevant knowledge, the benefits of energy efficient vehicles might not be sufficiently salient to them at the time of purchase. A great deal of work in behavioral economics identifies factors of this sort, which help account for the Energy Paradox.333 This point holds in the context of fuel savings (the main focus here), but it applies equally to the other private benefits, including reductions in refueling time and additional driving.334

Considerable research suggests that the Energy Paradox is real and significant due to consumers’ inability to value future fuel savings appropriately. For example, Sanstad and Howarth (1994) argue that consumers optimize behavior without full information by resorting to imprecise but convenient rules of thumb. Larrick and Soll (2008) find evidence that consumers do not understand how to translate changes in miles-per-gallon into fuel savings (a concern that EPA is continuing to attempt to address).335 If these arguments are valid, then there will be significant gains to consumers of the government mandating additional fuel economy.

The evidence from consumer vehicle choice models indicates a huge range of estimates for consumers’ willingness to pay for additional fuel economy. Because consumer surplus estimates from consumer vehicle choice models depend critically on this value, EPA would consider any consumer surplus estimates of the effect of our rule from such models to be unreliable.

Pages 49603-49604

Although EPA has not found an updated survey of these values, a few examples suggest that the existing consumer vehicle choice models still demonstrate wide variation in estimates of how much people are willing to pay for fuel savings. For instance, Espey and Nair (2005) and McManus (2006) find that consumers are willing to pay around $600 for one additional mile per gallon.339 In contrast, Gramlich (2008) finds that consumers’ willingness to pay for an increase from 25 mpg to 30 mpg varies between $4,100 (for luxury cars when gasoline costs $2/gallon) to $20,560 (for SUVs when gasoline costs $3.50/gallon).340

As noted, lack of information is one possible reason for the variation. Consumers face difficulty in predicting the fuel savings that they are likely to get from a vehicle, for a number of reasons. For instance, the calculation of fuel savings is complex, and consumers may not make it correctly.341 In addition, future fuel price (a major component of fuel savings) is highly uncertain. Consumer fuel savings also vary across individuals, who travel different amounts and have different driving styles. Studies regularly show that fuel economy plays a role in consumers’ vehicle purchases, but modeling that role may still be in development.342

If there is a difference between fuel savings and consumers’ willingness to pay for fuel savings, the next question is, which is the appropriate measure of consumer benefit? Fuel savings measure the actual monetary value that consumers will receive after purchasing a vehicle; the willingness to pay for fuel economy measures the value that, before a purchase, consumers place on additional fuel economy. As noted, there are a number of reasons that consumers may incorrectly estimate the benefits that they get from improved fuel economy, including risk or loss aversion, poor ability to estimate savings, and a lack of salience of fuel economy savings.

Considerable evidence suggests that consumers discount future benefits more than the government when evaluating energy efficiency gains. The Energy Information Agency (1996) has used discount rates as high as 111 percent for water heaters and 120 percent for electric clothes dryers.343 In the transportation sector, evidence also points to high private discount rates: Kubik (2006) conducts a representative survey that finds consumers are impatient or myopic (e.g., use a high discount rate) with regard to vehicle fuel savings.344 On average, consumers indicated that fuel savings would have to pay back the additional cost in only 2.9 years to persuade them to buy a higher fuel-economy vehicle. EPA also incorporate a relatively short ‘‘payback period’’ into OMEGA to evaluate and order technologies that can be used to increase fuel economy, assuming that buyers value the resulting fuel savings over the first five years of a new vehicle’s lifetime. This assumption is based on the current average term of consumer loans to finance the purchase of new vehicles. That said, there is no consensus in the literature on what the private discount rate is or should be in this context.

A detailed discussion of the state of the art of consumer choice modeling is provided in the DRIA. For this rulemaking, EPA is not able to estimate the consumer welfare loss which may accompany the actual fuel savings from the proposal, and so any such loss must remain unquantified. EPA seeks comments on how to assess these difficult questions in the future.