Warning: Obsessive data post to follow, probably of minimal interest to most normal people.
Brad Plumer has an interesting item today suggesting that once you account for all the electricity used to produce gasoline, electric cars not only use less gasoline than regular cars, they use less electricity too.1 Interesting! But something else in his post caught my eye: according to the Bureau of Transportation Statistics, the average new passenger car in 2010 got about 33.7 miles per gallon. Really? We just bought a new car a few months ago, and while we were doing our comparisons I was surprised at just how few cars were available that got really high mileage. So where does that 33.7 figure come from?
Well, Brad links to this table from BTS, which does indeed promote the 33.7 mpg number. But how did they come up with that? According to footnote C:
Assumes 55% city and 45% highway-miles. The source calculated average miles per gallon for light-duty vehicles by taking the reciprocal of the sales-weighted average of gallons per mile. This is called the harmonic average.
OK, so what's the source? Here it is:
1995-2009: Ibid., Highway Statistics (Washington, DC: Annual Issues), table VM-1, available at http://www.fhwa.dot.gov/policyinformation/statistics.cfm as of Apr. 20, 2011.
Great. So let's take a look at Table VM-1. It provides a figure of 23.8 mpg for all light-duty vehicles on the road in 2009. There's nothing there about about 2010 passenger cars in particular. So where does their data come from? Here's footnote 1:
The FHWA estimates national trends by using State reported Highway Performance and Monitoring System (HPMS) data, fuel consumption data (MF-21 and MF-27), vehicle registration data (MV-1, MV-9, and MV-10), other data such as the R. L. Polk vehicle data, and a host of modeling techniques. Starting with the 2009 VM-1, an enhanced methodology is used to provide timely indictors on both travel and travel behavior changes.
Hmmm. Table MF-21 estimates total 2009 gasoline usage of 132.8 billion gallons on highways and 3.8 billion gallons elsewhere. Table MF-27 provides a similar number for 2008.That's it. Table MV-1 informs us that there were 134.8 million automobiles registered in 2009. Tables MV-9 and MV-10 provide registration numbers for trucks and buses. None of that is helpful.
So apparently BTS's actual methodology is based on HPMS data "and a host of modeling techniques," not anything in those tables. That means I have no way to check their work. Still, does that 33.7 mpg figure seem credible? Using their 55%/45% split, that would mean an average city EPA rating of about 29 mpg and a highway rating of 39 mpg. That sure seems high to me. DOE's search site won't let me plug in those exact numbers, but when I ask for a list of all cars rated above 30 city and 40 highway, I get a grand total of 13 hits — and of those, eight are either Volkswagens or Smart cars, neither of which has a huge sales presence in the United States. What's more, as near as I can tell, not a single one of the top ten sellers in the United States in 2010 had a combined mileage of 33.7 mpg, and according to Ward's Automotive, only 4% of auto buyers in 2010 purchased cars with a combined mileage over 30 mpg, let alone 33.7 mpg.
Bottom line: If I had to guess, I'd say that somewhere between 2-5% of passenger car sales in 2010 had a combined mileage of 33.7 mpg. The average mileage of 2010 cars just has to be way less than that. If anyone has better data on this, please let me know.
1Or maybe not. In an update, Brad says this: "According to this Argonne study — and this analysis by the Department of Energy’s Jacob Ward— it takes about 6 kwh of energy to refine a gallon of gasoline, not 6 kwh of electricity, as I originally stated." Electric cars still come out looking pretty good, but quite that good.