The Environmental Protection Agency plans to change the 30-year-old method of estimating fuel economy ratings, the Boston Globe reports. Three changes are at the core of the proposal: --Alter testing to reflect today's more aggressive and high-speed driving habits, as well as address congestion in cities and suburbs. --Account for vehicles driven in cold climates, where fuel economy suffers. --Calculate the impact of accessories, such as air conditioners, that cut fuel economy. A Consumer Reports survey of 303 vehicles tested for the model years 2000-2006--each driven 8,000 to 10,000 miles--found that, in 90 percent of the cases, EPA mileage estimates were inflated, in some cases grossly so, the Globe reports. The auto companies use the same mileage estimates to meet Corporate Average Fuel Economy standards. Today's rules require a company's fleet of cars to average 27.5 miles per gallon and light trucks, including SUVs, 21.0 miles per gallon. Some Consumer Reports tests showed vehicles falling 40 and 50 percent below EPA estimates in city fuel economy. Consumer Reports buys the vehicles it tests and, over months of real-world driving, calculates its own mileage figures. Vehicles tested for EPA ratings are tested not by the EPA, but by auto manufacturers using EPA standards. The firms submit their results to the agency, which duplicates the tests in 10 percent of cases to check for accuracy. Why are the figures so inflated? --EPA testing is done on dynamometers, not pavement. --The highway test is at an average speed of 48 miles per hour on a smooth road. --The tests do not account for extreme temperatures, the use of air conditioners, bad road conditions, or increased urban and suburban traffic jams. --Today’s drivers reduce their mileage by being far more aggressive. The EPA said it hopes to have proposed regulations published by the end of this year, the Globe reports.