Monday, 19 September, 2005

EPA Changing Mileage Testing

The Environmental Protection Agency is in the process of changing the way that they compute the fuel economy ratings on new cars.  The agency has long been criticized for publishing unrealistic numbers due to their test methods.  For example, assuming average highway speeds of 48 MPH, under-estimating the emount of time spent idling during city driving, not testing at temperature extremes, and not testing with the air conditioning on.  Even though the EPA's published information says quite clearly that the numbers are intended only as guidelines for comparison shopping, consumers and consumer advocates get bent out of shape when their actual mileage varies significantly (it's almost always lower) from the EPA estimates.

Auto manufacturers like the EPA way of calculating fuel economy because it's overly generous and makes the cars look better than they actually are.  In addition, the EPA numbers are used in calculating the federally-mandated Corporate Average Fuel Economy standards.  Companies already have difficulty meeting those standards.  More realistic numbers from the EPA would put every car company in violation of the CAFE standards.

This is an interesting case of one number (two numbers, actually--the EPA "miles per gallon" ratings for city and highway driving) being used for two different things.  As far as auto manufacturers and the EPA are concerned, it's "just a number" to show that one car uses more or less fuel than another.  Testing is done on dynamometers (treadmills for cars) that are calibrated to simulate real-world conditions--most likely ideal conditions.  The car never runs on real pavement, at temperature extremes, in real traffic, or anything else.  It's not a realistic test of how the car would operate out on the road, now was it intended to be.  But it does exactly what it was designed to do:  provide an idea of relative performance.

Consumers see those two numbers on the new car sticker and automatically think that they will get that kind of gas mileage.  "The sticker says I should get 30 MPG on the highway."  Most consumers don't bother themselves with the fine print that says "your mileage may vary," or bother to take into account that their lead foot driving habits and jackrabbit starts (called "stoplight drag racing" these days) might affect the fuel economy.  Something as simple as under-inflated tires, something I see quite regularly, can have a huge effect on the amount of gas burned per mile.

I don't see how the EPA can develop a standard and method of testing that will meet the expectations of both groups.  The only way would be to create a huge series of tests that take into account vehicle maintenance, driving conditions, and driving habits.  Such testing would be prohibitively expensive for manufacturers and the resulting table of numbers would be so confusing that consumers' eyes would glaze over.  The average car shopper would just take the largest number in the table, use that to comparison shop, and then complain when the car they buy doesn't meet the expectations.  Not that I'm cynical or anything.

A better idea would be for consumers to spend five minutes reading and understanding the EPA guidelines, and spend another hour or two reading reviews of the cars that they're looking at in order to get an idea of how the car performs in real driving conditions.  With that information perhaps buyers would have more realistic expectations and could make informed decisions when purchasing.