When a distribution this skewed is observed, it is easy to justify an air pollution program that identifies the gross polluters and targets them for treatment. It is correspondingly hard to justify programs that treat all cars as equal (oxygenated fuels, periodic mandatory emission testing, ride-sharing, etc.)

In 1989, we added a video camera that takes pictures of the rear of the vehicle for license plate information. From the license plate, we can obtain the vehicle make and model year. By now, we have measured more than two million vehicles in 20 countries.

Results from three countries are shown in Figure 3. The filled squares are data from 1991 in Los Angeles. New vehicles have low average emissions. As the vehicles get older, the average emissions increase. Notice that there is no discernable break in 1974 or 1980 when new technologies (catalysts, 1974; closed-loop computer systems, 1980) were introduced. The line close to the L.A. data was obtained in 1991 in Sweden. Sweden introduced catalysts 50% in 1987 and 100% in 1988. The break is clearly discernable, and Swedish catalyst-equipped cars have lower average emissions (by half) than similarly equipped vehicles in Los Angeles. There are a number of social/personal reasons to expect better car maintenance in Sweden. Not the least of which, my Swedish friends assure me, is that there is no word in Swedish for "tampering" with your emission control equipment.

If, as we believe, good maintenance is even more important than catalysts, then as L.A. cars age, one might expect to see the (apparently badly maintained) catalyst-equipped cars in L.A. having higher emissions than non-catalyst cars in Sweden. This effect is observed in the 1975-81 model years. Contrasting with the lower two lines is the upper line of data from the United Kingdom. The U.K. introduced catalysts in 1990, but it is apparent that my home country suffers from a fatal combination of both poor technology and poor maintenance.