This report uses emissions test result data from 1997 to evaluate the effectiveness of theenhanced I/M program in reducing vehicle tailpipe emissions in Phoenix, Arizona. The analysisis based on a comparison of initial and final test results for individual vehicles that received theirinitial I/M test in 1997. Two types of tests are performed on vehicles subject to I/M testing inPhoenix; the idle and loaded idle test is required of 1980 and older vehicles, while 1981 andnewer vehicles must take the IM240 test. Significant differences between the two types of testrequire that the emissions of the two fleets be analyzed separately. Arizona allows vehicles to fast pass or fast fail the IM240 test; in order to compare emissions ofvehicles tested over different portions of the IM240, we must convert these “short test” results tofull IM240 test equivalents. A relatively simple method to make this conversion is used; acomparison of this method with other more detailed methods indicates that all methods tend tounderestimate full IM240 emissions using fast pass/fast fail emissions results. The analysis doesnot consider the effect of the I/M program on reducing evaporative HC emissions. Comparison of initial and final IM240 tests indicates that the program is reducing the averageper vehicle emissions by 16% for HC, 17% for CO, and 7% for NOx, for the entire vehicle fleet.After weighting per vehicle emissions by estimated annual miles traveled, the fleetwideemissions reductions are 2.3 tons per day (14% reduction) for HC, 34 tons per day (15%reduction for CO), and 2.3 tons per day (7% reduction) for NOx. CO and NOx reductions appearto be substantially larger for cars than for light duty trucks. Per vehicle emissions of the loadedidle fleet are reduced by 15% for HC and 23% for CO. About 11% of all vehicles fail their initial IM240 emissions test; the failure rate is slightly higherfor passenger cars (12%) than for light duty trucks (8%). The initial failure rate for the loadedidle test is 37%. Of the vehicles that fail their initial test, only 70% received a final passing testthrough March 1998; 30% did not receive a final passing test through March 1998. Becausewaivered vehicles are not identified in the data, the actual percentage of No Final Pass vehicles islikely to be closer to 26%. The percentage of No Final Pass cars is greater than the percentage ofNo Final Pass trucks. The percent reductions in loaded idle emissions for Final Pass vehicles tend to increase by modelyear, with larger reductions for newer vehicles. There is a large increase in percent reduction formodel year 1974 through 1980 vehicles, presumably due to stricter cutpoints applied to thosevehicles. The percentage reductions of IM240 Final Pass vehicles from model years 1981through 1993 are fairly constant by model year. HC and CO emission reduction percentagestend to increase after model year 1993. We use a relatively crude method to estimate total emissions and emission reductions in tons perday for the loaded idle fleet, in order to estimate the tonnage reductions for the entire PhoenixI/M program. We estimate that the program reduces the emissions of the fleet reporting for I/Mby 3.0 tons per day for HC, 38 tons per day for CO, and 2.6 tons per day for NOx. The majorityof the estimated emissions reductions comes from the IM240 fleet: 76% for HC, and 88% for COand NOx. The estimated percent reduction in total emssions is 15% for HC, 13% for CO, and7% for NOx. The estimated effectiveness of the I/M program depends on whether the No Final Pass vehicleshave been permanently removed from the I/M area, or if they continue to be driven in the I/Marea. The effectiveness of the program on the IM240 fleet nearly doubles if one assumes that allIM240 No Final Pass vehicles have been permanently removed from the area. Analysis of 1995IM240 test data and remote sensing data indicate that about half of the No Final Pass vehiclescontinue to be driven in the I/M area. If this information is correct for vehicles tested in 1997,the 1997 I/M program resulted in a 22% reduction in HC and CO, and a 9% reduction in NOxfrom the IM240 fleet. These percentage reductions are equivalent to 3.0 tons per day for HC andNOx, and 48 tons per day for CO. Analysis of a single year of I/M program test data can only provide a partial understanding of theprogram’s effectiveness in reducing emissions. Tracking of individual vehicles over several I/Mcycles can reveal important information on long-term effectiveness of vehicle repair, andchanges in the fleet reporting for I/M testing. In addition, an independent source of on-roademissions tests, such as from a remote sensing measurement program, can provide additionalinformation on repair effectiveness, the effect of pre-test repairs on emissions, and the numberand emissions of vehicles avoiding the I/M program.