VI.  Response to Issues Raised in OMB Return Letter About Preliminary Determination


    Pursuant to section 6(a)(3) of E.O. 12866, NHTSA is required to provide a written response to the points made by OMB in its February 12 return letter. As noted above, OMB stated in its return letter that: NHTSA should base its decision about the final rule on overall safety, instead of tire safety; while direct TPMSs can detect under-inflation under a greater variety of circumstances than indirect TPMSs, the indirect system captures a substantial portion of the benefit provided by direct systems; NHTSA should consider a fourth alternative for the long-term requirement, a one-tire, 30 percent compliance option, indefinitely, since it would allow vehicle manufacturers to install current indirect TPMSs; NHTSA, in analyzing long-term alternatives, should consider both their impact on the availability of ABS as well as the potential safety benefits of ABS; and that NHTSA should provide a better explanation of the technical foundation for the agency's safety benefits estimates and subject those estimates to sensitivity analyses.


    A. Criteria for Selecting the Long-Term Requirement

    1. Tire Safety and Overall Vehicle Safety

    OMB stated in its return letter that "a rule permitting indirect systems may provide more overall safety than a rule that permits only direct or hybrid systems." OMB said:

    Although direct systems are capable of detecting low pressure under a greater variety of circumstances than indirect systems, the indirect system captures a substantial portion of the benefit provided by direct systems. Moreover, allowing indirect systems will reduce the incremental cost of equipping vehicles with anti-lock brakes, thereby accelerating the rate of adoption of ABS technology. . . . Both experimental evidence and recent real-world data have indicated a modest net safety benefit from anti-lock brakes.

    While NHTSA's general obligation under the Vehicle Safety Act is to improve overall vehicle safety, it is mindful that its specific, immediate obligation in this rulemaking is to comply with the mandate of section 13 of the TREAD Act. The agency is seeking to comply with the mandate and safety goals of the TREAD Act in a way that encourages innovation and allows a range of technologies to the extent consistent with providing drivers with sufficient warning of low tire pressure under a broad variety of the reasonably foreseeable circumstances in which tires become under-inflated.

    2. Statutory Mandate

    Section 13 of the TREAD Act mandated the completion of "a rulemaking for a regulation to require a warning system in new motor vehicles to indicate to the operator when a tire is significantly under inflated" within one year of the TREAD Act's enactment. As noted below, the agency tentatively believes, based on the current record, that a four-tire, 25 percent under-inflation requirement would best meet the mandate.


    B. Relative Ability of Direct and Current Indirect TPMSs to Detect Under-Inflation

    As noted above, current indirect TPMSs work, in part, by adding the speeds of diagonal sets of tires and subtracting the sum of one set from the sum of the other. As a result, if all four tires are significantly under-inflated, and the difference in the tire pressures is not 30 percent or greater, current indirect TPMSs will not provide a warning. Similarly, if two tires on the same axle or same side of the vehicle are significantly under-inflated, current indirect TPMSs will not provide a warning.

    These combinations of significantly under-inflated tires occur frequently enough that current indirect TPMSs would have provided a warning in only about 50 percent of the instances in which NHTSA found significant under-inflation in the February 2001 NCSA survey. Conversely, current direct TPMSs would have provided warnings in all those instances.

    The following figures indicate how often current direct and indirect TPMSs would provide warnings when a vehicle has at least one tire that is at least 30 percent below the placard pressure.

    Of the 5,967 passenger cars in the February 2001 NCSA survey, 1,199 (20 percent) had at least one tire that was at least 30 percent below the placard pressure. Current direct TPMSs would have provided a warning in every case, while current indirect TPMSs would have provided a warning in only 653 cases (54 percent).

    Of the 3,950 light trucks in the NCSA survey, 789 (20 percent) had at least one tire that was at least 30 percent below the placard pressure. Current direct TPMSs would have provided a warning in every case, while current indirect TPMSs would have provided a warning in only 359 cases (46 percent).

    Thus, of the total 9,917 passenger cars and light trucks in the NCSA survey, 1,988 (20 percent) had at least one tire that was at least 30 percent below the placard pressure. Current direct TPMSs would have provided a warning in every case, while current indirect TPMSs would have provided a warning in only 1,012 cases (51 percent).

    Current indirect TPMSs would have failed to provide a warning in the remainder of the cases for various reasons. Many of the vehicles had one tire that was 30 percent below the placard pressure, but not 30 percent below the pressure in the other tires. As noted above, current indirect TPMSs require at least a 30 percent differential in tire pressure before providing a warning. Other vehicles had more than one tire that was 30 percent below the placard pressure. As noted above, current indirect TPMSs cannot detect when all four of a vehicle's tires, or two tires on the same side of the vehicle or the same axle, are under-inflated.

    The absence of a warning in approximately 50 percent of the instances of significant under-inflation is a matter of concern given that many drivers will rely on a TPMS instead of regularly checking their tire pressure. Data from the July 2001 BTS omnibus survey indicate that 65 percent of people would be less concerned, to either a great extent or a very great extent, with routinely maintaining the pressure of their tires if their vehicle were equipped with a TPMS. (34)


    C. Analysis of a Fourth Alternative Long-Term Requirement: One-Tire, 30 Percent Under-Inflation Detection

    As explained above in section V.A., "Alternative Long-Term Requirements Analyzed in Making Preliminary Determination," NHTSA analyzed three alternatives: a four-tire, 20 percent alternative; a three-tire, 25 percent alternative and a four-tire, 25 percent alternative.

    OMB recommended that the agency analyze a fourth alternative that would require a vehicle's TPMS to warn the driver when the pressure in any one of the vehicle's tires is 30 percent or more below the vehicle manufacturer's recommended cold inflation pressure for the tires, or a minimum level of pressure specified in the standard, whichever pressure is higher. (This alternative is referred to below as the "one-tire, 30 percent alternative.") The agency's analysis of the benefits and costs of this alternative follows.

    The agency estimates that the one-tire, 30 percent alternative would prevent 79 fatalities and prevent or reduce in severity 5,176 injuries. The agency estimates that the average per vehicle cost of this alternative would be $33.34. Since approximately 16 million light vehicles are produced for sale in the United States each year, the total annual cost of this alternative would be $533 million. The agency estimates that the average per vehicle maintenance cost would be $13.50, (35) and that the average per vehicle fuel and tread life savings over the lifetime of the vehicle would be $2.06 and $0.65, respectively. Thus, the net per vehicle cost of this alternative would be $44.13, and the total annual net cost would be $706 million. The net cost per equivalent life saved would be $5.8 million.


    D. Impact of One-Tire, 30 Percent Alternative on Installation Rate of ABS

    OMB said that NHTSA should analyze the impact of adopting its long-term regulatory alternatives as well as an additional long-term alternative, a one-tire, 30 percent alternative, on the installation rate of ABS. Since the additional alternative is the only one that would permit compliance by means of installing current indirect TPMSs, and since OMB's suggestion that a TPMS standard could induce increased installation of ABS is dependent upon the manufacturers' being able to install that type of TPMS, NHTSA's analysis focuses on that alternative.

    The agency believes there is no reliable basis for concluding that permitting current indirect TPMSs to comply would lead to a significant increase in installation of ABS in light vehicles for the following reasons.

    First, the final rule does not mandate the installation of ABS. Vehicle manufacturers always have the option of providing a measure that exceeds NHTSA's standards. However, nothing in the final rule requires manufacturers to install ABS.

    Second, the rulemaking record does not contain a reliable basis for concluding that manufacturers will voluntarily install ABS in significantly more light vehicles in response to being permitted to install current indirect TPMSs. When the Alliance addressed the issue of increased voluntary installation of ABS in its September 6, 2001 comments, it said only that a manufacturer "may well" opt to make ABS standard equipment on models for which optional ABS is currently available and is currently in high market demand. Further, only one manufacturer, Toyota, indicated that it might make ABS standard equipment on more vehicles if indirect TPMSs were allowed. Toyota provided this indication not in its written comments, but orally in a meeting with the agency. Nothing requires Toyota to make ABS standard equipment.

    Third, several manufacturers orally indicated that they would not install ABS on their light trucks even if indirect TPMSs were allowed. General Motors (GM) and Ford told NHTSA that they would install a direct TPMS on their trucks, rather than a four-channel ABS and indirect TPMS, because ABS was significantly more expensive. Further, the agency notes that in April 2002, GM announced that it would cease offering ABS as standard equipment on a number of its less expensive models of cars to make those models more price competitive.

    Fourth, it is not economically reasonable for manufacturers to install ABS voluntarily on significantly more vehicles in response to being permitted to install current indirect TPMSs. In the absence of written comments from individual manufacturers indicating that they are very likely to increase voluntarily their installation of ABS if allowed to install current indirect TPMSs, NHTSA may not simply assume that manufacturers will elect to spend $240 per vehicle to install ABS to save $53, the difference between the cost of a direct TPMS ($66) and an indirect TPMS ($13). The market for ABS has been static for several years, with the installation rate at about 63 percent. Absent a market demand for more installations, a manufacturer would not gain a market advantage by increasing the percentage of its vehicles with ABS.

    In NHTSA's Final Economic Assessment (FEA), the agency states that although a manufacturer may elect to increase the installation of ABS, it is solely a marketing decision. (36) The influence, if any, this rulemaking might have on their marketing decisions is purely speculative. There are many factors that influence a manufacturer's decision to install equipment. Cost impact is only one of them.


    E. Overall Safety Effects of ABS

    In addition to recommending that the agency assume that the adoption of the one-tire, 30 percent compliance option would induce vehicle manufacturers to increase their installation of ABS, OMB also recommended that the agency take into account the potential safety benefits of ABS when estimating the benefits of that option. OMB suggested that ABS could reduce fatalities in light vehicles.

    NHTSA has analyzed ABS and has determined that there is currently no statistically reliable basis for concluding that ABS reduces fatalities in light vehicles for the following reasons.

    First, NHTSA has analyzed the impacts of ABS on light vehicle fatalities for the past decade, with mixed findings. (37) In general, test track results indicate that ABS is a very promising technology that enables drivers to keep vehicles under control under adverse road conditions. Under some pavement conditions, ABS allows the driver to stop a vehicle more rapidly while maintaining steering control, even during panic braking.

    However, the agency's analysis of real world crash data shows that, on balance, ABS has not been proven, thus far, to be greatly beneficial in real world fatal crashes.

    NHTSA explored the desirability of requiring ABS on light vehicles in an ANPRM issued in 1994 (59 Fed. Reg. 281; January 4, 1994) in response to the National Highway Traffic Safety Administration Authorization Act of 1991. (Public Law 102-240, December 18, 1991). The Act directed the agency to consider the need for any additional brake performance standards for passenger cars, including ABS standards. The ANPRM solicited comments about whether rulemaking was warranted to require that all light vehicles be equipped with ABS. It also posed a number of questions relative to the regulatory approaches that might be employed if requirements were imposed; the types of performance tests that might be used; varieties of ABSs that might be appropriate; and regulatory implementation strategies and schedules that might be employed if requirements were established.

    Two years later, the agency issued a notice announcing that it had decided to defer indefinitely a decision whether to require equipping light vehicles with ABS. (61 Fed. Reg. 36698; July 12, 1996) In that notice, the agency stated that it was currently "inappropriate" to mandate ABS for the following reasons:

    (1) most studies that have analyzed the accident involvement experiences of ABS-equipped light vehicles have found mixed patterns, with a reduction in accidents in some crash modes but an increase in accidents in other crash modes, (2) even without a Federal requirement, a significant majority of light vehicles will be voluntarily equipped with ABS, (3) and requiring ABS on those light vehicles that will not be equipped with ABS would result in significant costs that, on balance, cannot be justified at this time.

    In the 1996 notice, the agency lowered the prediction that it had made in its 1994 ANPRM that the rate of voluntary ABS installation in passenger cars would increase from 55 percent in 1994 to 85 percent in 1999. Given that there had been almost no increase in the rate between the 1994 model year and 1995 model year, the agency suggested in the 1996 notice that the rate in 1999 could be as low as 70 percent. Even that reduced figure has been shown by subsequent events to be overly optimistic. In 2000, the rate had reached only 63 percent for passenger cars.

    The agency noted in the 1996 notice that the costs of bringing the percentages up to 100 percent for both passenger cars and light trucks could be very high, over $1.5 billion annually.

    Since the 1996 notice, NHTSA has conducted additional studies. In one study, NHTSA measured the braking performance of a group of ABS-equipped production vehicles over a broad range of maneuvers on different road surfaces. Results of this study showed that for most maneuvers, ABS-assisted stops yielded shorter stopping distances in comparison to non-ABS vehicles. (38)

    NHTSA has conducted several studies to examine possible reasons for the absence of overall safety benefits. One possible reason is that drivers are not adequately familiar or have inadequate or incorrect knowledge on the use of ABS. The agency has examined this possibility by conducting a national telephone survey to assess drivers' knowledge of ABS, its functionality and their expectations of its effects on vehicle performance. The results showed that, although most drivers had heard of ABS, many did not know what it did or how it affected vehicle performance. (39)

    The agency also investigated whether the apparent increase in single vehicle crashes was due to driver "oversteering" in crash-imminent situations. The steering capability could have contributed to vehicles going off of the roadway during crash avoidance maneuvers. However, this steering activity was not found to result in a significant number of road departure crashes in NHTSA's research. (40)

    The agency also evaluated possible ABS-related behavioral adaptation of drivers through the collection of more detailed data about the driving behavior of subjects in a naturalistic research setting. This study did not indicate any statistically significant trend towards behavioral adaptation by drivers of ABS equipped vehicles in comparison to others. (41)

    It is clear from the above comprehensive agency research efforts during the past five years that the agency still cannot explain why ABS systems do not produce the benefits anticipated from test track performance. Similarly, research by others has not yet succeeded in providing an explanation. Efforts by NHTSA and others continue today to try to explain this phenomenon.

    Second, OMB's apparent conclusion that increased installation of ABS in light vehicles could have a modest net safety benefit is based upon data that are not statistically significant. Those data are taken from a study by Charles M. Farmer for the Insurance Institute for Highway Safety (IIHS). (42)

    In the April 15, 2000 edition of its Status Report, IIHS said the following about the study:

    New evidence suggests that cars with antilock braking systems no longer are disproportionately involved in certain types of fatal crashes. However, antilocks still aren't producing reductions in overall fatal crash risk……

    …… As before, vehicles with antilock brakes were less likely than cars with standard brakes to be in crashes fatal to occupants of other vehicles. At the same time, the vehicles with antilocks no longer were found to be overinvolved in crashes fatal to their own occupants. Particularly important is the reduction in single-vehicle, run-off-the-road crashes.

    The data from the Farmer study are set forth in the table below:


      All Crashes 95 Percent Confidence
    Bounds
      Fatalities in ABS Cars
    Fatalities in Non-ABS Cars
    Lower Upper
    1. GM cars in 1993-95 1.03 0.94 1.12
    2. GM cars in 1996-98 0.96 0.87 1.05
    3. GM cars in 1993-98 0.99 0.93 1.05
    4. Non-GM cars in 1986-95 1.16 (Significant) 1.06 1.27
    5. Non-GM cars in 1996-98 0.91 0.77 1.06
    6. Non-GM cars in 1986-98 1.09 (Significant) 1.01 1.18


    A ratio of 1.0 in the second column means that ABS did not have any effect on fatalities. A ratio above 1.0 indicates a higher risk of fatalities in ABS-equipped vehicles, while a ratio below 1.0 indicates a lower risk of fatalities in ABS equipped vehicles.

    In order for the ratio for any group of vehicles to be statistically significant, both the lower and upper confidence bounds for that group must be either below 1.0 or above 1.0. This is true for only two groups of vehicles in the table: those in row 4, non-GM cars in 1986-95, and those in row 6, non-GM cars in 1986-98. For both of these groups, fatalities increased in ABS-equipped vehicles. Thus, in no subset of vehicles in the Farmer study is there any statistically significant advantage for ABS-equipped vehicles in crash fatalities.

    OMB interpreted the study to indicate a 4-9 percent reduction in fatalities in ABS-equipped vehicles. (43) However, NHTSA does not believe that these data are statistically significant because one confidence bound is below 1.0 and the other is above 1.0. Thus, these alleged benefits are more than 5 percent likely to be due purely to chance. (44)

    Mr. Farmer, the study's author, has indicated to NHTSA that people might have learned how to better use ABS by calendar years 1996-98, so that they were no longer at as great a risk of run-off-the-road fatal crashes as in prior years. (45) Even so, Farmer never stated in his study that ABS reduced fatalities. Regarding the Non-GM cars in 1996-98, he stated, "When all fatal crash involvements were considered, disregarding in which vehicle the fatalities occurred, the risk ratio was slightly lower than, but not significantly different from, 1.0."

    Third, the most recent NHTSA study showed an improved picture regarding benefits and disbenefits compared to earlier studies, but still no overall benefits in fatal crashes. (46) The study examined ABS effects separately for passenger cars and light trucks for five types of crashes: frontal impacts, side impacts, rollover, run-off-the-road, and pedestrian.

    The study found that, when both non-fatal and fatal crashes were combined, there were reductions in crashes for vehicles equipped with ABS. ABS was found to result in statistically significant reductions in crashes for most types of crashes, except side impact crashes, especially those involving cars.

    However, when only fatal crashes were considered, there were not any statistically significant overall reductions of those crashes for ABS-equipped vehicles. In fact, the only statistically significant finding was that fatal light truck rollover crashes increased in vehicles with ABS as compared to vehicles without ABS. (That did represent an improvement over a 1998 study (47) that found statistically significant increases for several types of crashes.) No statistically significant effects, positive or negative, were found for any type of fatal passenger car crashes or for other types of fatal light truck crashes.

    It is unclear whether the evidence in recent studies represents a statistical aberration relative to earlier studies or whether it is indicative of a real and positive trend. NHTSA will continue to monitor the real world performance of ABS on light vehicles. As with all protective devices, NHTSA plans to update its estimates for ABS as more data become available. If NHTSA obtains data enabling it to show that ABS reduces net fatalities and is cost/beneficial in light vehicles, the agency will consider initiating a separate rulemaking to address the issue of whether to require their installation.


    F. Technical Foundation for NHTSA's Safety Benefit Analyses

    OMB recommended that NHTSA better explain the technical foundation for the agency's estimates of safety benefits and subject those estimates to sensitivity analyses. (48) Since conducting these desired sensitivity analyses is relevant primarily to making a decision about the TPMS requirements for the long-term, the agency believes that its decision to postpone the final decision on TPMS requirements to the second part of this final rule makes it unnecessary to conduct additional sensitivity analyses at this time.

    The agency will complete its new study of TPMS by March 1, 2004. In this study, NHTSA will examine whether the tire pressure of vehicles without any TPMS are substantially closer to the vehicle manufacturer's recommended pressure than the tire pressure of vehicles with TPMSs, especially TPMSs that do not comply with the four-tire, 25 percent compliance option. If necessary, the agency will perform sensitivity analyses on these data.

    OMB specifically questioned the estimates of safety benefits that NHTSA made based on reduced skidding and better control, since these estimates were based on the Indiana Tri-level study published in 1977. The agency does not have later data of this quality on the effects of under-inflation on crashes. The agency has started to collect tire pressure data as part of its NASS-CDS data collection. However, NASS-CDS is not a system designed to determine the cause of a crash. Thus, NHTSA does not anticipate receiving significant further data on this issue. (49) However, if this issue becomes a critical element for the decision for the second part of this final rule, the agency will perform sensitivity analyses on the data from the 1977 study.

    OMB also noted NHTSA's use of Goodyear data, rather than VRTC data, on the effects of under-inflation on stopping distance. As explained in greater detail in the FEA, the agency did not use the VRTC data because of its concerns with the way in which the tests were performed. (50) The agency believes that the Goodyear test methodology adequately addressed these concerns. (51)

    In addition, OMB questioned the agency's use of the Goodyear data from a minivan to represent passenger cars. The critical element that is being measured is the difference in the tire's response when under-inflated. It is true that the absolute stopping distance will vary by vehicle weight and other vehicle performance characteristics. However, these same characteristics will influence both the properly inflated and the under-inflated tests in a similar fashion. Therefore, while Goodyear's test sample was confined to only two vehicles (a Dodge Caravan and a Ford Ranger), the differences measured under various inflation levels should still be indicative of the effect that could be expected.

    Finally, OMB questioned NHTSA's assumption that under-inflation is involved in 20 percent of blowouts that cause crashes. The agency does not know precisely how many blowouts that cause crashes are influenced by under-inflation. As noted above in Section III.D.1., "Reduced Vehicle Safety - Tire Failures and Increases in Stopping Distance," while the only tire-related data element in the agency's crash databases is "flat tire or blowout," even in crashes for which a flat tire or blowout is reported, crash investigators cannot tell whether under-inflation contributed to the blowout. The agency's best estimate is that under-inflation plays a role in 20 percent of blowouts that cause crashes.

    In making this estimate, the agency was mindful of the fact that many blowouts occur when one tire is punctured, begins to lose air at a rate somewhat faster than the normal rate due to natural causes, and then fails after being driven for some time while under-inflated. In these cases, a TPMS meeting either compliance option would be able to warn the driver of the under-inflated tire before the tire failed, possibly avoiding a crash.

    NHTSA emphasizes that the choice of 20 percent as its estimate of the percentage of under-inflation's involvement in blowouts that cause crashes made little difference in the agency's benefits analyses. As noted below in Section VIII.A.3., "Flat Tires and Blowouts," the agency estimates that the number of fatalities prevented per year due to reductions in crashes involving blowouts and flat tires will be 39 if all light vehicles meet the four-tire, 25 percent compliance option, and 32 if all light vehicles meet the one-tire, 30 percent compliance option. The choice of a somewhat higher or lower figure for the percentage of under-inflation's involvement would change only negligibly the relative benefits of the two compliance options.




    34 NHTSA notes that in its prepared statement submitted in connection with the February 28, 2002 hearing before the House Committee on Energy and Commerce on the TREAD Act, OMB stated:

    35 If the one-tire, 30 percent alternative were the only alternative available to vehicle manufacturers, the agency anticipates that the approximately 1/3 of vehicles not equipped with ABS would nevertheless comply by means of direct TPMSs. The approximately $40.91 of maintenance costs for each of those vehicles, if averaged over the entire fleet, is approximately $13.50.

    36 A copy of the FEA has been placed in the docket.

    37 See "Preliminary Evaluation of the Effectiveness of Antilock Brake Systems for Passenger Cars," NHTSA, December 1994, DOT HS 808 206. This study is available from the National Technical Information Service (NTIS) or NHTSA's Technical Reference Library.

    38 "NHTSA Light Vehicle Antilock Brake System Research Program Task 4: A Test Track Study of Light Vehicle ABS Performance Over a Broad Range of Surfaces and Maneuvers," January 1999, DOT HS 808 875, available at
    http://www-nrd.nhtsa.dot.gov/vrtc/ca/capubs/NHTSAabsT4FinalRpt.pdf.

    39 "NHTSA Light Vehicle Antilock Brake System Research Program Task 2: National Telephone Survey of Driver Experiences and Expectations Regarding Conventional Brakes versus ABS," November 2001, DOT HS 809 429, available at
    http://www-nrd.nhtsa.dot.gov/vrtc/ca/capubs/abssurvey_rptfinal.pdf.

    40 "Driver Crash Avoidance Behavior with ABS in an Intersection Incursion Scenario on Dry Versus Wet Pavement," (SAE Paper No. 1999-01-1288), available at
    http://www-nrd.nhtsa.dot.gov/vrtc/ca/lvabs.htm.

    41 "NHTSA Light Vehicle Antilock Brake System Research Program Task 7.1: Examination of ABS-Related Driver Behavioral Adaptation -- License Plate Study," November 2001, DOT HS 809 430, available at
    http://www-nrd.nhtsa.dot.gov/vrtc/ca/capubs/abs71.pdf.

    42 "New Evidence Concerning Fatal Crashes by Passenger Vehicles Before and After Adding Antilock Braking System," Charles M. Farmer, Insurance Institute for Highway Safety, February 2000. A copy of this study has been placed in the docket. (Docket No. NHTSA-2000-8572-206).

    43 The 4 percent figure is based on data for GM cars in 1996-98, while the 9 percent figure is based on data for non-GM cars in 1996-98.

    44 Most statisticians consider data that are more than 5 percent likely to be due purely to chance to be statistically insignificant.

    45 Mr. Farmer indicated this in an ex parte conversation with Jim Simons of NHTSA on February 14, 2002. (Docket No. NHTSA-2000-8572-210.)

    46 "Analysis of the Crash Experience of Vehicles Equipped with All Wheel Antilock Braking Systems (ABS) -- A Second Update Including Vehicles with Optional ABS," NHTSA, DOT HS 809 144, September 2000. A copy of this study has been placed in the docket. (Docket No. NHTSA-2000-8572-205.) It is also available at
    http://www-nrd.nhtsa.dot.gov/vrtc/ca/capubs/lvabstask1_crashdatareport.pdf.

    47 "An Analysis of the Crash Experience of Passenger Vehicles with Antilock Braking Systems-An Update," NHTSA, DOT HS 808 758, August 1998.

    48 When performing a sensitivity analysis, the agency changes assumptions it has made and then calculates differences in its benefits estimates. For example, the agency assumed that 20 percent of blowouts are caused by low tire pressure. If the agency performed a sensitivity analysis, it could change that assumption to 10 percent or 30 percent and then calculate a potential range of benefits.

    49 Although these data probably will not indicate whether low tire pressure caused a crash, the agency is collecting these data to determine the extent of the correlation between tire pressure and skidding/loss of control crashes.

    50 For example, the VRTC only tested new tires, not worn tires that are more typical of the tires on most vehicles. In addition, the NHTSA track surface is considered to be aggressive in that it allows for maximum friction with tire surfaces. It is more representative of a new road surface than the worn surfaces experienced by the vast majority of road traffic.

    51 For example, Goodyear tested tires with two tread depths: full tread, which is representative of new tires, and half tread, which is representative of worn tires.