V. BENEFITS ANALYSIS
Human Factors Issues
The Tire Pressure Monitoring Systems (TPMS) will provide notification to drivers that their tire pressure has dropped below the level recommended by the manufacturer. However, driver response to this information may vary depending upon the nature of the information provided by the TPMS. NHTSA believes that almost all drivers will respond in some manner to the warning, but the level of information presented to the driver by different display systems may result in different behavior by drivers.
The direct measurement systems could display individual tire pressures and tell the driver which tire(s) are low. Although individual tire pressures are not proposed to be required, this analysis assumes in Compliance Option 1 that all of the vehicles will be supplied with direct measurement systems that will display individual tire pressures because it will be helpful to drivers in terms of fuel economy, tread wear and safety. This was done because of uncertainty regarding the exact nature of displays that manufacturers will install. The indirect and hybrid measurement systems can only provide a warning lamp that tire pressure is low. Compliance Options 2 and 3 assume all vehicles will be equipped with only a warning lamp.
We anticipate that drivers will react differently to the different amounts of information. Some drivers will keep track of the individual tire pressures and will add pressure to their tires whenever necessary, say at 10 percent below placard, even before the warning is given. These drivers will accrue more safety benefits and more benefits in terms of fuel economy and tread life than drivers that wait longer for a warning. On the other hand, some drivers who currently check their own tires frequently enough to avoid significant under-inflation may start to rely on the TPMS to indicate under-inflation, rather than checking their tires frequently and filling them up whenever they were below the placard level. We believe this would happen more often under Compliance Options 2 and 3, where only a warning lamp comes on when tire pressure goes below a specified threshold, rather than under Compliance Option 1, where individual tire pressures could be monitored continuously. These drivers would actually accrue fewer safety, tread wear and fuel economy benefits than they did without the TPMS.
The agency has little information that would help it estimate how a TPMS would affect overall driver tire maintenance behavior. A survey question in the Bureau of Transportation Statistics Omnibus Survey of July 2001 asked 1,004 respondents “To what extent do you agree that an indicator lamp in your vehicle that warns the driver about under-inflation in any of the vehicle’s tires would allow you to be less concerned with routinely maintaining the recommended tire pressure?” The responses were 40 percent to “a very great extent”, 25 percent to “a great extent”, 18 percent to “some extent”, 7 percent to “a little extent”, and 10 percent to “no extent”. Putting this information together with survey data from the tire pressure survey, where one-third of those surveyed indicated that they check their tire pressure at least once a month, indicates that some people would check their tire pressure less frequently.
The agency has some information that would help it estimate what percent of drivers would put to use the information on individual tire pressures. From the agency’s tire pressure survey, we found that about one-third of the interviewed drivers indicated that they check their tire pressure once a month or more frequently. For Compliance Option 1, we assume that one-third of the drivers would pay attention to the individual tire pressure information provided on a monitor and would refill their tires when they were 10 percent below the placard. This means that if the average passenger car tire placard is 30 psi, we assume for Compliance Option 1 that one-third of the drivers would refill their tires when they get to 27 psi. The other two-thirds of the drivers would refill their tires when the warning is given at 25 percent below placard, or 22.5 psi for the average passenger car.
The second question is whether drivers, given a warning, will stop and inflate their tires back to the placard pressure. We do not expect driver compliance with the TPMS telltale, which is amber or yellow, to be 100 percent. In the Final Economic Assessment, we assumed that 95 percent of drivers will fill the low tire(s) to make sure they don’t get a flat tire and be stranded somewhere. Given just a telltale, the driver will probably need to check all the tires. Given a reading of tire pressure on all four tires with a direct measurement system, the driver will know which tire(s) are low and need to be filled.
This assumption was based on NHTSA’s own estimates and a study relating to the Cycloid Pump. “Examining the Need for Cycloid’s Pump: An Analysis of Attitudes and a Study of Tire Pressure and Temperature Relationships”, December 7, 2001 by the University of Pittsburgh Department of Mechanical Engineering, Department of Industrial Engineering. This study included a survey of people’s attitudes. The survey was not a random survey of consumers representing a national picture. The 225 respondents to the survey were:
One of the questions was:
Q21. Would you respond to a dashboard warning lamp informing you that your tire pressure was low?
Note that there were several questions before this one on how often do you check your tire pressure, when was the last time you checked your tire pressure, what is the recommended tire pressure in your vehicle, etc. These types of questions set up the respondents to thinking that tire pressure is an important topic worthy of checking out.
While this is not a random sample, the question format may have biased the responses, and driver’s actual deeds are often different from their telephone response, the response is overwhelming and leads some small credence to a very high estimate (our initial estimate was 95 percent of drivers will respond to a warning lamp).
In 2003, NHTSA collected information on direct and indirect systems, in terms of tire pressure and asked the owners several questions. This report is still in progress. Preliminary results from questions in this survey to determine consumer reaction to existing TPMS systems indicated that in almost 95% of cases where vehicles had direct systems, and the driver was given a low tire pressure warning, the drivers responded by taking appropriate action. These preliminary survey results thus validate NHTSA’s initial assumption. However, considering that these are all new vehicles and relatively expensive vehicles that have a direct TPMS, and that typically the reactions of purchasers of more expensive vehicles to behavioral warnings will be higher than the reactions of the average or second-time owners, we have assumed a more conservative 90 percent response rate to a warning.
In the Final Economic Assessment we assumed that there will be a natural process whereby, people fill up their tires and then the tires lose air over time. Thus, the benefits of the system are going from the level of pressure in the tire survey to an average level of pressure between times the tires are refilled using the following assumptions:
|Steady State psi Level for Passenger Cars||Steady State psi Level for Light Trucks|
|Compliance Option 1||27.0 psi||31.5 psi|
|Compliance Option 2||26.3 psi||30.6 psi|
|Compliance Option 3||26.3 psi||30.6 psi|
Skidding and Loss of Control
For loss of control crashes, speed is the most critical factor. Excessive speed alone can cause a loss of control in a curve or in a lane change maneuver. Tread depth, inflation pressure of the tires, and road surface condition are the most notable of a long list of factors including vehicle steering characteristics and tire cornering capabilities that affect the vehicle/tire interface with the road. In the Indiana Tri-Level Study, under-inflation was not considered a contributing factor to a crash when there was high speed involved. It was only considered when the tires were significantly under-inflated (an undefined term generally taken by the investigators to mean at least 10 to 15 psi below recommended pressure). Still, it is hard to know whether correcting this one problem area could result in the collision being avoided or reduced in severity. That is one reason why under-inflation was never cited as the definite cause of a crash. We tried to consider this by comparing under-inflation as a percentage of all of the probable causes in crashes. Certainly, reducing under-inflation is an important area and a move in the right direction. However, it is difficult to determine what the effectiveness of increasing tire pressure would be on these crashes. The following discussions describe how inflation pressure affects these crash types to the extent known.
Skidding and/or loss of control in a curve
Low tire pressure generates lower cornering stiffness because of reduced tire stiffness. When the tire pressure is low, the vehicle wants to go straight and requires a greater steering angle to generate the same cornering force in a curve. The maximum speed at which an off-ramp can be driven while staying in the lane is reduced by a few mph as tire inflation pressure is decreased. An example provided by Goodyear shows that when all four tires are at 30 psi the maximum speed on the ramp was 38 mph, at 27 psi the maximum speed was 37 mph, and at 20 psi the maximum speed was 35 mph while staying in the lane. Having only one front tire under-inflated by the same amount resulted in about the same impact on maximum speed. But, the influence of having only one rear tire under-inflated by the same amount was only about one-half of the impact on maximum speed (a 1.5 mph difference from 30 psi to 20 psi).
The agency also has run a series of tests to examine the issue of decreases in tire pressure on vehicle handling. A 2001 Toyota 4-Runner was run through 50 mph constant speed/decreasing radius circles to see the effects of inflation pressure on lateral road holding. We examined lefthand turns from 0 to 90 degrees handwheel angle for tire inflation pressures varied from 15 to 35 psi. The data indicate to us that in on-ramps/off ramps, tire inflation pressure is a critical factor in vehicle handling. The data show how much friction the vehicle can utilize, in terms of lateral acceleration (g’s), before it slides off the road. The more lateral g’s the vehicle can utilize, the better it stays on the road. So, if you are going around an off-ramp and need to turn the wheel 50 degrees at 50 mph, you can utilize 0.27 g’s at 15 psi, or you can utilize 0.35 g’s at 30 psi.
Skidding and/or loss of control in a lane change maneuver
In a quick lane change maneuver, under-inflated tires result in a loss of tire sidewall stiffness, causing poor handling. Depending upon whether the low tire(s) are on the front or rear axle impacts the vehicle’s sensitivity to steering inputs, directional stability, and could result in a spin out and/or loss of control of the vehicle.
Skidding and/or loss of control benefits estimate
In Chapter IV, we estimated a target population for skidding and loss of control crashes for under-inflated tires of 247 fatalities, 23,100 injuries and 53,130 property-damage-only crashes. The agency assumes that 90 percent of drivers will fill their tires back to placard pressure.
It is difficult to determine the effectiveness estimate, (i.e. , what percent of the crashes would be avoided by just improving low tire pressure). For this analysis, we assume 20 percent effectiveness to go from a very low pressure, where a warning would be given, to the steady state condition, although it could potentially be much higher. Thus, the benefits by Compliance Option are shown in Table V-1. An example calculation resulting in the estimated 44 fatalities is (247*.90*.20*.99 to account for one percent current compliance).
|PDO||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Non-Fatal Inj.||Total|
Note that the benefits are the same for all the Compliance Options, since they all require warnings at 25 percent below placard pressure. It is assumed that the benefits would come from increasing tire pressure from a low state to a pressure close to placard pressure. This reflects the finding that the levels of under-inflation in the Indiana Tri-Level study were higher than 25 percent to have under-inflation reported as a probable cause.
Tires are designed to maximize their performance capabilities at a specific inflation pressure. When tires are under-inflated, the shape of the tire’s footprint and the pressure it exerts on the road surface are both altered. This degrades the tire’s ability to transmit braking force to the road surface. There are a number of potential benefits from maintaining the proper tire inflation level including reduced stopping distances, better handling of the vehicle in a curve or in a lane change maneuver, and less chance of hydroplaning on a wet surface, which can affect both stopping distance and skidding and/or loss of control.
The relationship of tire inflation to stopping distance is influenced by the road conditions (wet versus dry), as well as by the road surface composition. Decreasing stopping distance is beneficial in several ways. First, some crashes can be completely avoided by stopping quicker. Second, some crashes will still occur, but they occur at a lower impact speed because the vehicle is able to decelerate quicker during braking.
In Chapter III, a variety of stopping distance test results are discussed. For the Preliminary Economic Assessment, NHTSA examined test results submitted by Goodyear Tire and Rubber Company as well as tests conducted at its own Vehicle Research Test Center (VRTC). In tests conducted by Goodyear Tire and Rubber Company, significant increases were found in the stopping distance of tires that were under-inflated. By contrast, tests conducted by NHTSA at their VRTC testing ground found only minor differences in stopping distance, and in some cases these distances actually decreased with lower inflation pressure. The NHTSA tests also found only minor differences between wet and dry surface stopping distance. It is likely that some of these differences were due to test track surface characteristics. 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. The Goodyear tests may also have been biased in other ways. Their basic wet surface tests were conducted on surfaces with .05” of standing water. This is more than would typically be encountered under normal wet road driving conditions and may thus exaggerate the stopping distances experienced under most circumstances. A general problem that applied to both data sets was that they measured stopping distance impacts for new tires only, while most vehicle miles are traveled on tires that are worn down to a level that is somewhere between full and minimal tread depths. Since tread depth and tread profile can greatly influence both water retention and tire friction, this could have a significant impact on estimates of tire pressure on stopping distance. Generally speaking, the Goodyear test results implied a significant impact on stopping distance from proper tire pressure, while the NHTSA tests implied these impacts would be minor or nonexistent at lesser water depths. The PEA estimated stopping distance impacts using the Goodyear data to establish an upper range of potential benefits. A lower range of no benefit was implied based on the NHTSA test results.
In the earlier PEA and in a subsequent memo to the docket (Docket No. 8572-81), NHTSA expressed concern regarding the adequacy of the currently available test data. In response, Goodyear conducted a new and comprehensive series of tests to evaluate the effects of tire inflation pressure on stopping distance. The Goodyear tests were conducted using two different vehicles (Dodge Caravan and Ford Ranger), two different tires (P235/75R15 Wrangler and 215/70R15 Integrity), three inflation pressures (35, 28, and 20 psi), two tread depths (full tread and half tread), and three water depths (dry, .02 inches, and .05 inches). In addition, the tests were run with vehicles with ABS and without ABS. The stopping distance was collected from 45 mph to 5 mph. Goodyear found that collecting the data at 5 mph reduced the variability in the results as compared to a full stop to 0 mph. A separate set of traction truck tests were also run to establish peak and slide coefficients of friction for these tires under similar circumstances but at speeds of 20, 40, and 60 mph.
NHTSA examined the new data submitted by Goodyear and determined that it provided a much more comprehensive data set than was used previously for the earlier PEA. The variety of water depths and tread depths were particularly important to resolving critical concerns with the initial data sets used in the earlier PEA. During the comment period, NHTSA contracted with the National Oceanic and Atmospheric Administration (NOAA) (See Docket No. 8572-167) to develop a data base that could be used to analyze the relative frequency of rainfall intensity in the U.S. Based on these data, the conditions which are likely to produce a surface water depth level of .05 inch, which was the basis for the original Goodyear tests, only occur about 10 percent of the time that it rains. Thus, the addition of a second lesser water depth test of .02 inch was critical to measuring the impact on crashes that occur under most wet road conditions. The new Goodyear data also confirmed that tread depth has a significant influence on stopping distance. Overall, the new test data provided a comprehensive picture of the impacts of tire inflation on stopping distance, and were relatively free of the contradictions found in the earlier data sets. For these reasons, NHTSA based the final analysis on the new data set provided by Goodyear, rather than average the results of the two previous conflicting sets of data.
Impact Speed/Injury Probability Model
In order to estimate the impact of improved stopping distance on vehicle safety, NASS-CDS data were examined to derive a relationship between vehicle impact speed (delta-V) and the probability of injury. Following is a description of the derivation of this model.
Data: From 1995-1999 CDS, all passenger vehicle occupants involved in crashes where at least one passenger vehicle used brakes.
MAIS 4+, MAIS 5+, and fatal injuries was calculated for each delta-V between 0 and 77 mph. The percent probability risk of each MAIS j+ injury level at each delta-V i mph is defined as the number of MAIS j+ injury divided by the total number of occupants involved at i mph delta-V. If j = 0 represents MAIS 0 injuries and j = 6 represents fatalities, the probability of injury risk can be represented by the following formula:
i = 0 to 77, j = 0 to 6
p+i,j = percent probability risk of MAIS j+ injuries at i mph delta-V,
x i,j = the number of j+ injuries (i.e. , MAIS 0, MAIS 1+, MAIS 2+, …, fatal) at i mph delta-V
Ti = total number of occupants at i mph delta-V
Note that p+i,0 = percent probability risk of MAIS 0 injuries at i mph delta-V and p+i,6 = percent probability risk of fatalities at i mph delta-V. Ii,0 = the number of MAIS 0 injuries and Ii,6 the number of fatalities at i mph delta-V.
(2) The risk-prediction curve for each j injury level was derived using a mathematical modeling process. The process used delta-V as the independent variable (i.e. , predictor) and p+i,j as the dependent variable and modeled all the data points (delta-V, percentage risk) for each j injury level. For example, for MAIS 1+ injuries, the process used the data points: (0, p+0,1), (1, p+1,1), (2, p+2,1), …, (75, p+75,1), (76, p+76,1), (77, p+77,1) to derive the MAIS 1+ risk curve. Table V-2 shows all the risk-prediction formula. These formulas were developed under two assumptions: a) no one was injured at 0 mph, i.e. , p+0,0 = 100 percent, and p+0,j = 0 percent for j=1…6, and b) everyone was assumed to have at least MAIS 1 injuries for 36 mph and higher delta-V, i.e. , p+i,0 = 0 , for i >=36 mph. This assumption was based on the injury distribution derived from 1995-1999 CDS.
|Injury Level||Risk-Prediction Formula|
(3) The percent probability risk pi,j was calculated for individual MAIS level. For MAIS 0 (j=0) and fatal injuries (j=6), pi,0 = p+i,0 and pi,6 = p+i,6 . The percentage risk for each MAIS 1 to MAIS 5 injury level is the difference between the two predicted risks. Thus, pi,1 (risk of MAIS 1 at i mph delta-V) = p+i,1 - p+i,2, pi,2 = p+i,2 - p+i,3, pi,3 = p+i,3 - p+i,4, pi,4 = p+i,4 - p+i,5, and pi,5 = p+i,5 - p+i,6.
(4) Adjusted total row percent risk to 100 percent. Because of statistical measurement variation and predicting errors, the row risk percentages at some delta-Vs do not add to 100 percent. To adjust to a total of 100 percent for these delta-Vs, an adjustment factor (fi) is applied to every risk probability. The adjustment factor is 100/(actual total percentage), i.e. ,
where j = 0…6.
The adjusted risk probabilities for i mph delta-V would be fi * pi,j. For example, at 10 mph delta-V, f10 = 100/85 = 1.1765. The risk probability for MAIS 0 becomes 52.5 (= 44.6*1.1765) and MAIS 1 becomes 43.5 (= 37.0*1.1765). These adjusted risk probabilities are higher than those predicted by the original curves listed in Table V-2. However, the general shape of each curve does not alter significantly. Table V-3 shows the adjusted percent probabilities of risk. Note that cell probabilities were rounded to the nearest tenth. Therefore the sum of the individual cells may not total exactly 100 percent.
Once this relationship was established, crash data from 1999 CDS and FARS were distributed across this matrix to establish a “base case” injury distribution. This was done separately for 3 different groups of crashes stratified according to the speed limits on the roadways where crashes occurred. The roadway stratification was selected because stopping distances are largely dependent on initial pre-braking travel speed, and speed limits were assumed to provide a reasonable stratification for this variable. However, actual travel speeds differ from speed limits. For this analysis, it was assumed that actual travel speeds were 5 mph higher than the mean speed limit in each category. The 3 speed limit categories were 0-35mph, 36-50mph, and 51 mph and over. The mean speed limits for each category were 30, 44, and 57. There were only minor differences between speed limits for wet and dry surfaces, or for passenger cars and LTVs. Therefore, the same average speed limit is used regardless of road surface or vehicle type. Allowing for a 5 mph difference for travel speed, the three assumed average speeds that represent the speed limit categories are 35, 49, and 62 mph.
|Delta-V (mph)||MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal||Total|
|Table V-3 |
Adjusted Percent Probabilities of Injury Risk,Cont.
Separate target populations were also derived for passenger cars and LTVs, and for crashes that occur on wet and dry pavement. These distinctions were necessary because stopping distance is strongly influenced by pavement conditions and vehicle characteristics. In addition, LTVs have significantly different levels of under-inflation than passenger cars and this impacts calculations of delta-V reductions. Note that the presence or absence of anti-lock brakes also has a significant influence on stopping distance. However, because reliable data on the presence of these systems is not included in crash databases, these differences will be accounted for at a different stage of the analysis. A total of 12 separate target population cells were thus produced. The fatalities and injuries for each cell are summarized in Table V- 4 for passenger cars and Table V-5 for LTVs. Table V-6 summarizes the target populations across all passenger vehicles.
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal||Total|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal||Total|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal||Total|
The impact of small reductions in stopping distance will, in most cases, result in a reduction in the impact velocity, and hence the severity, of the crash. However, in some cases, reduced stopping distance will actually prevent the crash from occurring. This would result, for example, if the braking vehicle were able to stop just short of impacting another vehicle instead of sliding several more feet into the area it occupied.
The benefits that would accrue from preventable crashes would only impact that portion of the fleet that:
Data from NHTSA’s tire pressure survey (see Table III-1) indicate that 26 percent of passenger cars and 29 percent of LTVs have at least one tire that is 25 percent or more below recommended placard pressure. For these vehicles, notification of this under-inflation would not be given until the system is triggered. For example, under the proposed requirements, a direct TPMS will trigger at 25% below placard pressure, or roughly 22.5 psi for passenger cars and 26.25 psi for trucks. The portion of the vehicle fleet that is below these levels will potentially experience some reduction in crash incidence due to improved stopping distance. However, in order to experience this reduction in stopping distance, the driver must respond to the warning. For the March 2002 Final Economic Assessment, NHTSA assumed that 95 percent would respond to a warning and refill their tires back to the placard level.
Preliminary results from a recent survey conducted to determine consumer reaction to existing TPMS systems indicated that in 95% of cases where vehicles had direct systems, the drivers responded by taking appropriate action. These preliminary survey results thus validate NHTSA’s initial assumption. However, the vehicles that have existing TPMS tend to be more expensive luxury vehicles that are typically purchased by upper income populations. Since these groups are typically more safety conscious than lower income groups, it is likely that the survey results imply a lower level of response for the overall driving public. Based on this, the overall response rate across all income groups will be estimated to be 90%.
The portion of crashes that would actually be preventable is unknown. However, an estimate can be derived from relative stopping distance calculations for vehicles that were involved in crashes. The average stopping distance was calculated for the existing crash-involved vehicle fleet, and for that fleet if they had correct tire inflation pressure. The method used to calculate these stopping distances is described later in this section of the analysis. The results indicate that the existing passenger car fleet would, on average, experience a stopping distance of 86.5 feet, while the crash-involved LTV fleet experienced an average stopping distance of 91.9 feet. These differences between passenger car and LTV stopping distances reflect the distribution of injuries by speed and road conditions for each vehicle type. By contrast, the average stopping distance for passenger cars with correctly inflated tires would be 85.2 feet, while for LTVs it would be 90.7 feet.
In theory, current crashes occur under a variety of stopping distances but if these distances were shortened due to improved inflation pressure then a portion of these crashes would be prevented. Crashes could be prevented over a variety of travel speeds and braking distances. For example, a vehicle might be able to avoid an intersection crash by slowing quickly enough to miss a speeding vehicle running a red light. In an angular head-on crash, better braking could reduce the chance of two vehicles striking their corners, given that crash avoidance maneuvers are also taking place. An example for rear impacts could involve sudden braking to avoid a vehicle swerving to cross lanes on an interstate highway. We anticipate that a large portion of the fatality and serious injury benefits for crash avoidance would occur in intersection crashes, since both vehicles are moving at high speeds, and a small change in braking efficiency could result in the avoidance of a high-impact crash.
NHTSA does not have data that indicate average stopping distance in crashes. Under these circumstances, it is not unreasonable to assume that crashes are equally spread over the full range of stopping distances. Under this assumption, the change in stopping distance under proper inflation conditions can be used as a proxy for the portion of crashes that are preventable. With equal distribution of crashes across all stopping distances, the portion of crashes that occur within the existing stopping distance that exceeds the stopping distance with correct pressure represents the portion of crashes that are preventable. For passenger cars, this portion is (86.5-85.2)/86.5 or 1.38 percent of all current crashes. For LTVs, this portion is (92.0-90.7)/92.0 or 1.36 percent.
Benefits from preventable crashes were thus calculated as follows:
Ip(s)= Preventable injuries of severity (s)
Pp = portion of crashes that are preventable
I(s)= Existing injuries of severity (s)
Pu = portion of vehicles with under-inflated tires that will receive notification from TPMS
Pr = portion of drivers who will respond to the TPMS notification
The results of this analysis are shown for passenger cars under Compliance Options 2 and 3 in Table V-7. The results for LTVs are shown in Table V-8, and for all passenger vehicles Table V-9. Results for Compliance Option 1 will be summarized at the end of this section, but will not be demonstrated. Note that these results have been adjusted to reflect a small amount of overlap that occurred in the separate examination of passenger car and LTV crashes, as well as potential overlap with“loss of control”crashes, which are accounted for separately in a previous section. A combined adjustment factor of .959 was applied to account for this overlap. This factor was derived by comparing the sum of the two separate crash counts to a total count based on all passenger vehicles. These estimates were also adjusted to reflect the impact of threshold braking, as well as current compliance. These concepts are discussed in detail in the following section on non-preventable crashes.
The benefits from preventable crashes, shown in Tables V-7, 8 and 9 were assumed to occur over all crash types and severities. This assumption recognizes that there are a variety of crash circumstances for which marginal reductions in stopping distance may prevent the crash from occurring. Crash prevention may be more likely under some circumstances than others. For example, it is possible that a larger portion of side impacts might be prevented than head-on collisions. In side impacts where vehicles are moving perpendicular to each other, improved braking by one vehicle reduces the speed at which it enters the crash zone and potentially allows the second vehicle to move through the crash zone, thus avoiding the impact. In a head-on collision, both vehicles are moving toward the crash and a reduction in stopping distance for one vehicle may be less likely to avoid a high-speed crash than in the case discussed above for side impacts. Further, if a separate analysis were conducted for different crash types and severities, the portion of crashes prevented would be greater for crashes at higher speeds. However, NHTSA does not have sufficient information to conduct a separate analysis of each crash circumstance and has used an overall estimate across all crash types instead.
Note that this analysis only addresses injury crashes. Property-damage-only crashes would also be impacted by proper tire inflation. These crashes are addressed separately in a later section of this analysis.
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
In the vast majority of crashes, small changes in stopping distance will not prevent the crash, but will reduce the speed at impact and thus the severity of the crash. As noted above, 1.38 percent of braking passenger cars and 1.36 percent of braking trucks could have avoided crashes with proper tire inflation. The remaining 98.6 percent of passenger car and LTV crashes would still occur, but at a reduced impact speed. To estimate the impact of reduced crash speeds, changes in stopping distance will be estimated and used as inputs to recalculate impact speeds for the population of non-preventable crashes. These changes in impact speeds will then be used to redefine the injury profile of this crash population shown in Table V-3, and safety benefits will be calculated as the difference between the existing and the revised injury profiles.
Stopping distance can be computed as a function of initial velocity and tire friction. The formula for computing stopping distance is as follows:
SD = Vi2/(2*g*Mu*E)
SD =Stopping Distance (in feet)
Vi = initial velocity (mean speed limit for specific data group + 5 mph)
g = gravity constant (32.2 ft/second squared)
Mu = tire friction constant (ratio of friction force/vertical load )
E = ABS braking efficiency (estimated @ 0.8)
About a third of all passenger vehicles sold in the U.S. do not have anti-lock brakes, although the portion is higher in the on-road fleet. For these regular braking systems, the term for anti-lock brake efficiency (E) would not be used.
The value of Mu is dependent on surface material (concrete, asphalt, etc.), surface condition (wet vs. dry), inflation pressure, and initial velocity. Based on data provided by The Goodyear Tire and Rubber Company in response to the NPRM, NHTSA developed a model that predicts Mu based on Vi and inflation pressure. Separate models were developed for Mu at both peak (the maximum level of Mu achieved while the tire still rotates under braking conditions) and slide (the level of Mu achieved when tires cease to rotate while braking (i.e., skid)). The peak models are used for vehicles with antilock brake systems. The slide models are appropriate for vehicles with non-antilock brake systems. The models are as follows:
For Wet surface conditions
Mp = 0.83140+(.0037109*ip)-(0.0038408*Vi)+(0.000023292*Vi2)
Ms = 0.55093+(0.0029423*ip)-(0.0036979*Vi)-(0.000020146*Vi2)
For Dry surface conditions
Mp = 0.978764+(.002557*ip)-(0.005542*Vi)+(0.0000470863*Vi2)
Ms = 0.717073+(0.000618*ip)-(0.005242*Vi)+(0.000082917*Vi2)
Mp = Mu peak value
Ms = Mu slide value
ip = inflation pressure (psi)
Vi= initial vehicle speed (mph)
Note that the wet surface condition model is based on 2 separate models. One was derived from the Goodyear tests conducted with .05 inches of water, and one with .02 inches of water. As noted previously, data from NOAA (See Docket No. 8572-167) indicate that only about 10 % of rainfall events occur at rates that would be necessary to produce .05 inches of water on road surfaces. The 2 wet condition models were therefore weighted to produce a single model based on weights of 90% for the .02 inch model and 10% for the .05 inch model
Mu Surface Adjustments
The above formulae were derived from tests conducted on a Traction Truck surface (this is a specific surface calibrated to specifications of OEM customers). In order to relate them to real world surfaces, predicted values from the formulas were compared to actual test results obtained using the same tires mounted on vehicles. The vehicles used were a Dodge Caravan with a 215/70R15 Integrity tire, and a Ford Ranger with a P235/75R15 Wrangler tire. Generally, the Integrity tests were intended to represent passenger cars while the Wrangler tests were intended to represent LTV performance. The tests were all run with an initial velocity of 45 mph, with braking measured down to 5 mph. Goodyear did not record data to a complete stop. In order to compare the predicted stopping distance results from the Mu regressions to real world results, braking distance was measured using the following equation:
SD = (Vi2 -Vii2)/(2*g*Mu*E)
SD = braking distance
Vi = initial speed before braking
Vii = speed to which vehicle braking is measured
This is a simple modification of the formula previously discussed for stopping distance. The Vii term is necessary to adjust for the 5 mph braking limit in the vehicle tests. Mu peak and slide values were estimated for each of the 3 psi levels used in the Goodyear vehicle tests at 45 mph. The resulting predicted SDs were then compared to the actual stopping distance found in the corresponding vehicle tests. The actual SDs were weighted to reflect an average of the full and half tread tests. Weighting factors for the actual SDs were derived from tread depth data obtained in NHTSA’s tire inflation survey. Full tread for the Integrity tire (assumed to represent passenger tires) was 10/32 inch and half tread was 5/32 inch. For the Wrangler tire (assumed to represent LTVs), full tread was 13/32 inch and half tread was 6.5/32 inch). Data from the NHTSA survey indicate that about 2/3 of all vehicle tires had tread depths more similar to the½tread level and about 1/3 had tread depths more similar to the full depth levels.
A comparison of the predicted and actual weighted SDs indicated close similarity across the three different psi levels. Therefore, factors were averaged across the 3 levels. However, they differed significantly by tire type, surface condition, and for peak vs. slide. Overall, the results of this comparison indicate that factors of from roughly 1.3 to 1.8 are required to adjust the stopping distances predicted using the Mu-based algorithms. The Wrangler factors were applied to LTV estimates and the Integrity factors were applied to passenger car estimates. Wet and dry factors were also applied to their corresponding cases. Peak factors were applied to vehicles with antilock brakes, while slide factors were applied to vehicles without antilock brakes. The factors used are summarized in Table V-11.
Anti-lock and Normal Braking Systems
Roughly 2/3 of all passenger vehicles sold in the U.S. have anti-lock brakes, but the portion is smaller in the on-road fleet. For vehicles with anti-lock brake systems, Mp is used to calculate stopping distance because it represents the peak controlled braking force that anti-lock brakes attempt to maintain. For vehicles with regular brake systems, Ms is used because it represents the level of friction encountered under normal braking by most drivers without assistance from anti-lock brakes. Also, for these regular braking systems, the term for anti-lock brake efficiency (E) would not be used.
Changes in stopping distances were then used to calculate the decrease in crash forces (measured by delta-V) that would occur due to the decrease in striking velocity of the vehicle. The formula used to calculate striking velocity is:
V(d) = velocity of vehicle at distance d after braking
Vi = initial velocity before braking
a = deceleration
d = distance traveled during braking of vehicle
In this case, V(d)is a measure of the speed at which the vehicle with under-inflated tires would be traveling when it reaches the distance at which it would have stopped had its tires been correctly inflated (d). Deceleration (a) is calculated for the vehicle with under-inflated tires. The derived formula for deceleration is:
a = (V(d)2-Vi2)/(2*d)
Since V = 0 at d, the formula becomes:
a = (Vi2)/(2*d) (the negative sign that would precede the formula indicates deceleration and will be ignored from this point on)
The distance over which a is calculated is the stopping distance for the vehicle with under-inflated tires. This will be designated as SDu. The formula thus becomes:
a = (Vi2)/(2*SDu)
SDu = stopping distance with under-inflated tires
The striking velocity is then expressed in mph by multiplying by 1/ 5280 ft.*3600 sec. hour. The delta-V experienced by each vehicle would be dependent on vehicle mass. For this analysis, the mass of each vehicle was assumed to be equal, giving a delta-V of 1/2 V(d) for each vehicle or:
DELTA-V = (V(d)*3600/5280)/2
DELTA-V = the change in velocity resulting from increased tire pressure.
The base case target population represents the injury profile that results from the fleet of passenger vehicles that were on the road at that time. In order to determine the inflation pressure that exists in that fleet, NHTSA conducted a survey of both recommended and actual inflation pressures on vehicles. (Details of that survey are discussed elsewhere in this analysis). The results of the survey indicate that 74% of all passenger vehicles are driven with under-inflated tires. However, because TPMS would not notify drivers of low pressure until it dropped 25% below placard, no stopping distance benefits would accrue to vehicles with smaller tire pressure deficits. Weighting factors were derived from the tire pressure survey to represent the affected population under this requirement. The distribution of each level of under-inflation is shown in Table V-12. The left column indicates the average under-inflation of the 4-tires, given that one tire was under-inflated by 25 percent or more.
|Under-Inflated||Percent Under-Inflated||Percent Under-Inflated|
As noted previously, the value of Mu in the formula for stopping distance is dependent on inflation levels. For each speed limit category, a set of delta-Vs corresponding to each under-inflation level was calculated. In each case, an average placard pressure of 30 psi was assumed for passenger cars. For LTVs, an average pressure of 35 psi was assumed. The rates of under-inflation in Table V-12 were used to weight the change in delta-V that results from each corresponding psi under-inflation level to an overall weighted average change across all levels. The resulting changes in delta-V are summarized in Table V-13 for each passenger car and LTV target population category for vehicles with ABS systems, non-ABS systems and combined systems, based on weighting factors representing the relative portion of the vehicle fleet that has Anti-lock brakes. Note that these estimates do not reflect any impact for vehicles with inflation levels that are less than the assumed set point for the TPMS system. This analysis assumes a set point of 25 percent below the placard pressure, or 7.5 psi based on the assumption of a 30 psi recommended pressure. Benefits would only accrue to those tires that are more than 7.5 psi beneath their recommended pressure. For LTVs, benefits would accrue for those tires that are more than 8.75 psi beneath their recommended pressure.
Calculation of Safety Benefits
Safety benefits were calculated by reducing the delta-V for each injury by the appropriate level for each specific target population category shown in Table V-13. The injury totals for each delta-V category were redistributed according to the injury probabilities of the reduced delta-V level. This resulted in a new injury profile. Totals for each injury severity category were then compared to the original injury totals to produce the net benefits from reducing delta-Vs. An example of the original target population distribution and the revised distribution is shown in Tables V-14 and V-15. Note that the revised distribution shown in Table V-15 represents a whole number delta-V change (in this case, 6 delta-V). Since actual average reductions were fractional, interpolation was used to calculate the results of the fractional reductions. These interpolated results are reflected in Table V-16. Table V-20 summarizes the results for all scenarios for passenger cars under Compliance Alternative 2.
By comparing current tire pressure levels to placard, benefit estimates reflect raising pressure levels to the placard level and retaining them there. However, over time tire pressure will drop back down to the threshold notification level again and drivers will again fill their tires to the placard level. Over time, the benefit that drivers obtain will be an average of the benefits from the various levels above the notification threshold. For this analysis, it was assumed that pressure loss is roughly constant at one psi per month and a revised average psi level was calculated for passenger cars and LTVs under each Alternative. These averages were previously shown in Table V-1x. Under the assumption that there is a reasonable correspondence between changes in delta-V and safety benefits, changes in Delta-V were recalculated based on the averages in Table V-1x. This was done by substituting the new average psi levels for theplacard pressure in previous calculations. An additional adjustment was made to reflect the impact on that portion of the fleet for which at least one tire was below the notification threshold, but for which the average psi across all 4 tires fell above the revised average psi level but below the placard level. This was done because these cases would be excluded by calculations based on 4 tire average psi levels below placard. The output from this process was a set of factors that were used to modify the results. These factors typically reduced benefit calculations based on full placard inflation levels to about 60% of their full placard level. The results of applying these factors are shown in Table V-21.
Adjustments to Non-Preventable Crash Safety Benefits
A number of adjustments must be made to the benefit estimates in Table V-19. These include:
|Delta-V||MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal||Total|
|Delta-V||MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal||Total|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
Braking Distance Distribution
Table V-16 represents safety impacts that would occur from the reduced stopping distance of a tire at the point where it would stop if pressure were corrected. It represents the maximum change in delta-V that would occur in cases where the actual braking distance in the crash just equals the correct stopping distance. In reality, crashes occur over a variety of braking distances, and the change in delta-V is a direct function of this distance. This relationship is illustrated in Figure V-2 below. The change in delta-V is virtually non-existent in crashes where braking distance is minimal, but becomes significant as the distance traveled during braking increases.
Generalized Relationship Between Change in
Delta-V and Traveling Distance
To account for the variety of possible outcomes, a factor was calculated based on the relationship between calculated delta-V changes and travel distance. The techniques used to calculate this factor are fully described in Appendix A. The results indicate that the impacts over the variety oftravel speeds would be about 7 percent of those based on maximum impact for both passenger cars and for LTV’s.
Properly Inflated Vehicles
As previously mentioned, 26 percent of all vehicles have no tires under-inflated. In addition, many vehicles have a level of under-inflation that would not trigger a warning from the TPMS. The target population used in the above calculations assumes a full fleet of under-inflated vehicles and must be adjusted for the portion of the fleet that is not under-inflated, and that will be notified of the problem. The portions differ by Alternative and vehicle type. Based on NHTSA’s tire pressure survey 26 percent of passenger cars and 29 percent of light trucks would benefit from a TPMS.
Table V-16 also represents the benefits that would accrue if all drivers responded immediately to the TPMS and inflated their tires to the proper level. Since this is unlikely to occur, an adjustment was made to represent the driver response rate, which, based on preliminary results from a survey of TPMS equipped vehicles, the agency estimates to be 90 percent. (The preliminary survey results indicated a response of 95 percent. Ninety percent was use in this analysis in the belief that the survey results, which were self reported, may have overstated actual response rates. This also provided a conservative estimate of benefits).
Overlapping Target Populations
As previously noted separate target populations were derived for passenger cars and light trucks because the under-inflation profile is different for these vehicle types. These populations were stratified based on the vehicle braking. However, a comparison of the two separate injury counts to a single count done for any passenger vehicle indicated that a small amount of double counting resulted from a simple addition of the two separate braking vehicle populations. Based on this comparison, an adjustment factor of .9685 was applied to the benefit estimates to eliminate the overlap. In addition, there is potential overlap between the target population examined here and the one used to calculate“out of control”crash impacts earlier in this analysis. To adjust for this overlap, an analysis of overlapping cases was conducted and an adjustment of 1% (i.e., a factor of .99 was applied) was made to reflect these cases.
Driver Response- Braking Threshold
When drivers are faced with potential crash circumstances, they apply their brakes at a rate that reflects both their perceptions of the need to stop and the vehicles actual response to this need. Theoretically, braking systems should be capable of the needed response, if drivers apply it, up to a threshold at which the tires loose their friction capabilities. On dry pavement, this would occur when tires exceed their peak coefficient of friction and start to skid rather than grip the pavement. In this analysis, it will be assumed that during emergency braking, all potential inadequacies in braking performance, including those caused by underinflated tires, will be perceived by drivers and that they will respond by applying more pressure to the brakes to compensate. Under these circumstances, any small impacts to stopping distance due to changes in the tire pressure that would occur prior to skidding on dry pavement would be compensated for by the driver. However, when skidding occurs, the driver can no longer compensate for such changes. To reflect this, CDS data from 1995-1999 was examined to determine what portion of fatalities and injuries occurred in crashes in which skidding occurred on dry surfaces. This analysis indicated that 72% of fatalities and 54% of injuries that occurred on dry pavement happened in crashes with skidding. On dry pavement, only these crashes with skidding would benefit from the TPMS. These factors were thus applied to all dry pavement stopping distance benefits. Given the high level of skidding involved on dry pavement, this analysis assumes that all crashes that occur on wet pavement involve some level of skidding and thus would benefit from TPMS. This may slightly overstate the impacts of TPMS in wet pavement crashes.
About one percent of the new car fleet already has a direct monitoring system. This portion of the fleet would not require costs or experience benefits from this rulemaking. A total of 5 percent of the fleet has either an indirect system (4%) or a direct system (1%). However, the current indirect systems would not meet the requirements of this final rule.
The above 6 adjustments were accomplished by multiplying the results in Table V-16 by factors of .07, .26, .95, .9589, .72 or .54 (dry pavement only), and .99 to account for current compliance. Similar adjustments were made for each vehicle type and Compliance Option. Table V-17 summarizes the total adjusted non-preventable crash benefits for passenger cars under Compliance Option 2. Table V-18 summarizes the benefits from non-preventable crashes under Compliance Option 2 for LTVs. Table V-19 summarizes the non-preventable benefits for all vehicle types under Compliance Option 2. Table V-20 summarizes total safety benefits for all crashes (Preventable and Non-Preventable) for passenger cars under Option 2. Table V-21 summarizes the Total safety benefits for all crashes for LTVs under Option 2. Table V-22 summarizes the total potential stopping distance impacts for all crashes and all vehicle types under Option 2. Note that safety benefits would be identical for Compliance Options 2 and 3. Table V-23 shows the potential stopping distance impacts across all crashes and vehicles for Compliance Option 1, which assumes a continuous readout of tire pressure is provided.
The results indicate a potential safety impact under Compliance Options 2 and 3 of 40 fatalities eliminated and roughly 3,500 nonfatal injuries prevented or reduced in severity from improved stopping distance. The safety impact of Compliance Option 1 would be 43 fatalities and about 3,700 nonfatal injuries prevented.
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
|MAIS0||MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Fatal|
Flat Tires and Blowouts
There are many factors that influence crashes of these types. For blowouts, there is speed, tire pressure, and the load on the vehicle. Blowouts to the front tire can cause roadway departure, or can cause a lane change resulting in a head-on crash. Blowouts in a rear tire can cause spinning out and loss of control. As discussed in the target population section, a target population can be estimated for tire problems, but the agency doesn’t know how many of these crashes are influenced by under-inflation. However, reducing under-inflation will be a real benefit in reducing flat tires and blowouts. The agency’s best estimates of these effects are discussed below.
The target population is 414 fatalities and 10,275 non-fatal injuries that occur annually in light vehicles in which the cause of the crash is a flat tire/blowout. It is difficult to determine the impacts of under-inflation. Puncture is the most common reason for a blowout. However, there are also many cases where a tire is punctured, loses air, and then fails later after being driven a distance under-inflated. In these cases, a TPMS would provide information of the low tire pressure before the tire failed. We are assuming that under-inflation is involved in 20 percent of the cases that caused the crash. At the same time, we realize that the influence that under-inflation has on the chances of a blowout are influenced by the properties of the tire. Thus, we believe that better tires could take care of 50 percent of this problem and are assigning this value to the tire upgrade rulemaking. In conclusion, it is estimated that 41 fatalities (414 x .2 x .5) and 1,028 injuries are caused annually by flat tires/blowouts, where under-inflation is the cause of the flat tire/blowout. At the same time we estimate that there are 41 fatalities and 1,028 injuries in the target population for better tires brought about by the tire upgrade rulemaking. The agency assumes that 90 percent of drivers will fill their tires back to placard pressure when given a warning. For this situation, the agency does not believe that the steady state analysis has any impacts on the benefits. Any tire above the warning level is not very susceptible to a flat tire, and it probably doesn’t matter whether the tire is at a placard level of 30 psi or at a steady state level of say 27 psi in terms of its likelihood of failing due to a flat tire. We also apply a .99 factor to take into account the one percent of the fleet that already has a direct measurement system.
Thus, the benefits for flat tires/blowouts for Compliance Options 1 through 3 are the same:
37 lives saved (41 x .90 x .99) and 916 injuries reduced (1,028 x .90 x .99)
Non-fatal injuries are divided into the AIS levels based on the injury levels in 1995-98 NASS-CDS distribution of injuries in vehicles with flat tires causing the crash. These are: AIS 1 = 80.1 percent, AIS 2 = 14.4 percent, AIS 3 = 3.5 percent, AIS 4 = 1.5 percent, AIS 5 = 0.5 percent.
Total Quantifiable Safety Benefits
Table V-24 provides the total quantifiable safety benefits by Compliance Option adding together the benefits for skidding/loss of control, stopping distance, and flat tires/blowouts.
|MAIS 1||MAIS 2||MAIS 3||MAIS 4||MAIS 5||Non-Fatal Total||Fatal|
Fuel Economy Benefits
Correct tire pressure will improve a vehicles’fuel economy. Current radial tires are a vast improvement over the old-fashioned bias-ply tires, yet they still use more fuel when they are run under-inflated, although not as much as bias-ply tires. According to a 1978 report , fuel efficiency is reduced by one percent (1%) for every 3.3 pounds per square inch (psi) of under-inflation. More recent data provided by Goodyear indicates that fuel efficiency is reduced by one percent for every 2.96 psi of under-inflation, fairly close to the 1978 estimate.For this analysis, we assumed that there was no effect of tire over-inflation, and that savings only started once the warning went on. In other words, if the placard pressure were 30 psi, and a warning were given at 22.5 psi (25 percent below placard), no benefits are assumed for those vehicles that have tires with lowest pressure above 22.5 psi. However, there is a benefit for those vehicles with continuous displays, in that their steady state psi position is higher in Compliance Option 1, than in Compliance Options 2 and 3. Data from the tire pressure survey was used to estimate the average under-inflation of all 4 tires for those vehicles for which a warning would be given. Table V-25 provides the average under-inflation and the percentage of the fleet that would get a warning by the TPMS. All the Compliance Options are the same because they give a warning at 25 percent below placard.
|Passenger Cars Average psi below placard of those vehicles warned||Percent of Fleet Affected||Light Trucks Average psi below placard of those vehicles warned||Percent of Fleet Affected|
|Compliance Option 1||6.8 psi||26%||8.7 psi||29%|
|Compliance Option 2||6.8 psi||26%||8.7 psi||29%|
|Compliance Option 3||6.8 psi||26%||8.7 psi||29%|
Tables V-26 and V-27 show the weighted vehicle miles traveled by age of vehicle for passenger cars and light trucks. They also show the 7 percent discount rate and the assumed price of gasoline. The projected price of gasoline was taken from a DOE projection from January 2001 . It excludes fuel taxes, at $0.38 per gallon, since these are a transfer payment and not a cost to society. The projections were for gasoline prices to steadily decline from 2001 through about 2005 when they will level off. A second group of adjustments were made to the price of gasoline to account for environmental costs and international oil market costs.
One product of the combustion of hydrocarbon fuels, such as gasoline and diesel, is CO2. The environmental and economic consequences of these releases are not included in the price of gasoline. While there are estimates of these consequences in the literature, the administration has not taken a position on their costs. Using estimates from the literature would result in very little savings on a per vehicle basis and they have not been included in this analysis. A second environmental cost of gasoline use relates to the hydrocarbon and toxic chemical releases from the gasoline supply chain, including oil exploration, refining, and distribution. Marginal costs of these activities combined have been estimated at $0.02 per gallon. 
The Organization of Petroleum Exporting Countries (OPEC) operates as a cartel that restricts the supply of oil to escalate the price above the free-market level. The greater the consumption of oil, the higher will be the price. Since the higher price of oil applies to all oil imports from OPEC, not just the increased oil use, the financial cost to the United States exceeds the market payment for the increased amount. Leiby et al.,  estimated this impact to be $3.00 per barrel. This equates to $0.07 per gallon ($3/42 gallons per barrel). The impact is dependent upon the amount of oil saved. The $0.07 per gallon is about right for the savings of this program.
Thus, the price of gasoline has been reduced by $0.38 per gallon to account for taxes, and has been increased by $0.09 per gallon to account for environmental and economic considerations.
|Survival Probability||Weighted Vehicle
|Survival Probability||Weighted Vehicle
The baseline miles-per-gallon figure for cars was 27.5 mpg at placard inflation, and for light trucks was 22.2 mpg (the MY 2007 light truck standard) at placard inflation. A sample calculation for passenger cars for Compliance Option 1 is:
The average of all four tires on a passenger car that would be warned based on our survey would be 6.8 psi lower than placard. The average steady state condition after TPMS are in place would be 3.0 psi lower than placard. Thus, the incremental steady state improvement of the TPMS is 3.8 psi (6.8–3.0). Since 1 percent fuel efficiency is equivalent to 2.96 psi lower, the average passenger car with a warning would get 1.0128 percent (3.8/2.96) higher fuel economy when re-inflated. With a baseline of 27.5 mpg, the average fuel economy of those vehicles warned that increased their tire pressure up to placard would be 27.5 * 1.012838 = 27.853 mpg. Based on our estimated vehicle miles traveled by age, scrappage by age, a 7 percent present value discount rate and estimated fuel costs per year, the baseline passenger car (at 27.5 mpg discounted by 15 percent to account for real on-road mileage) would spend $3,968.88 present value for fuel over its lifetime. Those drivers warned who filled up to placard pressure and achieved 27.853 mpg (discounted by 15 percent to account for real on-road mileage) would spend $3,918.58 for fuel over their car’s lifetime. The difference is $50.30. Since 26 percent of the fleet get a warning, and it is assumed that 90 percent of the drivers would fill their tires to placard, the average benefit is $11.77 ($50.30*0.26*0.90). The estimated benefit for each subgroup under the different compliance options is shown in Table V-28, under a 7 percent and 3 percent discount rate.
|Passenger Cars||Light Trucks|
|3% Discount||7% Discount||3% Discount||7% Discount|
|Compliance Option 1||$14.35||$11.77||$31.28||$24.53|
|Compliance Option 2||$11.74||$9.62||$25.95||$20.34|
|Compliance Option 3||$11.74||$9.62||$25.95||$20.34|
Weighting light trucks (9/17)  and passenger cars (8/17) and taking into account the one percent of the fleet that already has a direct measurement system results in the following overall benefit in fuel economy shown in Table V-29.
|Average Passenger Vehicle
3% Discount Rate
|Average Passenger Vehicle
7% Discount Rate
|Compliance Option 1||$23.08||$18.34|
|Compliance Option 2||$19.07||$15.14|
|Compliance Option 3||$19.07||$15.14|
Since there are fuel economy improvements, there are comparable savings in gasoline usage. Fewer gallons of gasoline used mean fewer emissions. Table V-30 shows the lifetime gallons of gasoline saved per vehicle for the different Compliance Options. These per vehicle estimates are multiplied by 17 million vehicles. Assuming constant vehicle sales from year to year, once all vehicles in the fleet meet the standard, the annual gasoline savings are equal to the lifetime savings of fuel of one model year. The rule of thumb for equating gasoline savings to emissions savings used by the Department of Energy is that for every billion gallons of gasoline saved, emissions are reduced by 2.4 million metric tons carbon equivalent (MMTCE).
|Passenger Cars||Light Trucks||Average Savings
Per Light Vehicle
|Compliance Option 1||16||37||27|
|Compliance Option 2||13||31||22|
|Compliance Option 3||13||31||22|
|Average Gasoline Savings
Per Light Vehicle
|Annual Gasoline Savings
(Millions of Gallons)
|Annual Emissions Reduction
|Compliance Option 1||27||459||1.10|
|Compliance Option 2||22||374||0.90|
|Compliance Option 3||22||374||0.90|
Driving at lower inflation pressure impacts the rate of tread wear on tires. This will cause tires to wear out earlier than necessary and decrease tread life. When a tire is under-inflated, it puts more pressure on the shoulders of the tire and does not wear correctly. This analysis will attempt to quantify the impact of increased tread wear on consumer costs.
Based on data provided by Goodyear (see Docket No. NHTSA-2000-8572-26), the average tread life of tires is 45,000 miles and the average cost is $61 per tire (in 2001 dollars).
For Compliance Option 1
Assuming a direct measurement system, the TPMS warns the driver anytime a tire is 25 percent or more below the placard and the driver inflates all of the tires back to the placard levels, then we can estimate the impact on tread life using the following calculations.
Goodyear provided data estimating that the average tread wear dropped to 68 percent of the original tread wear if tire pressure dropped from 35 psi to 17 psi. Goodyear also assumed that this relationship was linear. Thus, for every 1 psi drop in inflation pressure, tread wear would decrease by 1.78 percent [(100-68%)/(35-17psi)]. These effects would take place over the lifetime of the tire. In other words, if the tire remained under-inflated by 1 psi over its lifetime, the tread wear would decrease by 1.78 percent or about 800 miles (45,000*0.178).
Data from our tire pressure survey indicated that 1,575 out of 5,967 passenger car tires (26 percent) had at least one tire under-inflated by 25 percent or more below the placard level. The average under-inflation of the 4 tires for these vehicles was 6.8 psi. Based on our steady state assumptions discussed earlier, the average psi of the fleet under Compliance Option 1 would improve by 3.8 psi (placard pressure is 30 psi, steady state pressure under Compliance Option 1 is 27.0 psi, thus a 3.0 psi difference; 6.8 – 3.0 = 3.8 psi improvement). Thus, on average, passenger cars lose an estimated 3,040 miles (3.8 * 800 miles) of tread life for each tire due to the way they are currently under-inflated that could be remedied under Compliance Option 1 if everyone filled all their tires back up to the placard pressure when they were notified by a TPMS. If we assume that 90 percent of the people actually inflate their tires properly, then on average 2,736 miles of tread life would be saved per tire.
If the average current lifetime of tires is 45,000 miles at current inflation levels, the average lifetime could be 47,736 miles with a TPMS. The agency estimates that the average lifetime per passenger car is 126,678 miles. Thus, currently the average car would have 3 sets of tires on their car over its lifetime (new, at 45,000 miles, and at 90,000 miles) and with TPMS the average car would have 3 sets of tires purchased (new, at 47,736 miles, and at 95,472 miles). The benefit to consumers is the delay in purchasing those tires and getting interest on that money at an assumed 3 percent or 7 percent rate of return. Using a mid-year 3 percent and 7 percent interest rate and discount rate, the discounted present value of these delayed tire purchases is estimated to be $5.71 at a 3 percent discount rate and $10.23 at a 7 percent discount rate for those passenger cars that would be notified by a TPMS that they are under-inflated. Since 26 percent would be notified, the present discounted benefits are $1.48 ($5.71 * 0.26) at a 3 percent discount rate, $2.66 ($10.23 * 0.26) at a 7 percent discount rate, and 711 miles (2,736 * 0.26) of tread life.
For light trucks, data from our tire pressure survey indicated that 1,148 of 3,950 light truck tires (29 percent) had at least one tire under-inflated by 25 percent or more compared to the placard. The average under-inflation of the 4 tires for these vehicles was 8.7 psi. Based on our steady state assumptions discussed earlier, the average psi of the fleet under Compliance Option 1 would improve by 5.2 psi (placard pressure is 35 psi, steady state pressure under Compliance Option 1 is 31.5 psi, thus a 3.5 psi difference; 8.7 – 3.5 = 5.2 psi improvement). Thus, on average, light trucks lose an estimated 4,160 miles (5.2*800) of tread life for each tire due to the way they are currently under-inflated that could be remedied if everyone filled all their tires back up to the placard pressure when they were notified by a TPMS. If we assume that 90 percent of the people actually inflate their tires properly, then on average 3,744 miles of tread life would be saved per tire.
If the average current lifetime of tires is 45,000 miles at current inflation levels, the average lifetime could be 48,744 miles with a TPMS. The agency estimates that the average lifetime per light truck is 153,706 miles. Thus, the average light truck would have 4 sets of tires on their truck over its lifetime (new, at 45,000 miles, at 90,000 miles, and at 135,000 miles) and with a TPMS the average light truck would have four sets purchased (new, at 48,744 miles, at 97,488, and at 146,232 miles). Using the same methodology as for passenger car tires, the benefit in delaying purchasing tires is estimated to be a present discounted benefit of $23.33 at the 3 percent discount rate and $31.54 at the 7 percent discount rate. Since in 29 percent of the vehicles at least one tire is under-inflated by 25 percent or more, the average benefit for light trucks is estimated to be $6.77 ($23.33 * 0.29) at the 3 percent interest and discount rate, $9.15 ($31.54 * 0.29) at the 7 percent interest and discount rate, and 1,086 miles (3,744 * 0.29) of tread life.
The weighted tread life savings for passenger cars and light trucks after considering current compliance for Compliance Option 1 is $4.24 [($1.48 * 8/17) + ($6.77 * 9/17)]*.99 at the 3 percent interest rate and discount rate and $6.03 [($2.66 * 8/17) + ($9.15 * 9/17)]*.99 at the 7 percent interest rate and discount rate and 900 [(711 * 8/17) + (1,086 * 9/17)]*.99) miles of tread life.
For Compliance Options 2 and 3
Data from our tire pressure survey indicated that 1,575 out of 5,967 passenger car tires (26 percent) had at least one tire under-inflated by 25 percent or more below the placard level. The average under-inflation of the 4 tires for these vehicles was 6.8 psi. Based on our steady state assumptions discussed earlier, the average psi of the fleet under Compliance Options 2 and 3 with direct systems would improve by 3.1 psi (placard pressure is 30 psi, steady state pressure under Compliance Options 2 and 3 is 26.3 psi, thus a 3.7 psi difference; 6.8 – 3.7 = 3.1 psi improvement). Thus, on average, passenger cars lose an estimated 2,480 miles (3.1 * 800 miles) of tread life for each tire due to the way they are currently under-inflated that could be remedied if everyone filled all their tires back up to the placard pressure when they were notified by a TPMS. If we assume that 90 percent of the people actually inflate their tires properly, then on average 2,232 miles of tread life would be saved per tire.
If the average current lifetime of tires is 45,000 miles at current inflation levels, the average lifetime could be 47,232 miles with a TPMS. The agency estimates that the average lifetime per passenger car is 126,678 miles. Thus, currently the average car would have 3 sets of tires on their car over its lifetime (new, at 45,000 miles, and at 90,000 miles) and with TPMS the average car would have 3 sets of tires purchased (new, at 47,232 miles, and at 94,464 miles). The benefit to consumers is the delay in purchasing those tires and getting interest on that money at an assumed 3 and 7 percent rate of return. Using a mid-year 3 and 7 percent interest rate and discount rate, the discounted present value of these delayed tire purchases is estimated to be $4.68 at the 3 percent discount rate and $8.42 at the 7 percent discount rate for those passenger cars that would be notified by a TPMS that they are under-inflated. Since 26 percent would be notified, the present discounted benefits are $1.22 ($4.68 * .26) at the 3 percent discount rate, $2.19 ($8.42 * .26) at the 7 percent discount rate and 580 miles (2,232 * 0.26) of tread life.
For light trucks, data from our tire pressure survey indicated that 1,148 of 3,950 light truck tires (29 percent) had at least one tire under-inflated by 25 percent or more compared to the placard. The average under-inflation of the 4 tires for these vehicles was 8.7 psi. Based on our steady state assumptions discussed earlier, the average psi of the fleet under Compliance Options 2 and 3 would improve by 4.3 psi (placard pressure is 35 psi, steady state pressure under Compliance Options 2 and 3 is 30.6 psi, thus a 4.4 psi difference; 8.7 – 4.4 = 4.3 psi improvement). Thus, on average, light trucks lose an estimated 3,440 miles (4.3*800) of tread life for each tire due to the way they are currently under-inflated that could be remedied if everyone filled all their tires back up to the placard pressure when they were notified by a TPMS. If we assume that 90 percent of the people actually inflate their tires properly, then on average 3,096 miles of tread life would be saved per tire.
If the average current lifetime of tires is 45,000 miles at current inflation levels, the average lifetime could be 48,096 miles with a TPMS. The agency estimates that the average lifetime per light truck is 153,706 miles. Thus, the average light truck would have 4 sets of tires on their truck over its lifetime (new, at 45,000 miles, at 90,000 miles, and at 135,000 miles) and with a TPMS the average light truck would have four sets purchased (new, at 48,096 miles, at 96,192, and at 144,288 miles). Using a mid-year 3 and 7 percent interest rate and discount rate, the discounted present value of these delayed tire purchases is estimated to be $18.76 at the 3 percent discount rate and $26.02 at the 7 percent discount rate for those light trucks that would be notified by a TPMS that they are under-inflated. Since in 29 percent of the vehicles at least one tire is under-inflated by 25 percent or more, the average benefit for light trucks is estimated to be $5.44 ($18.76*0.29) at the 3 percent interest and discount rate and $7.55 ($26.02 * 0.29) at the 7 percent interest and discount rate and 896 miles (3,096 * 0.29) of tread life.
The weighted tread life savings for passenger cars and light trucks after considering current compliance for Compliance Options 2 and 3 are $3.42 [($1.22 * 8/17) + ($5.44 * 9/17)]*.99 at the 3 percent interest rate and discount rate and $4.98 [($2.19 * 8/17) + ($7.55 * 9/17)]*.99 at the 7 percent interest rate and discount rate and 740 miles [(580 * 8/17) + (896 * 9/17)]*.99) of tread life per tire.
Table V-32 shows the tread life savings per vehicle after considering current compliance.
|3 % Discount Rate||7% Discount Rate|
|Compliance Option 1||$4.24||$6.03|
|Compliance Option 2||$3.42||$4.98|
|Compliance Option 3||$3.42||$4.98|
There are other potential non-quantified benefits of increasing tread wear. Some people would not have to purchase the last set of tires for a vehicle if they were going to scrap the vehicle soon, or if it were totaled in a crash shortly before they were going to purchase new tires. So, there will be cases where the total purchase price of tires $244 ($61 per tire * 4) will be saved. However, we can’t estimate the frequency of that occurrence.
Property Damage and Travel Delay Savings
Reduced stopping distance, blowouts, and loss of control in skidding will prevent crashes and reduce the severity of impacts and the injuries that result. Property damage and travel delay will also be mitigated by these improvements. To the extent that crashes are avoided, both property damage and travel delay will be completely eliminated. Crashes that still occur but at less serious impact speeds will still cause property damage and delay other motorists, but to a lesser extent than they otherwise would have.
NHTSA has developed data on the cost of both travel delay and property damage stratified by injury severity on a per-person injured basis . Travel delay is defined as the value of lost time experienced by motorists not involved in a crash, but who are delayed in traffic congestion resulting from these crashes. Property damage is the value of vehicles, cargo, and other items damaged in traffic crashes. The number of injuries prevented, as well as the number of PDO crashes prevented in out of control skids, has already been estimated in Chapter V. An estimate of total PDO involved vehicles was derived as follows.
Table V-33 summarizes the injuries that would be prevented by TPMS. To estimate the impact on PDOs, it will be assumed that PDOs are reduced in the ratio of overall occurrence of PDOs to injuries. The PDO cost data mentioned above is expressed in terms of per damaged vehicle. Therefore, PDOs will be measured in these same units. The number of vehicles involved in crashes that produced the injury savings in Table V-34 is estimated based on the average number of police reported injuries per vehicle (1.35) . The results are shown on the Injury Vehicles line in that Table. For the stopping distance and blowout categories, these estimates were then multiplied by the overall ratio of PDO involved vehicles to injury involved vehicles  to estimate the total number of police reported PDOs that would be prevented. The out of control skidding category was handled differently because a specific estimate of PDO crashes prevented was derived in Chapter V. For this category, the number of crashes prevented (9,994) was multiplied by the overall ratio of PDO vehicles per crash .
PDOs are notoriously underreported. Many localities don’t even record crashes unless they involve some variable damage threshold and often drivers involved in single vehicle PDOs will leave before police arrive. Overall, NHTSA estimates that only 52% of the vehicles involved in PDO crashes get reported . An adjustment was made to the 3 PDO vehicle totals to reflect this. The final results are shown in the Total PDO Vehicles line of Table V-33.
The MAIS0 line in Table V-33 represents uninjured ccupants that are present in vehicles that avoid crashes due to TPMS. They are estimated based on the ratio of uninjured occupants to injured occupants in police reported crashes . They are included here because the unit property damage and travel delay costs used in this analysis were distributed over all occupants and to fully account for all savings in these avoided crashes they must also be accounted for.
|Crash Involvements Prevented By TPMS||Property Damage & Travel Delay|
|Injured Persons in Crashes Due To:
|Ratio PDO/Inj Veh||4||4|
|Total P.R. PDO Veh||4589||17511||2807|
|Total PDO Vehicles||8825||33675||5398||47897||$2,352||$112,657,838|
|Total Including PDOs||$153,160,408|
Table V-33 also lists the total per-case Travel Delay and Property Damage costs stratified by injury severity . The costs are expressed as per-injured person for all injury levels, and per damaged vehicle for PDOs. These unit costs were multiplied by the corresponding injury and PDO incidence savings to estimate total savings in travel time and property damage from crashes prevented by TPMS.
The impact on non-preventable crashes is more subtle and measuring it requires some assumptions regarding the nature of injury mitigation. The injuries prevented in non-preventable crashes are summarized in Table V-34. These represent the net impact on total injuries in each severity category after the severity of each crash was reduced. For all but minor injuries, this would typically involve a tumble-down effect, where injuries are reduced to a lower severity level rather than being eliminated entirely. Since the savings are a net result of this process, this means that the total number of injuries reduced in each category is really the sum of the savings in that category plus those injuries that tumbled-down into that category from a more severe level. To simulate this, it will be assumed that each injury mitigated will fall only one level. The second column in Table V-34 shows the resulting gross savings for each severity level. In the third column, the difference in unit costs of travel delay and property damage between the specific injury level and the next highest level are shown. These numbers represent the change in these costs that occurs from each reduction in injury levels. Total costs for each level are the product of these unit costs and the total injuries saved at that level.
Total travel delay and property damage cost savings from non-preventable crash severity mitigation is thus estimated to total $5.6 million. Total savings from all crash types, including preventable injury and PDO crashes would total $158,767,754. Since these savings would occur over the life of the vehicle, a discount factor will be applied to express their present value. At a 3% discount rate, the present value of total travel delay and property damage savings would be $130,713,492 (.8233 combined factor). At a 7% rate, this value would be $104,786,718 (.6600 combined factor). These are the estimates for Compliance Options 2 and 3.
For Compliance Option 1, the same methodology results in total savings from all crash types, including preventable injury and PDO crashes would total $160,871,765. At a 3% discount rate, the present value of total travel delay and property damage savings would be $133,445,724 (.8233 combined factor). At a 7% rate, this value would be $106,175,365 (.6600 combined factor).
Under-inflation affects many different types of crashes. These include crashes which result from:
The agency has quantified the effects of under-inflation in a crash involving skidding and loss of control, flat tires and blowouts, and the reduction in stopping distance. However, it cannot quantify the effects of under-inflation on hydroplaning and overloading the vehicles. The primary reason that the agency can’t quantify these benefits is the lack of crash data indicating tire pressure and how large of a problem these conditions represent by themselves, or how often they are contributing factors to a crash. The agency has just starting collecting tire pressure in its crash data investigations.
Skidding and/or loss of control from hydroplaning
The conditions that influence hydroplaning include speed, tire design, tread depth, water depth on the road, load on the tires, and inflation pressure. At low speeds (less than about 50 mph), if your tires are under-inflated, you actually have more tire touching the road. However, hydroplaning does not occur very often at speeds below 50 mph, unless there is deep water (usually standing water) on the road. As you get to about 55 mph and the water pressure going under the tire increases, an under-inflated tire has less pressure in it pushing down on the road and you have less tire-to-road contact than a properly inflated tire as the center portion of the tread gets lifted out of contact with the road. As speed increases to 70 mph and above and water depth increases due to a severe local storm with poor drainage, the under-inflated tire could lose 40 percent of the tire-to-road contact area compared to a properly inflated tire. The higher the speed (above 50 mph) and the more under-inflated the tire is, then the lower the tire-to-road contact and the higher is the chance of hydroplaning.
Tread depth has a substantial impact on the probability of hydroplaning. If you make a simplifying assumption that the water depth exceeds the capability of the tread design to remove water (which most likely would occur with very worn tires), then an approximation of the speed at which hydroplaning can occur can be estimated by the following formula:
Hydroplaning speed = 10.35 x inflation pressure 
Under this assumption of water depth exceeding the capability of the tread
design to remove water:
At 30 psi, hydroplaning could occur at 56.7 mphThis is presented to show the relative effect of inflation pressure on the possibility of hydroplaning.
At 25 psi, hydroplaning could occur at 51.8 mph
At 20 psi, hydroplaning could occur at 46.3 mph.
Overloading the vehicle
When a vehicle is overloaded, (too much weight is added for the suspension, axle, and tire systems to carry) and the tires are under-inflated, there is an increased risk of tire failures. This can result in a loss of control of the vehicle.
Potential Benefits for Antilock Brake Systems
If a manufacturer decided that the difference in the cost between an indirect and direct TPMS was enough to make antilock brakes a marketable feature for that vehicle, then it might decide to increase its use of ABS and use an indirect TPMS to meet the phase-in part of the final rule. The agency has been analyzing the safety impacts of ABS for several years. The initial findings  were mixed. Fatal crash involvements in multi-vehicle crashes on wet roads and fatal crashes with pedestrians and bicyclists were significantly reduced. However, these reductions were offset by a statistically significant increase in the frequency of single vehicle, run-off-road crashes (rollovers or impacts with fixed objects). The run-off-road crashes were surprising in view of the good performance of ABS in stopping tests conducted by the agency and others. The agency has spent several years trying to determine why run-off-road crashes have increased with ABS, without a satisfactory answer.
Two more recent studies of ABS have found no statistically significant fatality improvement with ABS. The Farmer study from IIHS  found the results shown in Table V-35. (A ratio of 1.0 means there is no effect on fatalities. Less than one is a reduction in fatalities, more than 1.0 is an increase in fatalities. In order for the results to be statistically significant, the confidence bounds would have to be both below 1.0 or both above 1.0). The only statistically significant findings were that fatalities went up in non-GM cars in calendar years 1986-1995 and overall from 1986-1998.
|All crashes||95 percent confidence bounds|
|GM cars in 1993-95||1.03||.94||1.12|
|GM cars in 1996-98||.96||.87||1.05|
|GM cars in 1993-98||.99||.93||1.05|
|Non-GM cars in 1986-95||1.16 (Significant)||1.06||1.27|
|Non-GM cars in 1996-98||.91||.77||1.06|
|Non-GM cars in 1986-98||1.09 (Significant)||1.01||1.18|
Farmer’s theory is that people learned how to use ABS better in calendar years 1996-98 and they were no longer overinvolved in run off the road fatal crashes. Farmer never states that ABS reduced fatalities. His statement on the GM cars for 1996-98 is "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".
The second recent analysis by Ellen Hertz (NHTSA) , in which she included optional ABS to get more cases, also resulted in no overall statistically significant findings for fatalities. ABS effects were examined separately for passenger cars and light trucks for five types of crashes (frontal impacts, side impacts, rollover, run-off-road, and pedestrian). The only statistically significant finding was that fatalities in light truck rollover crashes went up in ABS vehicles compared to non-ABS vehicles (see Table V-36). In this study, a negative is an improvement in safety (fewer fatalities) and a positive is an increase in fatalities.
|Point||95 percent confidence bounds|
|Frontal – PC||-4.9%||-19.9%||11.5%|
|Frontal – LTV||17.9||-7.1||49.6|
|Side Impact – PC||32.4||-1.0||77.2|
|Side Impact – LTV||-0.3||-42.3||72.2|
|Rollover – PC||12.3||-17.2||52.2|
|Rollover – LTV||106.5 (Significant)||49.2||185.9|
|Run-Off-Road – PC||-13.4||-28.1||4.2|
|Run-Off-Road – LTV||21.8||-12.6||69.5|
|Pedestrian – PC||-0.4||-16.3||18.4|
|Pedestrian – LTV||-22.7||-50.1||19.6|
The Hertz study did find that antilock brakes had an overall effect of reducing crashes, but not fatalities.If NHTSA believed that antilock brakes were cost/beneficial, we would consider requiring them to be installed. We have not considered requiring antilock brakes because we have not been able to show that they are beneficial in reducing fatalities. Reducing their costs, by offsetting the costs with a TPMS, does not affect our conclusions to date that we have not been able to prove that ABS reduces fatalities.
 Evaluation of Techniques for Reducing In-use Automotive Fuel Consumption; The Aerospace Corporation, June 1978. Original reference from Goodyear, pp 3-45.
 DOE Energy Information Administration, Annual Energy Outlook 2001, Table A3, Energy Prices by Sector.
 "Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards", National Research Council, July 2001, Pages 5-5 to 5-6.
 "Oil Imports: An Assessment of Benefits and Costs", P.N. Leiby, D.W. Jones, T.R. Curlee, and L. Russell, 1997, ORNL-6851, Oak Ridge National Laboratory, Oak Ridge, Tenn.
 We assume sales of 8 million passenger cars and 9 million light trucks for a total of 17 million vehicles annually.
 Blincoe et al, Ibid.
 Blincoe et al, Ibid.
 "Mechanics of Pneumatic Tires" edited by Samuel K. Clark of the University of Michigan, published by NHTSA, printed by the Government Printing Office in 1981.
 "Preliminary Evaluation of the Effectiveness of Antilock Brake Systems for Passenger Cars", NHTSA, December, 1994, DOT HS 808 206.
 "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.
 "Analysis of the Crash Experience of Vehicles Equipped with All Wheel Antilock Braking Systems (ABS) – A Second Update Including Vehicles with Optional ABS", NHTSA, September 2000,
Table of Contents