5. SAFETY ANALYSES

This section of the report details the findings of the economic analysis, which estimates the societal benefits and costs of implementing the NTSB recommendations. The primary benefits of the recommendations are the safety impacts of crash prevention, avoidance of high crash costs, avoidance of fines due to violations associated with brakes being OOA, and avoidance of costs resulting from OOS brake violations due to the improved manner in which brakes are tested and maintained on commercial motor vehicles. The primary costs of the recommendations include the following: the costs of driver and brake inspector training, per trip driver inspection costs, and the costs associated with brake inspector testing and certification.

Analysis of these benefit and cost items above have been quantitatively estimated over a 10-year analysis period. Since most of the costs are expected to occur on an annual basis, given that the NTSB recommendations are for an ongoing program while only the initial training costs will be borne in the first year, the examination is based on a 10-year analysis time frame, from 2005 to 2014. There are no capital costs (i.e., outlays for capital investment in equipment, buildings or property) associated with implementing the NTSB recommendation for certifying commercial motor vehicle brake technicians. A discount rate of 7 percent was used to construct the net present value estimate in the baseline scenario as prescribed in the OMB Circular A-94 (OMB, 2002). In order to examine the sensitivity of the analysis to discount rates, scenarios using discount rates of 4 and 10 percent were also examined. All benefit and cost items were combined to compute benefit-cost ratios (BCRs) as depicted in Figure 5-1. To the extent that benefits exceed costs, the BCR exceeded 1.0. BCRs of less than 1.0 indicate negative economic returns to society.

5.1 Background Crash Data for Benefits Assessment

This subsection establishes CMV brake data estimates for use in benefit-cost analyses. Estimates are provided for crashes prevented and mitigated through the implementation of an enhanced brake maintenance, certification, and inspection program for technicians and an enhanced brake inspection program for drivers. Estimates of the number of crashes that potentially could either be mitigated or prevented with the new program are based on a direct analysis of the LTCCS data, as well as the values provided by the motor carrier, insurance, and crash investigator interviews. Developing the predicted effectiveness of the enhanced inspection program in preventing or mitigating brake related crashes relies on engineering judgment. The study team drew on its expertise in brake function, crash analysis, and familiarity with how brakes operate in a crash environment.

Figure 5-1. Data Elements Combined to Determine Societal Benefit-cost Ratios

5.2 Benefits Assessment

The benefits associated with the NTSB recommendations are those related to safety as measured by the prevention of crashes involving CMVs, the avoidance of the monetary costs for these CMV crashes, the reduced number of OOS brake violations, and reduced levels of fines associated with violations resulting from brakes being OOA. These benefits are explored individually in the following sections.

5.2.1 Safety Impacts

This section examines the approach used to estimate the number of CMV crashes that would be prevented if improved brake maintenance and inspection procedures were implemented and the monetary benefits associated with avoiding CMV crashes. Data used in the study, along with sources, are also cited in this section of the report.

5.2.1.1 Avoidable Truck Crashes

Studies and databases surveyed for the project were examined to determine the most appropriate sources to use for estimating the number of truck crashes that may have been caused by brake problems. Databases describing the results of post crash inspections seemed to be most appropriate. The LTCCS and the MCMIS inspection file both include the results of post crash inspections. The LTCCS data files currently include data for 1,070 trucks (includes some data collected for trucks during a pilot study). The LTCCS shows that brake problems exist in about 30 percent (321 out of 1,070) of the post-crash inspections conducted for the study. MCMIS post-crash inspection data show that brake defects or deficiencies existing prior to the crash were found in only about 8 percent of the post-crash inspections.

An examination of data included in other studies in the project literature search indicates that the percentage for brake problems included in the LTCCS data analysis provides a reasonable estimate of brake problems that may cause a crash. For example, CVSA provides a detailed summary of data from Operation Air Brake for the years 1998 to 2004. During those years, Operation Air Brake checked 66,941 trucks during roadside inspections. Of these, 11,360 (16.97 percent) have had an OOS brake condition. The percentage of brake problems found in inspections of trucks after a crash would be expected to be much higher than those found in inspections of operating vehicles. In another study, 667 crashes involving 734 large trucks (greater than 10,000 pounds) in Washington State were investigated over a two-year period. When the sample was then limited to tractor trailers (about 60 percent of the crash sample), trucks with brake defects were present in 56 percent of the crash sample. This study did not use as representative a sample as the LTCCS and, therefore, may have found a higher percentage of brake problems than the LTCCS. In other words, the Washington State study did not use a representative statistical sampling approach to selecting the crashes to include and may have over-sampled crashes involving brake defects.

Due to the thoroughness of the LTCCS and the bounding results of other studies, the LTCCS was used for this study in order to estimate the number of brake-caused crashes that would be prevented or had their severity reduced if the brake training, certification, and maintenance program based on NTSB recommendations (described in Section 1.1) were introduced. The great majority of the crashes in the study involved trucks with air brakes. About 90.1 percent of the trucks inspected in the post-crash investigations had air brakes, about 7.6 percent had hydraulic brakes and the remainder, about 2.3 percent, had unknown brake types.

The LTCCS includes data for four types of truck brake problems related to a particular crash: brake failure, brake inoperable, brakes OOA, and brakes deficient. The problems are defined as follows:

• Out of adjustment: If any of the brakes were measured as OOA, then this variable would be recorded as present.
• Brakes inoperative: This problem is present if the brakes are not working for any reason. If the brakes are inoperative because they are severely OOA, they might be recorded in the OOA category.
• Brake system malfunction (failure): This variable establishes whether or not the vehicle experiences a braking system malfunction (total failure such as pedal to the floor) during the pre-crash phase (may not include a malfunction due to OOA).
• Brake system deficiency: Braking system deficiency records any problem other than brake OOA. It includes the following: worn pads, unmatched brakes, hose connection, air pressure, brake fade, etc.

In order to assess the contribution of each of the brake problems, the data were grouped in categories because, in some cases, more than one brake problem was identified for a particular vehicle. The brake problem types were grouped in a hierarchy according to the problem considered most important, or most likely to contribute to a crash: brake failure, brake inoperative, brake system deficiency, and brakes OOA.

Brake failure accounted for 12 crashes, although an additional brake problem also existed in two of the crashes. Similarly, the 23 "brakes inoperative" problems identified for a crash were grouped as one category despite the fact that 13 of the crashes also had another brake problem associated with them. For brake system deficiency, 97 of the 121 crashes where brake deficiency was listed had only that factor. Twenty-four of the crashes also included brakes OOA as a factor. Those 24 crashes were placed in the brake system-deficiency category because of the hierarchy used in the analysis. One hundred and sixty-five crashes were characterized by brakes OOA alone.

The next step was to determine the percentage of the crashes that were unlikely to be influenced by the truck's brakes. These were the crashes where another vehicle hit the truck from behind. About 23 percent of the crashes with brake problems fall into this category. Thus, the number of crashes eligible as candidates for causing a crash is reduced by 23 percent, reducing the 321 eligible crashes to 247. The distribution of the brake problems among the 321 crashes in the LTCCS is shown in Table 5-1 along with the number of eligible crashes with identified brake problems after the "truck was hit from behind" accident type was eliminated from the total crashes.

Next, the role that the various brake problems play in causing a crash was determined. Although more than one brake problem was sometimes associated with a crash, for this step, causation was linked to the most serious problem in the category. Because the data in the LTCCS provided insufficient detail concerning the behavior of individual brakes and the exact nature and severity of a particular brake problem, the Battelle Team was unable to estimate precisely the role that a brake mechanical problem may have played in leading to a crash. Therefore, since the team understood that other pre-crash factors such as human error also might be in play, engineering judgment was used to estimate the number of crashes that would either be avoided or mitigated if the brake problem could have been eliminated before the crash occurred. As shown in Table 5-1, the number in each brake category found in the first column already has been reduced by subtracting the percentage of crashes where the crash type involved another vehicle striking the CMV from behind. For each of the four categories of brake problems, engineering judgment was first applied to determine the likelihood that the crash would either have been prevented or mitigated if the brake problem had been eliminated. Next, the percentage of crashes that would have fallen into each of those two categories was estimated. The discussion below describes how the particular brake crash category was allocated into either the crash-prevented or crash impacts-mitigated categories.

• Brake system malfunction (failure): For this brake problem category, it was estimated that all (100 percent) of the nine crashes (excluding where the truck was hit from behind) would have been prevented if the total brake failure had been prevented through proper maintenance and inspection.
• Brakes inoperative: For the brakes inoperative category, where the brakes are not working, it was assumed that by preventing these problems, all 18 crashes (excluding where the truck was hit by another vehicle from behind) would have either been avoided or mitigated through proper maintenance and inspection. This percentage was further assumed to be about 50 percent where the crash would have been avoided (nine crashes) and 50 percent where the crash would have been mitigated (nine crashes).
• Brake system deficiency: The team evaluated this category to be the third most serious of the four brake problem categories. The team applied engineering judgment to conclude, based on the nature of brake problems in this group, that about half (50 percent) of the 93 crashes in this category (47 crashes) would either be avoided or mitigated if the brake system deficiency had been prevented through proper maintenance and inspection. It was further assumed that about half of the 47 crashes would be avoided and half mitigated if brake problems had been prevented.
• Out of adjustment: The team estimated that about 25 percent of the crashes identified in the brake OOA category (32 out of 127 crashed) would have been prevented or mitigated if this problem had been eliminated through proper maintenance and inspection. This percentage was selected because the post-crash inspection required only one brake to be identified as OOA to be placed in this category. However, if only one brake was OOA, the stopping power of the vehicle would not have been significantly reduced, and consequently would not have been the cause of the crash. Battelle also estimated that if this braking problem were eliminated, 16 of the 32 crashes would have been avoided and 16 mitigated. A better estimate could be placed on this category if the number of axles on each vehicle and the number of OOA brakes were known; however, the LTCCS did not provide this level of detail.

Table 5-2 shows the four categories of brake problems with the number of crashes that could have been caused by brake problems and the breakdown into whether preventing the brake problem would have resulted in either avoiding a crash or reducing a crash's severity. For the purposes of this analysis, the assumption was made that if a crash were mitigated by fixing or preventing a brake problem, then the crash impacts would have been reduced by 50 percent.

In order to estimate the percentage of truck crashes that could be averted by the implementation of an enhanced training and brake maintenance program, the number of crashes shown under the "Number in Category" column in Table 5-2 must first be converted to the "Number of Crashes Caused by Problem" as shown in column 3 of the table. The next step is to convert the number of crashes caused by the problem to either "Number of Crashes Avoided" (column 4) and/or "Crashes with Consequences Reduced" (column 5). For example, in order to calculate the numbers in Table 5-2 for "Brakes Inoperative" (column 1), the project team used the assumptions found under each brake problem category and discussed above. Therefore, for "Brake Inoperative," this is considered the second most serious brake problem and all of the eligible crashes characterized by this problem and found in the "Number in Category" column were estimated to have caused crashes. Table 5-2 shows that the 18 crashes in the "Number in Category" column remain 18 in the "Number of Crashes Caused by Problem" column. The next step is to allocate the number of estimated crashes caused by the problem to either crashes avoided or crashes with the consequences reduced. Based on the discussion of brakes inoperative above, the assumption was made that if the brake problem could have been prevented or repaired before a crash, then half of the crashes would have been avoided (9 crashes) and half of the crashes would have been mitigated (9 crashes).

In order to estimate the percentage of truck crashes that would either be avoided or mitigated once an enhanced training and brake maintenance program was implemented, the project team first estimated the percentage of crashes that would have been avoided if the new program was fully effective in preventing or repairing brake malfunctions. (Bus crashes are discussed in the next section.) To calculate this number, the total number of crashes avoided, as shown in Table 5-2, was compared with the LTCCS 1,070 sample number to arrive at a percentage of all crashes that would have been avoided. The percentage of crashes that would be avoided if the brake problem could be eliminated would be 5.3 percent of all crashes. This percentage is calculated by dividing the Total Number of Crashes Avoided in Table 5-2 by the total sample number. That is, by dividing the 57 crashes avoided by the 1,070 crashes in the sample.

Second, to obtain a percentage of the total number of crashes that would be mitigated, the total number of crashes mitigated as shown in Table 5-2 was compared with the LTCCS 1,070 sample number to arrive at a percentage of all crashes that would have been mitigated. The number of crashes listed under the Total for Consequences Reduced in Table 5-2 is 49. When 49 is divided by the 1,070 LTCCS sample total, 4.6 percent of the total crashes are estimated to have their consequences mitigated. Thus, about 10 percent (9.9 percent) of all crashes could either be avoided or mitigated if the enhanced brake maintenance and training program was 100-percent effective. This percentage (10 percent) falls within the range of 1 to 40 percent estimate of crashes that could be prevented that the project team received as estimates from interviewees with brake expertise.

Finally, in order to estimate the percentage of truck crashes that either would be avoided or mitigated once an enhanced training, certification, and brake maintenance program was implemented, the Battelle Team had to estimate the effectiveness of the new program. The team assumed that the new program would unlikely be 100-percent effective. Some quality problems will always exist and shortfalls may continue to exist in training and effectiveness of maintenance and inspection programs. Although the team had no precise metrics to calculate new program effectiveness, based on their experience working with changes in vehicle maintenance and inspection programs, the team applied their best judgment to develop the assumption that the new program would experience program shortfalls in about 30 percent of the brake program activities. Thus, the new program would have an estimated effectiveness of 70 percent. Therefore, the enhanced NTSB recommended brake program (See Section 1.1) would result in either avoiding or mitigating about 7 percent (10 percent × 70 percent) of the total crashes. This number still falls within the range obtained in the survey results. Although the 70-percent program effectiveness estimate represents the project team's best assumption, the project team believes it appropriate to perform a sensitivity analysis for a range of effectiveness levels. Section 5.4, Benefit-cost Analysis, examines cost-benefit for effectiveness levels of 50, 60, 80, and 90 percent as well as at the 70-percent program effectiveness level for both crashes avoided and crashes mitigated.

Utilizing the results from the above analysis of LTCCS, about 10 percent of the 416,000 police-reported crashes found in the most recent Large Truck Crash Fact Book, 2002 data could reasonably have been caused by faulty brakes and could either be avoided or mitigated if the new brake training and maintenance program was 100 percent effective. However, when the assumed 70 percent program effectiveness is used, only 7.0 percent of all crashes would be affected. When the data are examined in more detail, 3.6 percent of all crashes would be avoided. The percentage of crashes avoided is estimated by assuming that 70 percent of the 57 shown in Table 5-2 for "Total Crashes Avoided" (about 40) would actually be avoided under the enhanced program. This percentage is about 3.6 percent of the total crashes in the LTCCS sample. To calculate the estimated number of crashes avoided, 5.3 percent is multiplied by the 416,000 police reported crashes in 2002. The resulting number, 22,161, is then multiplied by the assumed 70 percent program effectiveness rate to give 15,513 annual crashes. When rounded, this results in about 15,550 crashes avoided. A similar calculation to estimate the number of annual crashes mitigated shows that about 3 percent of the 416,000 crashes would be mitigated after the 70 percent efficiency rate is applied. This means that about 13,340 annual crashes would have their consequences reduced by half under the enhanced program.

In order to break down benefits between the two parts of the NTSB recommendations—driver training and pre-trip inspections and brake inspector training and certification—a small number of trucking industry representatives were contacted in order to allocate the benefits between the two recommendations.^5-1 Each of the trucking industry representatives contacted stated that the vast majority of the benefits would result from the brake inspector training certification. These views were echoed at a presentation of preliminary findings.^5-2 Reasons given for the limited usefulness of driver inspections included:

• High driver turnover rates,
• Lack of knowledge concerning vehicle maintenance techniques,
• Limitations in union contracts,
• Physical constraints,
• Aversion to the liability risk, and
• Lack of willingness to crawl under vehicle.

5.2.1.2 Avoidable Bus Crashes

Total bus crashes are listed in National Transportation Statistics report for 2003 (BTS 2004). In 2002, there were an estimated 58,000 bus crashes. The National Transportation Statistics figure also includes transit and school buses. However, identifying the number of intercity buses subject to FMCSA rules involves identifying the number of crashes involving buses not subject to FMCSA rules and subtracting them from the total number of crashes. Transit and school buses operate in a completely different environment from intercity motor coaches and many other passenger CMVs subject to FMCSA's jurisdiction. Several sources including the American Bus Association, National Safety Council, FMCSA databases, and FARS were investigated to determine the number of intercity bus crashes alone. The project team determined that data from the Transit Bus Integrated Vehicle-Based Safety Systems (IVBSS) project provides the best available data and would be used to estimate the annual number of intercity bus crashes (FTA 2004). This project began its process to estimate intercity and intracity bus crashes by first calculating the number of transit and school bus crashes and then subtracting this number from all bus crashes. When these data are used, the annual number of intercity bus crashes is estimated at about 7,800. The assumption used for this analysis is that the same percentage of intercity bus crashes as was the case with trucks could be caused by brake problems. When the same percentage figure of crashes that could have been caused by brakes and the same estimated 70 percent program efficiency level are applied to the 7,800 crashes, about 291 crashes could have been avoided by the implementation of the enhanced NTSB recommended brake program. Similarly, about 250 crashes annually could have been mitigated if the new training, certification, and inspection programs were in place. As for truck crashes discussed in Section 5.2.1.1, the assumptions for distributing benefits between the two NTSB recommendations were also applied here. A sensitivity analysis of a range of program effectiveness levels is applied to bus crashes in Section 5.4.

5.2.1.3 Avoided Costs of CMV Crashes

The CMV crash data presented in the preceding sections are multiplied by the societal costs of such crashes to determine program benefits or, more specifically, the costs that would be avoided by reducing the number of CMV crashes. The most recent and comprehensive estimates of the societal costs of truck and bus accidents are presented in the Pacific Institute's Revised Cost of Large Truck and Bus-Involved Crashes, prepared for FMCSA in 2002 (Zaloshnja, 2002). The societal costs as estimated by this report associated with heavy truck crashes and bus crashes are $59,153 and $32,548, respectively.^5-3 These costs, which are represented in 2000 dollars, were inflated to 2005 dollars-inflated to $66,813 for large trucks and $36,763 for buses-based on growth in the Consumer Price Index. These costs comprise medical-related costs, emergency services, property and equipment damage, lost productivity (e.g., wages, fringe benefits, claims processing costs, litigation costs, crash investigation costs, recruiting and training replacement for disabled workers), and monetized quality-adjusted life years.

The number of truck- and bus-involved crashes expected to be averted with implementation of the recommendations are expanded over the 10-year time horizon using a truck motor carrier growth rate based on the American Trucking Associations' U.S. Freight Transportation Forecast to 2016 (ATA, 2005). Furthermore, the estimates of total bus crashes are expanded based on motor carrier growth as estimated based on historic growth rates of private commercial buses from 1993-2003. This data is found in Federal Highway Statistics in Table MV-10 (FHWA, 2004).

To construct estimates of the benefits associated with crash avoidance the total number of estimated crashes avoided due to the implementation of the NTSB recommendations-16,941 for trucks and 308 for buses in 2005-were multiplied by estimated societal costs per crash ($66,813 for trucks and $36,763 for buses). The number of avoidable truck crashes was assumed to grow by 3 percent annually based on forecast growth in the trucking industry.^5-4 The number of bus crashes avoided and mitigated through the implementation of the NTSB recommendations was estimated to grow by 1.9 percent annually based on historic growth rates in the number of intercity buses.^5-5 The dollar values estimated for crash mitigation is the product of the estimated number of mitigated crashes (14,563 for trucks and 265 for buses) and the societal costs per crash divided in half. As noted previously, we assume that mitigated crashes would generate roughly half the costs relative to those crashes that would be avoided.

Based on the aforementioned assumptions, total crash cost savings are estimated at $18.7 billion over the 10-year analysis time horizon. Of this amount, $18.5 billion or 99 percent are attributed to the reductions in large truck crashes while $176.6 million are attributed to intercity bus crashes. Discounting these values at 7 percent results in an estimated $13.8 billion in avoided costs due to enhanced safety over the 10-year (2005-2014) time horizon. These benefits are shown in Table 5-3.

5.2.2 Out-of-Service Violation Impacts

For OOS truck violations, several surveys such as Operation Airbrake could have been used to provide the basis for estimating how many OOS violations could have been prevented by the adoption of the enhanced maintenance and inspection activities. The assumption made for this analysis is that the MCMIS data for OOS brake-related violations is most appropriate because it reflects the actual inspections that a carrier was subjected to by a "qualified technician" in a year. Consequently, any reduction in OOS violations derived from the implementation of the enhanced brake maintenance and inspection activities should be estimated using the MCMIS OOS data. OOS data for 2002, 2003, and 2004 from MCMIS are shown in Table 5-4.

MCMIS data for 2002 to 2004 show that for an average year, a total of 2,142,246 non-crash truck inspections were conducted. Of these inspections, brake violations represent an average of about 9.8 percent or 209,940 brake-related OOS violations per year.

If the same program efficiency of 70 percent is applied to a new brake maintenance and training program, 70 percent of the OOS violations could be prevented by the implementation of the enhanced brake maintenance and inspection activities. By adopting the new program, this calculation would result in a reduction of 146,958 brake related OOS violations annually.

For the years 2002, 2003, and 2004, MCMIS non-crash motor coach inspection data show that for an average year, 9,493 non-crash bus inspections were conducted. Of these, about 5 percent or 487 inspections resulted in an OOS brake violation in an average year. If the same 70 percent program efficiency level was applied to the bus brake-related OOS violations, 341 OOS violations would be avoided in a year by adopting the enhanced training, certification, and inspection program.

The costs associated with OOS violations are based on the assumptions that the average OOS violation results in 8.7 hours of downtime and that being placed out of service costs an average motor carrier $57 per hour. ^5-6 Based on these assumptions, total savings resulting from a reduction in OOS violations as a result of the brake inspection program totals $851.2 million over the 10-year time horizon or $629.3 million in discounted present value terms.

5.2.3 Impacts on Citations involving Monetary Penalties

Some cost savings can be expected due to improved maintenance and driver inspections (and correction) of stroke adjustment as a result of fines not being assessed for brake OOA violations. These would be in addition to the savings for vehicles not placed OOS due to brake adjustment.

While the amount of a fine for a "stroke beyond the adjustment limit" violation varies from place to place, a check of eight jurisdictions resulted in an average fine per violation (including court costs) of roughly $100.

The number of vehicles that would see cost savings from improved maintenance and pre-trip inspection due to elimination of OOA violations is calculated based on the assumptions outlined below. The data are taken from Operation Air Brake statistics from the years 1998 through 2004, looking only at the U.S. checks and separate calculations are given for trucks and buses:

• The total number of trucks inspected: 111,053
• The total number of truck brakes checked: 866,627
• Average number of truck brakes per vehicle = 866,627/111,053 = 7.80
• The total number of truck OOS for brake adjustment = 14,412

Thus, this is a minimum of 2 × 14,412 = 28,825 brakes OOA on trucks OOS for brakes OOA. This assumes that the CVSA 20-percent rule ^5-7 was applied, so at least 2 brakes per truck had to be OOA. Another method would take 20 percent of 7.80 brakes average per truck × 14,412 trucks = 22,494 brakes OOA on trucks OOS for OOA brakes.

Data compiled from Operation Air Brake statistics indicate that the total number of brakes OOA was 54,228. We assume that the total number of brakes OOA (54,228) includes those on trucks placed OOS (28,825 or 22,494, depending on which calculation you use). Therefore, the number of brakes OOA, but not resulting in OOS (i.e., those which would have been subject to a citation and fine) would be 54,228 less 28,825 (or 22,494) = 25,403 (or 31,734). We divide this number of brakes by the number of brakes per truck (in order to not be OOS, each truck would only be allowed to have only one brake OOA) and get 25,403/7.80 = 3,255 trucks (or 4,066 using the second method).

This results in 3,255/111,053 = 2.9 percent (or 3.7 percent) of the trucks inspected being subject to a fine for brakes OOA. Using this percentage of the 2.1 million inspections per year and $100 dollar average fine, and a 70 percent efficiency of the "maintenance + driver inspections," a savings of $4.4 million per year (or $5.5 million using the second calculation method) are estimated. We use the mid-point of these two estimates (3.3 percent) in the benefits calculation resulting in an estimated $4.9 million in savings in Year 1. Total savings over the 10-year time frame are estimated at $56.6 million (nominal 2005 dollars).

The bus analysis follows the truck analysis:

• The total number of buses inspected: 76
• The total number of bus brakes checked: 350
• Average number of bus brakes per vehicle = 350/76 = 4.61
• The total number of bus OOS for brake adjustment = 11

Thus, this is a minimum of 2 × 11 = 22 brakes OOA on buses OOS for brakes OOA. This assumes that the CVSA 20-percent rule was applied, so at least 2 brakes per bus had to be OOA. Another method would take 20 percent of 4.61 brakes average per bus × 11 buses = 10 brakes OOA on buses OOS for OOA brakes.

The total number of brakes OOA, as estimated based on data compiled from Operation Air Brake statistics, was 31. We assume that the total number of brakes OOA (31) includes those on buses placed OOS (22) (or 10, depending on which calculation you use). Therefore, the number of brakes OOA, but not resulting in OOS (i.e., those which would have been subject to a citation and fine) would be 31 less 22 (or 10) = 8 (or 20). These numbers do not sum to 31 due to rounding. We divide the number of brakes by the number of brakes per bus (in order to not be OOS, each bus would only be allowed to have only one brake OOA) and get 8/4.61 = 2 buses (or 4 using the second method).

This results in 2/76 = 2.3 percent (or 5.8 percent) of the buses inspected being subject to a fine for brakes OOA. Using this percentage of the 9,500 inspections per year and $100 dollar average fine, and a 70 percent efficiency of the "maintenance + driver inspections," we get a savings of $15,000 per year (or $38,500 using the second calculation method). We use the mid-point of these two estimates (4.1 percent) in the benefits calculation resulting in an estimated $27,000 in savings in Year 1. Total savings over the 10-year time frame are estimated at $310,000 (nominal 2005 dollars).

5.3 Costs Assessment

Implementing the NTSB recommendations would generate training, inspection, testing, and certification costs as well as administrative costs associated with record keeping and retention. Training would be required of bus and truck drivers responsible for performing pre-trip inspections. Brake inspectors would train in preparation for the testing and certification process called for in the NTSB recommendation. Training costs would include the costs associated with preparing training materials, compensating internal trainers, paying fees to external trainers, and labor costs for staff receiving the training. Training costs would include those required to train both new and existing drivers in initial and refresher training courses. The brake inspector certification process also would result in costs associated with registration and testing fees. Finally, pre-trip inspections would take drivers away from revenue-generating activities and would, thus, result in additional labor costs. In each case, these costs would be borne by the motor carrier industry. This section of the report examines these costs and examines the costs each would impose on industry.

5.3.1 Training, Testing, and Certification Costs

The recommendations set forth by NTSB would require a sufficient level of training in brake systems for both CMV drivers and brake inspectors. The results of the training, testing, and certification surveys suggest that the intensity of training necessary for drivers and brake inspectors would be different reflecting their particular responsibilities. Drivers would need to attain an acceptable level of knowledge to ensure for proper maintenance of CMV brake systems by conducting pre-trip inspections to identify the presence of brake deficiencies.

Initial training costs in Year 1 for driver training and for brake inspector training are made up of the costs to deliver the training, fees paid to third-party vendors, and labor replacement costs. The costs to deliver and receive the training are based on input provided from training, testing, and certification surveys regarding the duration of the training, pay for trainers, and class size. Labor data for truck and bus drivers, as well as brake inspectors, were obtained from the following sources: the National Compensation Survey (Bureau of Labor Statistics [BLS]), Occupational Employment Statistics (Bureau of Labor Statistics), American Trucking Associations Driver Compensation Survey, and Bureau of Transportation Statistics Office of Motor Carrier Information's Financial and Operating Statistics. The total number of truck drivers who would need to be trained initially was estimated by FMCSA at 7.0 million. The total number of brake inspectors used in the analysis is 456,666, and is also based on data provided by FMCSA.

Respondents indicated that an initial training course taking 4 hours would be sufficient to train drivers. Brake inspectors, however, would need a higher level of understanding to properly inspect, maintain, service, and repair CMV brake systems and to pass testing for certification as recommended by NTSB. The respondents indicated that brake inspectors would be required to undergo 8 hours of initial training and 2 hours of testing and certification.

Training costs over Years 2-10 of the 10-year analysis timeframe include the additional training needed to account for periodic refresher courses. To determine refresher training costs, assumptions were required regarding the wage of the participants, average class size, frequency, and duration of the refresher training course. Truck driver turnover rates were obtained from comments provided on an initial draft of this report at the TMC Fall Meeting in Valley Forge, PA. These comments identified turnover rates by industry segment (LTL, truckload, private). This information was combined with ATRI market share data to estimate driver turnover rates for the entire industry. Bus driver turnover rates were assumed to mirror those of truck drivers, and brake inspector turnover rates are based on BLS averages for the transportation sector and information acquired through stakeholder interviews. Interview respondents indicated that the NTSB recommendations would require truck drivers to undergo 2 hours of refresher training every year, while brake technicians would take 2 hours of refresher training every 2 years.

The research team assumed that driver training for companies employing 20 or more individuals would be conducted in-house. The in-house training costs are a function of training material costs ($4 per student), the hourly cost of the trainer ($34.70 per hour), training time (4 hours for initial training and 2 hours for refresher training), and average class size (17.5 students per class). The basis for each assumption is detailed in Table 5-5. Based on these assumptions, the average cost to supply the course in-house is estimated at $12 per student for initial training and $8 per student for refresher training. Because small companies lack the resources and capabilities to perform training in-house, companies with 19 or fewer employees are assumed to rely on external trainers. Based on the data collected in support of the brake technician course cost analysis presented in Table 5-6, fees paid to external trainers are estimated at $15 per student per hour. The 2002 Economic Census indicates that roughly 24.9 percent of all individuals employed in NAICS 484 Truck Transportation and 14.4 percent of all individuals employed in NAICS 4852 Interurban and Rural Bus Transportation and NAICS 4855 Charter Bus Industry are employed by companies with fewer than 20 employees (U.S. Department of Commerce 2002). Based on these assumptions, the numbers of truck and bus drivers seeking external training in the first year following implementation are estimated at 1.7 million and 3,503, respectively. External training costs are estimated at $15 per student per hour based on a scan of training schools presented in Table 5-6.

The more complex needs associated with training and certifying brake technicians is assumed to be performed completely by third-party vendors. The fees associated with external training courses for brake technicians were determined through a scan of comparable programs offered through training schools and academic institutions around the United States, as highlighted in Table 5-6. The costs associated with these training courses ranged from $45 to $250, with most in the $150-$250 range. However, these programs are generally more comprehensive and take much more time than the amount most motor carriers view as necessary to complete the training considered in this study. The list of comparable classes presented in Table 5-6 ranges from simple pre-trip inspection courses to more comprehensive brake technician courses. When normalized on a per-hour cost basis, the hourly training cost for these courses generally falls between $10 and $20. Based on these data, this analysis assumes an average training course cost of $15 per student per hour or $120 for initial brake technician training and $30 for all refresher training courses. This cost does not include the certification testing nor does it cover the labor costs associated with the time brake technicians spend in class.

Training cost estimates also account for payments to workers who replace drivers attending training. These labor replacement costs are a function of the hourly rates paid to replacement labor and the time drivers spend in class. The overtime provisions of the Fair Labor Standards Act do not apply to CMV drivers. Therefore, it is assumed that labor replacement costs would be paid at standard hourly rates and motor carriers would not incur overtime costs. Based on these assumptions, the labor replacement costs, which include fringe benefits, are estimated at $19.00, $19.25, and $22.69 for truck drivers, bus drivers, and brake technicians, respectively.

Based on these assumptions, total training, testing, and certification costs are estimated at $5.4 billion over the 10-year time horizon in discounted 2005 dollars. Of that amount, $4.8 billion are attributed to the training of truck drivers, $13.3 million to training bus drivers, and $564.9 million to the training, testing, and certification of brake inspectors.

In the short-term, the availability of external training options for drivers and brake inspectors could be limited. Based on the assumptions outlined previously in this section, the NTSB recommendations considered in this report would result in 1.7 million drivers and 456,666 brake inspectors attempting to locate external training options to satisfy new training requirements. Furthermore, motor carriers across the country would require time to develop training curricula and associated materials in order to train approximately 5.3 million drivers in the first year following implementation of the NTSB recommendations.

In 2004, there were roughly 2,000 automotive service technician training programs that had been certified by the National Institute for ASE. The NATEF has also certified more than 100 schools offering M/H technician training programs. There are 700 testing centers nationwide used by ASE in the certification process. These and other training facilities could be used to satisfy the requirements outlined in the NTSB recommendations; however, there are concerns that implementing the NTSB recommendations could be difficult in terms of both the ability of external trainers to meet the demand in the short-term and the geographic distribution of trainers and training facilities.

The economic analysis presented in this report does not make any assumptions regarding the availability of training seats, nor does it examine the impact of the enhanced demand for training caused by the NTSB recommendations on the prices associated with external training classes and certification testing. In order to address industry concerns regarding the challenges associated with meeting the requirements outlined in the NTSB recommendations, FMCSA could consider an extended implementation period (e.g., 2-4 years). Extending the implementation period would ease the burden on industry by allowing motor carriers more time to develop internal training programs and enabling the training industry to expand the supply of training seats and address the issues relating to the geographic dispersion of motor carrier operations.

5.3.2 Inspection Costs

The NTSB recommendations would require that CMV drivers inspect vehicle brakes before each trip. To estimate the cost of pre-trip inspections, the research team calculated the product of the total number of axles on single-unit and combination trucks operating on U.S. highways, the average number of trips taken by heavy trucks annually in the United States, per-axle inspection time requirements, and the average wage of the truck drivers performing the inspections. This section of the report describes the methods used to determine the costs associated with pre-trip inspections and estimates total pre-trip inspection costs associated with the NTSB recommendations.

Total Number of Axles on Single-Unit and Combination Trucks

Since the interview survey respondents estimated pre-trip inspection costs on a per-axle basis (2 minutes per axle), it became necessary to determine the number of axles on single-unit and combination trucks operating in the United States. The Federal Highway Administration's (FHWA's) Highway Statistics presents annual data on the number of single-unit and combination trucks operating in the United States in Table VM-1 (FHWA, 2004). These values were combined with data presented in the 2002 Vehicle Inventory and Use Survey (VIUS) in order to estimate the number of axles on trucks requiring pre-trip inspections (U.S. Census Bureau, 2004). The results of this analysis are presented in Table 5-7. Note that the NTSB recommendations would not apply to vehicles weighing 10,000 pounds or less. Therefore, these vehicles were excluded from the analysis. As shown, the number of axles on single-unit and combination trucks was estimated at 16.4 million in 2003.

Annual Number of Trips Taken by CMVs

The 2002 Vehicle Inventory and Use Survey (VIUS) published a comparative summary of operation ranges for U.S. trucks registered in the 1997 and 2002 VIUS (U.S. Census Bureau, 2004). These estimates represent average miles per trip, as shown in Table 5-8. VIUS data indicate that roughly 50 percent of all trucks operate in an area of less than 50 miles, the average trip length for 16.8 percent of all trucks falls between 51 to 200 miles, and only about 9.5 percent of all trucks' average trip length exceeds 200 miles.

The drawback of using these percentages for the CMV brake inspection and maintenance study is that they do not represent trucking industry averages. Rather, VIUS statistics represent aggregate statistics of all types of trucks. This study, on the other hand, focuses on trucking and intercity bus industry motor carriers that are largely under-represented in VIUS data.

Forkenbrock, in a study titled "External Costs of Intercity Truck Transportation," used an ATA database principally comprised of Class I (annual revenues in excess of $10 million) carrier data to estimate the proportion of for-hire truckload general freight trucking operations falling into three ranges of operation categories: less than 250 miles, 250 to 500 miles, and over 500 miles (Forkenbrock, 1999). These data are shown in Table 5-9. The data set used in Forkenbrock (on pages 505 to 526), had a small number of Class II carriers (annual revenues between $3 and $10 million) and almost no Class III carriers (annual revenues of less than $3 million). Due to the bias inherent in the data used in the study, the percentage of trucks traveling in excess of 500 miles per trip would not appear to apply to the overall trucking industry.

Neither of the two studies outlined previously are directly applicable to this study; however, they could represent upper and lower bounds on most likely estimates. Thus, the research team initially adopted a mid-range estimate of 300 to 400 miles per trip on average for the CMVs studied here. However, when considering that long-haul trucking operations would be required to inspect the brake systems each day prior to departure, the most miles a truck could travel between inspections would be roughly 660 (60 miles per hour multiplied by 11 hours on the road). Thus, the low-end estimate was reduced to 200 miles. When combined with annual mileage estimates on a per-truck and bus basis appearing in Highway Statistics, this range produces an estimate of 68 to 136 inspections per year per truck and 21 to 43 inspections per year per bus.

Inspection Costs

The annual number of axles inspected is computed as the product of the number of annual inspections (68-136 as outlined in the previous section) and the total number of axles on single-unit and combination trucks (16.4 million as documented in Table 5-7). Based on these assumptions, the total number of truck axles requiring inspection was estimated at between 1.1 and 2.2 billion in 2005. Based on truck driver salaries previously noted in the training cost analysis (Table 5-5), total estimated costs associated with performing 2-minute pre-trip inspections on each axle would result in approximately $4.9 to $9.9 billion in inspection costs over the 10-year study time horizon.

The total number of bus axles inspected annually is a function of the number of inspections outlined in the previous section of this report (21-43 annually) and the total number of bus axles (estimated at 279,790). The cost to inspect each axle is a function of the time required to inspect it (2 minutes or 1/30th of an hour) and the hourly wage plus fringe benefits of bus drivers ($19.25 as documented in Table 5-5). Based on these assumptions, total inspection costs are estimated at between $25.2 and $50.3 million over the 10-year study time horizon.

Administrative Costs

There are administrative costs borne by the motor carrier industry associated with coordinating training for drivers and brake inspectors and maintaining training and certification records. To determine the costs associated with these activities, there were several questions posed to industry regarding the costs associated with these administrative functions. Unfortunately, the interview surveys failed to produce conclusive results in this area. Thus, based on limited data and additional analysis of the administrative requirements under the NTSB recommendations, the research team estimates administrative costs at 10 minutes per driver and brake inspector trainee. The annual administrative costs are, therefore, estimated as the product of the number of annual driver and brake inspector trainees and 10 minutes salary for executive secretaries and administrative assistants, estimated in the BLS OES at $22.64, including fringe benefits. Based on these assumptions, total administrative costs are estimated at $316.7 million over the 10-year time horizon (2005 to 2014). Of this amount, $303.7 million was attributed to administering the driver training programs, with the remainder ($13.0 million) tied to brake inspector training and certification.

5.4 Benefit-cost Analysis

The incremental costs incurred by the motor carrier industry to implement the NTSB recommendations in the base case analysis (70 percent program effectiveness) are summarized in Tables 5-10 and 5-11. Benefits associated with improved safety and reduced OOS violations are also included in these tables. The most significant costs are those incurred during pre-trip inspections, which total roughly $4.9 to $9.9 billion over the 10-year time horizon (2005 to 2014). Though each inspection would require only 2 minutes per axle, the costs of this requirement are significant because they would require billions of inspections to be performed over the next 10 years. The cost range is due to varying the assumption regarding the average distance per trip between 200 and 400 miles. Training, testing, and certification collectively represent the next highest cost item, totaling $5.4 billion over the 10-year analysis timeframe. Administrative costs are relatively low at $316.7 million over 10 years.

Overall, estimated benefits are higher than projected costs over the 10-year time horizon in the base case analysis when low-end cost assumptions are used, with discounted benefits (using a 7 percent discount rate) totaling $14.5 billion ($19.6 billion in nominal benefits), compared to $10.7 billion in costs. When the high-end cost assumption regarding the number of inspections is used, estimated costs grow to $15.6 billion and exceed estimated benefits by $1.1 billion. Thus, net benefits are estimated at ($1.1) to $3.8 billion. These numbers correspond with benefit-cost ratios of .93 and 1.36. These results demonstrated nearly no sensitivity to variations in discount rates. At a 10 percent discount rate, the benefit-cost ratio ranges between .92 and 1.35 and varies from .93 and 1.07 when a 4 percent discount rate is applied. The lack of discount rate sensitivity occurs because benefits and costs are realized relatively evenly over the 10-year time horizon, with the exception of relatively higher training costs in Year 1.

Table 5-12 presents the findings of the benefit-cost analysis when the ratio of program effectiveness is varied between 50 and 90 percent and the benefits and costs are split between the two NTSB recommendations (driver inspection and training and brake inspector training and certification) As discussed in Section 5.2, roughly 70 to 90 percent of the benefits associated with reducing the number of crashes, fines and violations are attributed to the proposed brake inspector training and certification program. The majority of the costs, however, were attributed to the driver training and pre-trip inspection recommendation.

The final step in conducting the benefit-cost analysis (BCA) is to compare the societal costs and benefits and to construct BCRs in order to quantify the extent to which benefits exceed or fall short of costs. A BCR is equal to the present value of benefits divided by the present value of costs. Thus, a BCR in excess of 1.0 demonstrates positive economic returns to society. When BCRs exceed 1.0, society experiences net benefits from the regulation (net present value of benefits = present value of benefits - present value of costs). Under the high-end estimate concerning the costs associated with pre-trip inspections (200-mile average trip), the NTSB recommendations produce net benefits with BCRs in excess of 1.0. Under the low-end estimate (400-mile average trip), the recommendations fail to produce net benefits when the program effectiveness drops below 76 percent (i.e., the BCR is below 1.0).

The driver training and pre-trip inspection recommendation fails to produce net benefits under any of the scenarios considered in Table 5-12. The driver training and pre-trip inspection recommendation produced BCRs ranging from 0.07 (50 percent ratio of effectiveness, 200-mile average trip) to 0.55 (80 percent program effectiveness, 400-mile average trip).

The brake inspector and certification recommendation, on the other hand, produced large net benefits, with BCRs exceeding 12.65 in all scenarios considered in this study. The BCRs range from a low of 12.65 (50 percent program effectiveness, 200-mile average trip) to 29.27 (90 percent program effectiveness, 400-mile average trip).