|This report is an archived publication and may contain dated technical, contact, and link information|
Publication Number: FHWA-HRT-04-142
Date: December 2005
Enhanced Night Visibility Series, Volume XI: Phase II—Cost-Benefit Analysis
PDF Version (443 KB)
PDF files can be viewed with the Acrobat® Reader®
CHAPTER 5—FINDINGS AND CONCLUSIONS
The procedures described above, applied to the data described above, generate an estimate of the cost of each VES and pavement marking system. They also generate an estimate of total annual crash costs and the percentage of crash costs most likely to benefit from enhanced night visibility.
The cost of each innovative technology, and of the possible combinations of technology, is estimated in two different ways. The first set of results, displayed in table 14 through table 16, forecasts the incremental annual costs of each technology in the steady state when 100 percent implementation is reached; the forecasts are without regard to the costs that accrued in getting to that point. Here the incremental costs refer to the excess above and beyond the costs of the benchmark HLB technology with nonfluorescent pavement markings. The second set of results, displayed in table 20 through table 22, forecasts the costs, discounted to present value, over the course of an assumed 20 year implementation.
Table 14 gives an estimation of the initial incremental cost of each VES per vehicle; table 15 does the same for each pavement marking system. The tables also show what the total annualized incremental cost of each system would be in the year 2020, if 100 percent implementation were achieved by that time. These costs are given in year 2003 U.S. dollars. They are incremental costs, that is, only those costs above and beyond the cost of HLB headlamps or nonfluorescent paint, which are used as benchmarks. For this reason the incremental cost of HLB, and of any system that costs as much as HLB does, is zero by definition; the same applies for nonfluorescent paint.
Table 16 shows the annual incremental cost of each combination of one VES with one pavement marking system. In other words, the dollar value entered in the “five UV–A + HLB” row and the “fluorescent thermoplastic” column is the sum of the total annualized incremental cost of the five UV–A + HLB system plus the total annualized incremental cost of the fluorescent thermoplastic system.
Break-Even Crash Reduction at 100 Percent Implementation
Table 17 shows the estimated annual cost of crashes in each of five critical event categories (compare figure 7 and figure 13). This presentation is intended to permit comparison with the similar tabulation used by Nitzburg et al.(1) The chief differences are that the table represents average crash costs in 1999-2001 rather than in the 1980s, and the crash costs are categorized by critical event and light condition rather than by crash geometry.
Table 18 shows what percentage reduction in the costs of total unlighted nighttime crashes each VES/pavement marking combination would need to achieve to create annual benefits (i.e., cost savings) that match its estimated annual incremental cost.
Table 19 shows what percentage reduction in the costs of total unlighted nighttime, dawn, and dusk crashes each VES/pavement marking combination would need to create annual benefits (i.e., cost savings) that match its estimated annual incremental cost.
Because the only combination of pavement marking material and VES that shows systematic improvements in sight distance for drivers of different age groups (see ENV Volumes III through VI) is the five UV–A lamps with halogen low-beam lamps plus the nonfluorescent pavement markings, only these systems are likely to yield positive crash cost reduction. When the sight distance findings are broken down by the type of object to be detected and related to the corresponding critical event category, three systems may be expected to create pedestrian crash cost savings: three UV–A + HLB, five UV–A + HLB, and IR–TIS; these three systems may also be expected to create cyclist crash cost savings. Lane departure crash cost savings may be expected for the five UV–A + HLB, hybrid UV–A + HLB, three UV–A + HLB, and IR–TIS.
It should be evident that if the ENV technologies affect night, dusk, and dawn crashes rather than night crashes alone, then the potential crash cost savings of each combination would be about 25 percent larger (see figure 7 and compare table 18 and table 19). The relative rankings of the systems remain unchanged.
The results in the tables that follow show costs discounted to the beginning of the first year of implementation at an interest rate of 4 percent per annum. Implementation is assumed to occur at the rate of 5 percent per year (i.e., an additional 5 percent of vehicles and 5 percent of highway miles are equipped each year) until full implementation is achieved at the end of 20 years.
Table 20 shows the incremental present discounted cost of each VES with conventional pavement markings. The table covers a 20-year period on unlighted highways when the system is introduced to the vehicle fleet, and then, in steps of 5 percent per year until 100 percent implementation is achieved; table 21 does the same for pavement marking systems. These costs are given in year 2003 U.S. dollars. They are incremental costs (i.e., only those costs above and beyond the cost of HLB headlamps or nonfluorescent paint) that are used as benchmarks. For this reason, the incremental cost of HLB (and any system that costs as much as HLB does) is zero by definition; the same goes for nonfluorescent paint.
Table 22 shows the present discounted cost of each combination of one VES with one pavement marking system. In other words, the dollar value entered in the “five UV–A + HLB” row and the “Fluorescent Thermoplastic” column is the sum of the incremental present discounted cost of the five UV–A + HLB system plus the incremental present discounted cost of the fluorescent thermoplastic system. Again, the assumption that the systems would penetrate the vehicle fleet and the unlighted highways in steps of 5 percent per year underlies the computations.
The percentage of effect on crash costs necessary to break even would tend to be slightly larger in the present discounted-value computation than in the steady-state computation for those technology combinations that include both UV–A headlamps and fluorescent pavement markings. (Put differently, any benefit-cost ratios that one might calculate would tend to be slightly smaller.) This slight difference results from the effect that the UV–A headlamps and the fluorescent pavement marking systems create when used in combination. At a constant implementation rate of 5 percent per year, the cost of these systems in combination grows at 5 percent per year also, while their positive effect on crashes (if any) grows very slowly at first.
The cost-benefit analysis in this report, adhering closely to the cost-benefit framework in the FHWA report A Safety Evaluation of UVA Vehicle Headlights,(1) permits a relatively straightforward comparison of the cost and benefit estimates produced for this study with the earlier estimates that Nitzburg et al. produced in their steady-state analysis. The FHWA report, based on engineering estimates and a very limited body of relevant literature, unavoidably lacked precision, and it is instructive to see how far the Smart Road tests and the reported equipment costs corroborate its estimates.
The reports differ on a couple of methodological points. First, Nitzburg et al. used the GES to tabulate estimated crash costs from a hybrid CDS/NASS file that they created to correct some shortcomings in the personal injury data that NASS provided.(1) The current report uses GES to estimate crash costs from a set of NASS files. Second, Nitzburg et al. tabulated the crash cost estimates in six categories defined by crash geometry.(1) The current report tabulates the crash cost estimates in five categories defined by critical event. The category definitions may not be important, but they may lead to different judgments about which nighttime crashes appear to be relevant, that is, have a potential for reduction.
A glance at table 17 shows that the current study’s estimate of total crash costs in unlighted conditions (dark, dawn, and dusk), $58.75 billion at 2003 prices, is reasonably close to the Nitzburg et al. estimate of $53.2 billion at 1995 prices.(1) Table 16, on the other hand, shows that the current study’s estimate of the costs of the ultraviolet and fluorescent technologies, $111 to $116 billion for the UV–A headlights and some $16 billion for the fluorescent markings, is two orders of magnitude greater than the Nitzburg et al. estimate of $1.3 billion for the UV–A headlights and $0.23 billion for the fluorescent markings.(1) The ENV study’s estimates of the cost of HID headlamps and IR imaging systems have no counterpart in Nitzburg’s FHWA report.
Traffic counts that indicate what fraction of vehicle miles traveled take place in clear, rainy, snowy, and foggy atmospheric conditions would permit a variant approach to the benefit calculation.
The analysis in this study postulates that the motoring public would realize the benefits of enhanced night visibility in the form of reduced crash costs. It is conceivable that some motorists would attempt to convert crash-cost savings into time-cost savings by driving faster. Any estimate of the cost savings based on constant traffic volume and speed must be considered a lower bound on the true benefits that might occur if motorists could capture additional net savings by trading safety for time.
Under the extreme assumption that motorists benefiting from one of the new night visibility technologies would choose to speed up so much that the risk of a crash remained exactly the same as before, the benefits of the new technology would accrue entirely in the form of travel-time savings. Estimating time savings would require an estimate of the vehicle miles traveled in each of the combinations of light conditions and weather (and, possibly, driver age and gender) by which the crash database can be categorized. Traffic counts that break down traffic volume on a road by the light conditions, weather, and driver age would minimize the number of assumptions and the margin of error in such a calculation.
Some information about the cost of fluorescent delineator posts was collected while completing the cost-benefit analysis; however, the effect of fluorescent materials on the distance at which a delineator post might be detected by a driver was not measured. Therefore, the cost-benefit analysis does not include an assessment of the potential effect of fluorescent delineator posts on future crash costs.
In principle, if delineator post detection distances were obtained from a future study and if information on the distribution of delineator posts on the Nation’s highways were collected, it would be possible to include the effect of fluorescent delineator posts in a study such as this one.
Topics: research, safety
Keywords: research, safety, Crash, Automobile, Benefit, Cost, Cost-Benefit Model, Detection, Fluorescent, Halogen, Headlamp, Night Vision, Nighttime, Road Marking, Ultraviolet, Visibility, Vision Enhancement System
TRT Terms: research, Safety and security, Safety, Transportation safety, Automobiles--Lighting--Cost effectiveness, Road markings--Cost effectiveness, Automobile driving at night, Night visibility, Headlamps, Benefit cost analysis