Center High Mounted Stop Lamps (CHMSL) have been standard equipment on all new passenger cars manufactured on and after September 1, 1985 for sale in the United States. They are required by an October 1983 amendment [11] of Federal Motor Vehicle Safety Standard 108 [5]. CHMSL have also been standard equipment on all new light trucks (pickup trucks, vans and sport utility vehicles)manufactured on and after September 1, 1993 for sale in the United States, following an April 1991 amendment [12] of FMVSS 108. CHMSL are red stop lamps mounted on the center line of the rear of a vehicle, generally higher than the stop lamps on the sides of that vehicle. They are activated when the driver steps on the brake pedal and they are off at other times. The purpose of CHMSL is preventing crashes by reducing the reaction time for drivers to notice that the vehicle in front of them is braking.

There are several hypotheses why CHMSL might stimulate a quicker reaction than conventional stop lamps. The central and raised location of CHMSL puts them "in an area of the forward visual field toward which a following driver most often glances [6]." Since its central location makes "the CHMSL separate and distinct from all other rear lamps and signals, any possible ambiguity of the signal is reduced," especially, the "likelihood that the signal will be interpreted as a directional signal [6]" (turn signal or tail lamp). The CHMSL, in combination with the two lower side mounted lamps, forms a triangle which could be an additional cue to get the driver's attention. The high mounting of the lamp might make it visible through the windows of a following vehicle and enable the driver of the third vehicle in a chain to react to the first car' s braking. Some drivers may interpret the high mounted lamp as a warning to keep their distance;

by following at a safer distance, they have more room to stop.

1.1 Evaluation of CHMSL

The Government Performance and Results Act of 1993 [16] and Executive Order 12866 (October 1993) [13] require agencies to evaluate their existing programs and regulations. The objectives of an evaluation are to determine the actual benefits - lives saved, injuries prevented, damages avoided - and costs of safety equipment installed in production vehicles in connection with a rule. This report tracks the effectiveness, benefits and costs of passenger-car CHMSL during 1986-95. At the beginning of that 10-year period, only a small number of cars with CHMSL were on the road, but by 1995, the majority of cars on the road were CHMSL-equipped. The report completes the evaluation of passenger car CHMSL, following up on NHTSA's preliminary evaluation based on CHMSL performance in mid 1986 [18] and interim evaluation based on 1987 performance [19]. In addition, this report presents early effectiveness results for light truck CHMSL, based on statistical analyses of crash data.

The CHMSL rule has evolved through the full cycle of experimental research, test fleets, regulatory analysis, rulemaking and evaluation. During 1974-79, experimental research with CHMSL-equipped passenger cars showed significant reductions in reaction time relative to conventional stop lamps [15], pp. III-19 - III-23, [6]. In 1976-79, NHTSA sponsored installation of CHMSL on test fleets comprising over 3000 cars. The CHMSL equipped cars had significantly fewer rear impacts than control groups with conventional lamps. The Regulatory Impact Analysis, published in 1983, included detailed projections of the crashes, injuries and damages that might be avoided with CHMSL, as well as a cost estimate [15]. It concluded that CHMSL would almost certainly be cost effective. When the CHMSL rule was promulgated in 1983 (with an effective date of September 1, 1985), a comprehensive evaluation plan [7] was published at the same time, outlining statistical and engineering analyses to determine the actual effectiveness and cost of production CHMSL. The earlier evaluations [18], [19] as well as this report follow the guidelines of the evaluation plan.

CHMSL retrofit kits are relatively easy to manufacture and install. Favorable public opinion and support by motorist groups such as the American Automobile Association helped create a substantial market for retrofit CHMSL. Inquiries to lamp manufacturers and the AAA revealed that approximately 4,000,000 CHMSL retrofit kits had been manufactured or imported into the United States by mid 1986 and most of them were installed on model year 1980-85 cars [18], p. 5. More cars have been retrofitted since then. It is likely that 10 percent of model year 1980-85 cars (or at least 1984-85 cars) had been retrofitted by mid 1987. As will be shown in Section 2.4, the percent of retrofits is one of the variables in the formula for estimating CHMSL effectiveness from crash data.

1.2 Results of earlier effectiveness and cost studies - passenger cars

The initial effectiveness studies were based on the 1976-79 test fleets [15],III-8 - III-18. The first test fleet consisted of Washington taxicabs [27]. Some taxis were equipped with CHMSL or other distinctive stop lamps, while a control group of the same makes, models and driver characteristics had conventional stop lamps. Drivers reported all crash involvements. The most important finding of the field test with Washington taxicabs was that the CHMSL equipped cars had 36 percent fewer rear impacts per million miles than the control group.

NHTSA validated the first study with a larger test fleet of telephone company cars (2,500 with CHMSL) in 4 regions of the United States [30]. The results were nearly identical: the CHMSL equipped cars had 35 percent fewer rear impacts per million miles than the control group.

The Insurance Institute for Highway Safety sponsored a field test with New York City taxicabs and obtained an average 34 percent reduction of rear-impact crashes [29]. (Actually, the three preceding studies measured effectiveness as a reduction of "CHMSL relevant" crashes, where the driver was braking before being struck. Since the data in those studies indicated that two-thirds of all rear impacts are "CHMSL relevant," and since no effect can be expected in crashes where the driver was not braking, the percentage reduction of all rear impacts is two-thirds the percentage reduction of "CHMSL relevant" rear impacts.)

NHTSA's Final Regulatory Impact Analysis (FRIA), dated October 1983 [15] based its effectiveness estimate on the field tests and its cost estimate on analyses of prototypes. Benefits (crashes, injuries and damages avoided) were projected from conservative assumptions about the types of crashes in which CHMSL would be effective. The main predictions of the FRIA were:

Overall effectiveness    33 percent reduction of rear-impact crashes (field test results rounded down)

Cost per car      $4.13-6.76 in 1982 dollars, which is equivalent to $6.34-10.37 in 1994 dollars

Damage reduction   $434 million per year (conservative estimate of $282 average damage per crash involved vehicle; conservative assumption of the number of rear impact crashes; no effectiveness assumed in rural crashes)

Injury reduction       40,000 per year (33 percent effectiveness was not assumed to apply to injury crashes; instead, NHTSA postulated a speed distribution for injury crashes and projected how improved reaction times with CHMSL would change this distribution)

NHTSA's preliminary evaluation of CHMSL [18] was based on police-reported crashes that occurred at the 50 National Accident Sampling System (NASS) areas during June-August 1986, a nationally representative data set. The involvement rate in rear impacts for model year 1986 cars (all CHMSL equipped) is compared to 1985 cars (mostly without the lamps). The sample included 1571 CHMSL equipped cars with rear impact damage (15 times larger than the sample in the field tests, but 1/50 as many cases as NHTSA's second evaluation report). The principal findings were:

Overall effectiveness 15 percent reduction of rear-impact crashes

Confidence bounds Not explicitly stated, but appear to be at least as wide as 7-22 percent if the data are treated as a simple random sample and perhaps wider, given that the data derived from a cluster sample

In specific situations CHMSL likely to be more effective in crashes involving 3 or more vehicles than in 2 vehicle crashes; no big differences in the effect of CHMSL between rural and urban areas

NHTSA's second evaluation report was based on a full calendar year 1987 of crash data from 11 States [19]. There were enough data to estimate the effectiveness of CHMSL with reasonably narrow confidence bounds. The statistical procedure was to compare the proportion of crashes that are rear impacts in model year 1986-87 cars (with CHMSL) to model year 1980-85 cars (without CHMSL), after adjusting those proportions for vehicle age effects. The cost of CHMSL was estimated by a detailed inspection of a sample of lamps from production vehicles, and a cost-effectiveness analysis was performed. The principal findings were:

Overall effectiveness 11.3 percent reduction of rear-impact crashes

Confidence bounds 8.8 to 13.8 percent, based on the State-to-State variation of the effectiveness estimate

In specific situations CHMSL somewhat more effective in 3+ vehicle crashes, in daylight crashes, rural roads, nonsignalized locations; equally or more effective in injury crashes than in property-damage crashes, but no effect on fatalities

Cost per car $10.48 in 1987 dollars, which is equivalent to $13.60 in 1994 dollars

Damage reduction $910 million per year (when all cars, no light trucks have CHMSL)

Injury reduction 79,000 - 101,000 per year

The unmistakable trend in these results is that CHMSL effectiveness declined: from 34-36 percent in the field tests, to 15 percent in the first evaluation, to 11.3 percent in the second. To some extent, perhaps, the conclusion may be hedged because the various studies derived from different types of data, but such differences alone could hardly explain the downward trend, especially when the second evaluation showed fairly similar levels of CHMSL effectiveness over a wide variety of crash types and severities. The results raised obvious questions. As more and more cars on the road have CHMSL, do drivers "acclimatize" to the lamps and pay somewhat less attention to them [17]? Would effectiveness continue to decline, and if so, what would be its "long-term" level? Needless to say, the results of NHTSA's second evaluation report did not yet constitute a "final" assessment of CHMSL effectiveness. As recommended in that report and in NHTSA's 1983 Evaluation Plan [7], p. 3, follow-up analyses needed to be performed.

At least two evaluations of the effectiveness of CHMSL have been performed by agencies outside the United States government, based on North American crash data. Transport Canada analyzed the effect of CHMSL in calendar year 1987, based on crash data from seven Provinces [28]. The time frame (CY 1987) and analysis methods were quite similar to NHTSA's second evaluation. So was the result: a 10.5 percent reduction of rear-impact crashes attributed to CHMSL.

However, the first comprehensive analysis to include data collected after 1987 was performed by Farmer at the Insurance Institute for Highway Safety, covering the 1986-91 time frame [9]. The data base consisted of over 400,000 property-damage liability claims filed with insurance companies by owners of model year 1984-87 passenger cars. The statistical procedure was to compare the proportion of claims that are rear impacts in model year 1986 cars (with CHMSL) to model year 1985 cars (without CHMSL), after adjusting those proportions for vehicle age effects. Over the full 1986-91 time frame, CHMSL was associated with a 5.1 percent reduction of rear-impact crashes. That's well below the 11.3 percent reduction found in NHTSA's second evaluation for 1987. Farmer did not have enough data to obtain accurate year-by-year estimates or to draw definitive conclusions about whether effectiveness was changing over time and whether or not it had reached its long-term level. Nevertheless, this authoritative study served notice that the long-term effect of CHMSL would be well below the levels seen in NHTSA's two evaluations, and quite possibly below 5 percent.

1.3 Extension of CHMSL to light trucks

Many of the safety standards that originally applied only to passenger cars were subsequently extended to light trucks (pickup trucks, vans and sport utility vehicles). The process accelerated after 1980, as light trucks became increasingly popular vehicles for personal transportation. To the extent that light trucks have the same types of crashes as cars, and similar driving exposure, it is often plausible to argue, without extensive additional research, that safety measures effective in cars are also likely to be effective in light trucks. In some cases, however, issues arise that complicate the feasibility of a standard in light trucks, or raise doubts about its potential benefits or costs.

Following the successful debut of CHMSL on passenger cars in model years 1986-87, it became natural to consider extending the requirement to light trucks. The principal issue that complicated the extension to light trucks was the location of the CHMSL. It was not a problem for the smaller vans and sport utility vehicles, whose backsides are similar to station wagons. But on pickup trucks, the question arose whether to locate the CHMSL near the cab roof, where it might be less than perfectly conspicuous to the following driver (too high and too far forward, especially relative to the other rear lights), or on the tailgate, where there might be problems of durability and cost, and where the effect would be lost when the tailgate is open. Also, on certain large vans, there was a question if the CHMSL would be too high up to be fully conspicuous.

In 1988, NHTSA conducted extensive tests of the reaction times of volunteers to simulated light trucks with CHMSL or with conventional brake lights. The reaction time for drivers following a truck with CHMSL was 0.09 seconds shorter than for drivers following a truck without CHMSL. That is just a slightly lower benefit than in passenger cars, where the reduction in reaction time with CHMSL was 0.11 seconds [14], p. 2.

On the basis of this research and other information, the Final Regulatory Impact Analysis (FRIA) for light truck CHMSL asserted that the problems concerning location could be resolved and that the lamps would be effective for trucks, but granted that it might be slightly less effective than in cars (i.e., 9/11 as effective) [14], p. 21. The FRIA also asserted that in most light trucks the cost of CHMSL would be about the same as in passenger cars, but granted that costs might increase by about 50 percent in trucks that are produced in multiple stages - i.e., where the truck is modified prior to sale, obscuring the original CHMSL and requiring it to be moved to another location [14], p. 30.

The April 1991 Final Rule allowed 2 years lead time, till September 1993, for installation of CHMSL. In fact, CHMSL were phased in over a three-year period, MY 1991-94, since some trucks already had them before the Rule was published, while others did not get them until the effective date. That contrasts with the nearly simultaneous implementation of CHMSL in passenger cars (MY 1986 in all models except Cadillacs and very few others). Dodge Caravan, Plymouth Voyager, Chrysler Town & Country and Ford Explorer were the first to get CHMSL, in MY 1991. Ford's full-sized pickup trucks, vans and sport utility vehicles got them in 1992, and all other make-models got them in 1993 or 1994.