Mandatory Passive Restraint Systems inAutomobiles: Issues and Evidence

September 1982

NTIS order #PB83-159822

TECHNOLOGYAND

HANDICAPPED PEOPLESEPTEMBER 1982

BACKGROUND PAPER #l: MANDATORY PASSIVERESTRAINT SYSTEMS IN AUTOMOBILES:

ISSUES AND EVIDENCE

Kenneth E. Warner, Ph. D.

Department of Health Planning and Administration

School of Public Health

University of Michigan, Ann Arbor, Mich.

OTA Background Papers are documents containing information that supplementsformal OTA assessments or is an outcome of internal exploratory planning and evalua-tion. The material is usually not of immediate policy interest such as is containedin an OTA Report or Technical Memorandum, nor does it present options for Con-gress to consider.

-.CONGRESS OF THE UNITED STATES

., ~ Office of Technology Assessment‘ Washington, D C 20510

—

Library of Congress Catalog Card Number 82-600546

For sale by the Superintendent of Documents,U.S. Government Printing Office, Washington, D.C. 20402

Foreword

Technology exerts a powerful influence over the lives of everyone, making life easier,more fulfilling, but sometimes more painful and frustrating. This is especially true forpeople with disabilities. The appropriate application of technologies to diminishing thelimitations and extending the capabilities of disabled and handicapped persons is oneof the prime social and economic goals of public policy.

The Federal Government is deeply involved in programs that affect the develop-ment and use of technologies for disabilities. Congress and other institutions have beenincreasingly interested in how well programs that directly develop technologies and sup-port their use have been performing. The Senate Committee on Labor and Human Re-sources requested the Office of Technology Assessment (OTA) to conduct a study oftechnologies for handicapped individuals. That study examined specific factors that af-fect the research and development, evaluation, diffusion and marketing, delivery, use,and financing of technologies directly related to disabled people. The problems and proc-esses of the development and use of technologies were analyzed in the context of societalallocation of resources and the setting of goals for public policy. The main report ofthe study Technology and Handicapped People was released in May 1982.

This case is background paper #1 of the study. A number of case studies will bepublished as part of the assessment, and each will be issued separately. The case studieswere commissioned by OTA both to provide information on specific technologies andto gain lessons that could be applied to the broader policy aspects of technology anddisability.

Drafts of each case study were reviewed by OTA staff; by members of the advisorypanel to the overall assessment, chaired by Dr. Daisy Tagliacozzo; by members of theHealth Program Advisory Committee, chaired by Dr. Frederick Robbins; and by nu-merous other experts in medicine, disability policy, Government, economics, publicinterest and consumer rights, and rehabilitation engineering. We are grateful for theirassistance. However, responsibility for the case studies remains with the authors.

#..

JOHN H. GIBBONS

Iii

Advisory Panel on Technology and Handicapped People

Daisy Tagliacozzo, Panel ChairDepartment of Sociology, University of Massachusetts, Harbor Campus

Miriam K. BazelonWashington, D.C.

Tom BeauchampKennedy Institute-Center for BioethicsGeorgetown University

Monroe BerkowitzBureau of Economic ResearchRutgers University

Henrik BlumUniversity of California, Berkeley

Frank BoweWoodmere, N. Y.

Jim GallagherMartha Porter Graham CenterUniversity of North Carolina, Chapel Hill

Melvin GlasserCommittee for National Health Insurance

Ralf HotchkissOakland Calif.

John KimberlyThe Wharton SchoolUniversity of Pennsylvania

Robert LeopoldDepartment of PsychiatryHospital of the University of Pennsylvania

LeRoy LevittMount Sinai Hospital

A. Malachi Mixon, IIIInvacare Corp.

Jacquelin PerryRancho Los Amigos Hospital

Barbara W. SklarMount Zion Hospital

William StasonVeterans Administration and Harvard School of

Public Health

Gregg VanderheidenTrace Research and Development CenterUniversity of Wisconsin

Michael ZulloCorporate Partnership ProgramU.S. Council for International Year of

Disabled Persons

OTA Staff for Background Paper #1

H. David Banta, Assistant Director, OTAHealth and Life Sciences Division

Clyde J. Behney, Health Program Manager

Clyde J. Behney, Project Director

Anne Kesselman Burns, Analyst

Chester Strobel, Analyst

Kerry Britten Kemp, Editor

Virginia Cwalina, Administrative Assistant

Mary E. Harvey, Secretary

Pamela J. Simerly, Secretary

OTA Publishing Staff

John C. Holmes, Publishing Officer

John Bergling Kathie S. Boss Debra M. Datcher Joe Henson

Chapter PageI. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2, AUTOMOBILE SAFETY PROBLEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Health Consequences of Accidents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7Record of Federal Safety Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Manual Seatbelts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3. ALTERNATIVES TO MANDATORY PASSIVE RESTRAINT SYSTEMS . . . . . . . . . . 15

Do Nothing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Information/Education Campaign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Mandatory Seatbelt-Use Laws . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16“Technological F ix’’.... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Economic Incentives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4. PASSIVE RESTRAINT SYSTEMS: THE BAG AND THE BELT . . . . . . . . . . . . . . . . . . 21

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21Safety and Effectiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Costs of Passive Restraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

5. COST-BENEFIT ANALYSES OF PASSIVE RESTRAINTS . . . . . . . . . . . . . . . . . . . . . . . 31

6. PHILOSOPHICAL AND ETHICAL ISSUES IN THE PASSIVERESTRAINT DEBATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

7. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

List of Tables

Table No.1.2.3.4 .

5 .6 .7.8 .

Motor VehicleMotor Vehicle

Fatalities in the United States by Age, 1979 . . . . . .Fatality Rates of Licensed Drivers by Sex and Age

Distribution of Injuries by Injury Severity . . . . . . . . . . . . . . . . . . . . .Existing Federal Standards Intended To Provide Protection in theEmergency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .SeatbeltSeatbeltRelative

Usage in 1977-78 by Car Model Year . . . . . . . . . . . . . . . . .Usage in 1977-78 by Sex, Region, and Capsize . . . . . . . .Strengths of the Two Passive Restraint Systems . . . . . . . .

Root Mean Square AIS Ratings for Occupants in Frontal Crashesby Body Region and Restraint System . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .Event of an. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . .With VDIs 3-5,. . . . . . . . . . . . . . . . .

Page778

9101021

2 5

vii

1 .

I ntrod uction

1

Introduction;

The OTA study Technology and HandicappedPeople concentrated on an examination oftechnologies developed for and used by peoplewith handicaps. An alternative, or complemen-tary focus, would be on technologies designed toprevent handicaps. There are literally hundredsof prevention technologies that constitute familiaraspects of 20th century American life. Many pre-vention technologies have derived from scientificresearch; the vaccines that prevent crippling child-hood diseases are among these. Others representa more prosaic application of common sense toeveryday problems; an obvious example is theubiquitous rubber bathtub mat. Some preventiontechnologies require the installation of elaboratesafety equipment, such as that found in a nuclearpower plant, others require only the installationof a simple idea in people’s minds—e.g., the ad-vice not to drink and drive.

Regardless of their diverse technical character-istics, all prevention technologies can be viewedin a common perspective vis-a-vis the class oftechnologies designed to treat, restore, rehabili-tate, and palliate handicapped individuals. Pre-vention technologies complement these “after-the-fact” technologies in the battle to ameliorate theconsequences of handicaps. But prevention tech-nologies also compete with treatment technol-ogies, for a primary goal of the former is to ob-viate the need for the latter. Thus, physicians canexhort their patients not to smoke, or years laterthey can attempt to assist them to learn to livewith emphysema.

Obviously, from a humane perspective, socie-ty wants to prevent all preventable handicaps. Buthere, as in so many other desirable human en-deavors, practical considerations enter—in par-ticular, what will prevention cost? In the coldestof analytical perspectives, the purely economiccosts of preventing handicaps can be compared

*NOTE: The writing of this background paper on passive re-straints in automobiles was completed in May 1982. As the paperwent to press in September 1982, the situation regarding passiverestraints had been further altered by court decisions but had notyet been finally resolved.

with the purely economic costs of dealing withpreventable handicaps after they are realized. Butthe challenge of social resource allocation deci-sionmaking calls for a more complex and some-what “warmer” cost-benefit calculus, one whichblends the economic and humanitarian concerns.In the effort to minimize the adverse consequencesof handicaps, analysts must develop measures (orat least concepts) of social cost which incorporateboth pecuniary and nonpecuniary costs, whichadd the costs of suffering to the costs of materials.

The purpose of this background paper is notto develop such a social cost-benefit calculus, noreven to array the attributes of alternative preven-tion and treatment technologies in a comparativeframework. Rather, the paper is intended mere-ly to complement the main body of the OTAstudy by introducing the prevention perspectivethrough a case study of a single prevention tech-nology. It is hoped that this will enrich policy-makers’ deliberations, implicit or explicit, on thesocial costs and benefits of alternative strategiesfor dealing with the problems of handicaps.

This case study examines issues in the debateon whether passive restraint systems—air bagsand automatic belts—should be required in allnew automobiles sold in the United States. In1977, Federal Motor Vehicle Safety Standard(FMVSS) 208, as amended, decreed that all newcars would have to have a passive restraint systemcapable of meeting a 30-mph crash performancerequirement by September 1, 1983 (1984 modelyear), with phase-in beginning with the largest1982 model cars by September 1, 1981.

On April 9, 1981, the National Highway Traf-fic Safety Administration (NHTSA) announceda delay of 1 year in implementation of FMVSS208, and new hearings were held in August 1981to consider whether the (delayed) rule should beput into effect or one of three alternatives shouldbe adopted. Two of the alternatives involved areordering of implementation dates for the varioussizes of cars; the third involved elimination of thepassive restraint requirement.

3

4 . Background paper #1: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

On October 29, 1981, NHTSA announced thatit was rescinding the requirement altogether. Law-suits to reverse the decision have been filed, andcongressional action with a similar intent has beenthreatened. Thus, though the passive restraint re-quirement has been eliminated by administrativefiat, the issue is not entirely dead.

Congressional hearings have been held onlyonce since the 1977 decision, but several agencyhearings have been held, and debate has ragedover issues as diverse as effectiveness, safety, cost,and individual liberty. The automobile, the quint-essential symbol of American affluence and in-dividualism, has thus become a battleground fora major political issue of the early 1980’s—gov-ernmental regulation.

The preceding paragraph should suggest themultidimensional significance of the outcome ofthe Federal rulemaking, but the raison d’etre ofpassive restraint systems links the debate integral-ly into a consideration of technology and handi-capped people: motor vehicle accidents annuallykill more than 50,000 Americans (over half ofthem frontseat occupants of vehicles) and inflictdisabling injuries (ranging from temporary andminor to permanent and serious) on an additional2 million people. Automobile accidents are theleading cause of death and accident-producedhandicaps among young people. Some analystshave estimated that passive restraint systems couldprevent up to half of the deaths of frontseat oc-cupants and over 100,000 moderate to critical in-juries each year.

The remaining chapters of this paper explorethe issues and evidence in the passive restraintsystem debate. Chapter 2 considers the nature andextent of the automobile safety problem, examin-ing accident, death, and injury data and review-ing the record of automobile safety standards anddevices, including the current (“active”) seatbeltsystem. Chapter 3 offers a glimpse at approachesto improving the use of passenger restraints bymeans other than mandating passive restraints.Attention then turns in chapter 4 to a descriptionof the two passive restraint systems—air bags andautomatic belts—and a presentation of data ontheir estimated costs, safety, and effectiveness inreducing deaths and disabilities. The fifth chapterreviews and compares cost-benefit analyses ofthese systems. Chapter 6 identifies and discussesthe philosophical and ethical issues related to man-

dating passive restraint systems. Concludingthoughts are presented in chapter 7.

In closing this introduction, it seems worthwhileto reflect momentarily on a peripheral but impor-tant theme—20th century technology and its im-pact on health (a theme which is well illustratedby the case of the automobile). Certainly, overall,technological developments in this century haveenhanced the quality of life and added years toits average duration. Many developments havedirectly attacked common sources of disabilityand mortality (polio vaccine is one prominent il-lustration), while other developments have re-duced health hazards indirectly (for example, theaffluence bred of modern industrialization has im-proved our diets and thereby rendered us moreresistant to disease). But 20th century technologyhas not invariably reduced health hazards. Rather,the history of new technology is one of continualredefinition of the types and sources of risks tohealth, with new hazards replacing old, and thegains generally exceeding the losses: the overalltrend in disability and mortality has been down-ward.

The automobile serves as a prime example ofthe complexity of modern technology’s role inhealth. Its invention introduced an era in whichthe time distance between a health crisis andcurative medical care would be reduced by criticalminutes, in which timely rescue from a burningbuilding would become increasingly feasible, andin which distribution of life-sustaining food andmedicine would occur ever more rapidly and in-expensively. Accompanying these health benefitsof motor vehicles, however, have been the signifi-cant health costs of street and highway travel, andthe deaths and injuries which reflect the size, struc-ture, and velocity of the vehicles, as well as char-acteristics of the roads and of the operators of thevehicles. The disproportionate impact of motorvehicle accidents on the young is particularlytragic, as thousands of lives are cut short in theirprime and healthy bodies are committed to dec-ades in beds and wheelchairs. The economic costsof treatment and rehabilitation as well as lostfuture productivity are substantial. The emotionaltoll is enormous. It is toward reducing theseburdens that the technology of passive restraintsis directed.

2Automobile Safety Problem

2 mAutomobile Safety Problem

HEALTH CONSEQUENCES OF ACCIDENTS*In 1979, motor vehicle accidents killed 51,900

Americans, including 9,400 pedestrians and42,500 nonpedestrians. The vast majority of thelatter group, 28,900, were occupants of passen-ger cars; 6,700 were in trucks, and 3,700 were onmotorcycles. An additional 2 million people re-ceived disabling injuries. All told, there were 18.1million accidents involving 29.7 million vehicles,the large majority of which resulted primarily inproperty damage or nondisabling injuries. Onein twelve vehicles registered in the United Stateswas involved in an accident, and a similar ratiocharacterized the fraction of the population in-volved in an accident. Additional consequencesincluded the following:

● 3.5 million hospital bed-days beyond initialemergency care;

● 35,700 person-years of work effort lost; and● an estimated $35 .8 b i l l ion in economic

costs, * * almost half of the total costs of alltypes of accidents.

The motor vehicle accident toll is not dis-tributed proportionately among the population.Over 40 percent of accident-involved drivers areunder the age of 25, an age group constituting justunder 23 percent of licensed drivers. Table 1shows the most tragic consequence of this phe-nomenon: the motor vehicle death rate for 15- to24-year-olds is twice the national average and fivetimes that of younger children. Table 2 demon-strates further that there is a strongly unequal sexdistribution of motor vehicle fatalities, with maledrivers’ age-specific death rates exceeding thoseof females by a factor of from 2.4 to 4.6. The

*Data in this section are from the National Safety Council (31)and the National Highway Traffic Safety Administration (29). Itshould be noted that many accident-related data are estimates andthat numerous data inconsistencies are found in the literature onmotor vehicle accidents.

**This figure includes lost wages, medical expenses, insurance ad-ministration costs, and property damage. It does not include police,fire, and court expenses, the value of lost cargo on commercial ve-hicles, etc. (31). A recent independent study estimated the costs ofmotor vehicle deaths and injuries at $20 billion (11).

Table 1 .—Motor Vehicle Fatalities in theUnited States by Age, 1979

Age Number of fatalities Fatality ratea

<5 years. . . . . . . . .5 to 14 years . . . . .

15 to 24 years . . . . .25 to 44 years . . . . .45 to 64 years , . . . .65 to 74 years . . . . .

>75 years. . . . . . . . . .Total . . . . . . . . . .

1,5002,900

18,90015,0007,9003,0002,700

51,900

9.68.4

45.725.018.019.628.823.6

aDeaths per 100,000 population.SOURCE: National Safety Council, Accident Facts (Chicago: NSC, 1980).

Table 2.—Motor Vehicle Fatality Rates ofLicensed Drivers, by Sex and Age

Fatality ratea

Age Males Females

<20 years. . . . . . . . . . . . . . 131.8 35.120 to 24 years . . . . . . . . . 107.9 23.625 to 34 years . . . . . . . . . 69.1 15.235 to 44 years . . . . . . . . . 52.1 13.745 to 54 years . . . . . . . . . 44.6 12.055 to 64 years . . . . . . . . . 36.4 12.5

365 years . . . . . . . . . . . . . . 37.0 15.6aDeaths per 100,000 population.

SOURCE: National Highway Traffic Safety Administration, Highway SafetyFacts 3 (Washington, D.C.: Department of Transportation, 1978).

worst rate for females (women under the ages of20) is less than the best rate for males (men 55to 64 years old). More than 1 in every 10015-year-old boys will die in an accident before theage of 25, a death rate 20 times higher than thatattributable to polio at its worst (13). *

Unfortunately for the purposes of this study,there are no good national data on the types and

● Age-specific motor vehicle death rates are in part a functionof exposure. For example, the number of passenger-miles per yearvaries from 16,000 for males in their early thirties to under 3,000for elderly women. From their mid-twenties to early sixties, menride from 50 to 100 percent more passenger-miles than women. Asa result, fatality rates per 100 million passenger-miles by age differfrom the simple age-specific fatality rates. The young males’ ratesremain the most socially alarming, but the highest rates per 100million passenger-miles belong to the very elderly (over 75) (6).

7

8 ● Background paper #l: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

numbers of handicaps that result from motor vehi-cle accidents. The National Highway Traffic Safe-ty Administration (NHTSA) is trying to refinedata collection to produce such information, buttoday’s data on disabilities do not permit a usefulassessment (34). Currently, data indicate severi-ty of injury but do not follow through on the out-comes of injuries.

Table 3 presents the percentage distribution ofinjuries in a recent NHTSA sample by injuryseverity. The data indicate that over 70 percentof all injuries are scored as minor on the Ab-breviated Injury Scale (AIS), a common index ofinjury severity. Under 12 percent of injuries arein categories (AIS 3 to 5) in which survival is prob-able or possible and in which serious handicapscould be a result. Given the total number of in-juries, however, this relatively small percentagestill represents many tens of thousands of people.Furthermore, moderate injuries (AIS 2) can alsoresult in disabling handicaps. Thus, while reliabledata on accident-produced handicaps are notavailable, injury severity data suggest the substan-tial probable burden.

Despite the “bad news” contained in the abovemotor vehicle accident data, there is also goodnews. During the 1970’s, total motor vehicledeaths declined 7 percent—dropping from 55,791deaths in 1969 to 51,900 in 1979. Given growthin the population and in the number of registered

Table 3.—Distribution of injuries by injury Severity

AIS PercentagePercentage in class

Code no. Definition of persons surviving

1 . . . . . . . . . Minor 71.1 99.9892 . . . . . . . . . Moderate 16.4 99.8783 . . . . . . . . . Serious, not life 8.1 99.158

threatening4 . . . . . . . . . Severe, life threatening 2.2 91.9785 . . . . . . . . . Critical 1.5 41.7996 . . . . . . . . . Maximum injury, 0.8 0.000

virtualIy unsurvivable

SOURCE: S. Partyka, “Effects of Traffic Accident Injuries on the WorkforceEstimated From NCSS and NASS Data” (Washington, D. C.: NationalHighway Traffic Safety Administration, National Center for Statis-tics and Analysis, June 1981).

vehicles, the decline in death rates was quitedramatic. For example, the death rate per 10,000registered vehicles dropped 37 percent, from 5.19to 3.26. In 1979, the death rate per 100 millionvehicle-miles stood at an all-time low of 3.4; inthe 1940’s and earlier, that rate was in the teens.The death rate per 100,000 population fell by 15percent over the course of the decade; this in-cluded a lo-percent decrease in the 15- to 24-year-old age category, the smallest age-specific decline.A variety of factors contributed to these improve-ments, including the national 55-mph speed limitimplemented in 1974, the increasing price ofenergy in the mid to late 1970’s, and automobilesafety features (47). We now turn to a look at therecord of Federal governmental regulation ofautomobile safety.

RECORD OF FEDERAL SAFETY STANDARDSAs substantial as the accident toll is today, there

is evidence that it would have been considerablygreater in the absence of existing Federal Govern-ment safety regulations. Analyzing data from theFatal Accident Reporting System of NHTSA,Robertson (37) has estimated that for the years1975 through 1978, some 37,000 fewer deaths oc-curred than would have been expected in theabsence of the Federal safety standards.

Robertson observed a total death rate of 5.5persons per 100 million vehicle-miles for cars notsubject to safety regulations, but a rate of only3.4 for cars meeting the Federal safety standards.A differential characterized all classes of victims

(including pedestrians, motorcyclists, and pedal-cyclists), but occupants of automobiles meetingthe standards realized the greatest benefit, witha death rate of 1.5 per 100 million vehicle-milescompared with 2.9 for occupants of vehicles notsubject to the standards.

Robertson did not examine the impact of safe-ty standards on injuries and disabilities, but aqualitatively similar benefit would be expected. *Other studies also have documented decreased oc-

● It is possible that a safety standard which reduced deaths mightthereby result in increases in nonfatal injuries. The evidence gatheredto date, however, suggests that reductions in fatal and nonfatal in-juries go hand-in-hand.

Ch. 2—Automobile Safety Problem

cupant deaths and nonfatal severe injuries asso-ciated with State and Federal safety regulations(7,16,21,35).

Most Federal safety standards are technologicaland fall into two categories: vehicle crash-worthiness and crash avoidance. Table 4 listsprominent examples of existing Federal MotorVehicle Safety Standards (FMVSSs) intended toprotect automobile occupants solely in the eventof an emergency; these standards are similar inintent to passenger restraint systems. It should benoted that nontechnological standards can havea comparable or greater effect in reducing the ac-cident health toll; the national 55-mph speed limitstands as perhaps the most prominent example.In a 1979 study, the National Safety Council(NSC) estimated that there would have been anadditional 5,500 motor vehicle, accident deaths peryear were it not for reduced speeds and a narrowerspeed distribution, both of which are attributableprimarily to the 55-mph law. NSC also estimated

that if all States had raised their speed limits to65 or 70 mph in 1978, an additional 5,200 to 7,800deaths would have resulted that year (47).

The magnitude of the remaining accident healthtoll suggests the considerable potential for usingadditional technology to further decrease high-way-produced deaths and disabilities. High usagerates of passenger restraint systems alone couldreduce the toll by half (see ch. 4). Belted occupantsof automobiles experience 50 percent fewer high-way deaths than unbelted occupants (40)—yet itis estimated that only 10 to 15 percent of thepopulation wears seatbelts, and the percentage hasbeen falling in recent years (15). Automatic (i.e.,passive) restraint systems are advocated precise-ly because of this failure of the vast majority ofthe population to use manual (i.e., active) belts.The reasons for, and implications of, the nonuseof manual belt systems are considered furtherbelow.

Table 4.—Existing Federal Standards Intended To Provide Protectionin the Event of an Emergency

FMVSSTitle Automatic performance required

105-75 . . . . . . . . .

110 ., . . . . . . . . . .

111 . . . . . . . . . . . .

201 . . . . . . . . . . . .

203 . . . . . . . . . . . .

205 . . . . . . . . . . . .

212 . . . . . . . . . . . .

215 . . . . . . . . . . . .

301 . . . . . . . . . . . .

Hydraulic brakesystems

Tire selection and rims

Rearview mirrors

Occupant protection ininterior impacts

Impact protection forthe driver from thesteering control system

Glazing materials

Windshield mounting

Exterior protection

Fuel system integrity

Requires split brake system for redundancy if primarysystem fails

Rim must retain tire from 60 mph to stop after rapiddeflation

Breakaway inside mounting for mirror

Contactable interior surfaces must be padded or meet per-formance requirements with headform impact at 15 mph

Steering assembly must absorb driver impacts under con-trolled crash criteria

Windshield has a high penetration resistant inner layer

Requires windshield mounting retain specified periphery ofwindshield in crashes

Provides vehicle protection in certain crash impacts

Provides protection against fuel systems rupture andleakage in crashes

SOURCE: W. Haddon, Submission of Documents to W. Coleman, Jr., Secreta~, Department of Transportation, Washington,D. C., Sept. 17, 1976.

10 ● Background paper #1: Mandatory passive Restraint Systems in Automobiles: Issues and Evidence

MANUAL SEATBELTSFront-lap seatbelts were first installed in all cars

as standard equipment in 1964, when 14 Statesrequired them (41). By the late 1960’s, lap andshoulder belts were required as standard equip-ment on all new cars sold in the United States.Thus, almost all cars on the road today areequipped with lap or lap/shoulder belts; the vastmajority have the lap/shoulder combination.Studies indicate that, when worn, belts reduce therisks of death and serious injury by 50 percentor more (28,40). Although most people may notbe familiar with the precise statistics, virtuallyeveryone is aware that “seatbelts save lives, ” asthe publicity slogan put it years ago.

Yet, according to recent observations of manualbelt use, only 11 percent of drivers wear their belts(15). Usage rates have fallen in recent years,following a brief period of increases in the mid-1970’s. The increases in the mid-1970’s may havebeen attributable in part to the ignition-interlocksystems that were installed on 1974 and some 1975model cars. The interlock systems provoked aloud and angry public response as drivers foundthemselves unable to start their cars when theyplaced cargo on the passenger seat (e.g., groceriesor the family dog) that exceeded the weight min-imum which activated the system. The congres-sional response was to prohibit the Departmentof Transportation from requiring the system onlater cars (41).

Table 5 shows how belt usage varied in 1977-78by automobile model year. The American Auto-mobile Association claims higher usage rates insome of the (then) newer model cars with im-proved belt systems (13). Table 6 indicates howbelt usage varied by sex, region of the country,and car size.

The result of such low usage rates is that man-ual belts in cars currently reduce the fatality andserious injury rates by less than 10 percent. Thus,a tremendous potential for saving lives and pre-venting injuries is going unrealized, despite theready accessibility of the technology and the rel-ative ease of its use.

Table 5.—Seatbelt Usage in 1977-78 by CarModei Year

Usage (percent)

Car model year Lap only Lap and shoulder Total

1 9 6 4 - 6 7 . . . . . . . . . , . 8 . 91968 -71 . . . . . . . . . . . 8.51972-73 . . . . . . . . . . . 12.71974 . . . . . . . . . . . . . . 2.71975 . . . . . . . . . . . . . . 1.71976 . . . . . . . . . . . . . . 1.41977 . . . . . . . . . . . . . . 1.11978 . . . . . . . . . . . . . . 1.2

—2.44.2

14.412.712.212.212.5

8.910.916.917.114.413.613.313.7

SOURCE: National Highwai{ Traffic Safety Adminlstration, Occupant Protec-t)orr Program Progress Report No. 2 (Washington, D. C.: Departmentof Transportation, April 1979).

Tabie 6.—Seatbeit Usage in 1977=78 by Sex,Region, and Car Size

Characteristic/Car model years Total usage (percent)

Sex, 1964-78:Male drivers . . . . . . . . . . . . . . . . . 12.6Female drivers. . . . . . . . . . . . . . . 16.4

Region of country, 1964-78:West coast. . . . . . . . . . . . . . . . . . 18.3All other regions (average) . . . . . 11.4

Car size, 1976-78:Subcompact . . . . . . . . . . . . . . . 19.5Compact . . . . . . . . . . . . . . . . . . 12.5Intermediate . . . . . . . . . . . . . . . 10.3Full , . . . . . . . . . . . . . . . . . . . . . . . 9.6

SOURCE: National Highway Traffic Safety Administration, Occuparrt Protec-tion Program No. 2 (Washington, D. C.: Department of Transporta-tion, April 1979).

What accounts for the extremely low usagerates? Two factors invariably cited in polls arediscomfort and inconvenience (28), although asurvey sponsored by General Motors (GM) hasidentified other factors as being of greater sig-nificance (e.g., fear of being trapped in a vehi-cle) (18). The issue of discomfort reflects belts’pressuring or abrading hips, chests, and necks andcreating an unpleasant sense of confinement orrestriction of movement. For some people, incon-venience refers simply to the minor effort involvedin buckling a well-functioning belt system, whilefor others it relates to difficulties in retracting thebelts or latching or releasing them.

Ch. 2—Automobile Safety Problem ● 11

In a study of its own, NHTSA concluded thatmany of the complaints about belts were wellfounded. A representative sample of Detroit-areadrivers identified moderate or serious problemswith comfort or convenience in all of the 30 carstested, with the best-performing car cited as hav-ing a problem in 35 percent of the trials and theworst-performing car cited in 85 percent. In an-ticipation of automatic belt systems, NHTSA hasbeen developing comfort and convenience speci-fications intended to address these problems (28).If the GM-sponsored study is correct, however,improvements in comfort and convenience maynot lead to significant increases in belt usage (18).

The decision not to wear a belt presumably re-flects a judgment that the disutility associated withdiscomfort, inconvenience, or other factors out-weighs the perceived utility of reducing risk.Arnould and Grabowski (3) offer two related ex-planations for why this judgment is so common.

The first explanation, referred to as the “insen-sitivity-to-low-probabilities” hypothesis, suggeststhat for very low-probability events (such as aserious car crash on a single outing), individualsbecome insensitive to the high potential cost ofnot protecting themselves and indeed may notcomprehend the meaning of the tiny probabili-ty; instead, they respond primarily to the unlike-lihood of the event.

Arnould and Grabowski discuss a study (44)which illustrates the principle: two groups of ex-perimental subjects were given data on the prob-ability of experiencing a fatal or disabling acci-dent. One group was given the figures for a life-time (SO years) of driving (a l-in-100 chance ofa fatal accident and a l-in-3 chance of at least onedisabling injury), while the other group receivedthe same information calculated on a per-trip basis(where the odds of the accident outcomes areminuscule). Compared with the latter group, thegroup given the lifetime figures responded by in-dicating a much greater increase in expected seat-belt usage and a greater disposition toward seat-belt laws.

Arnould and Grabowski also present evidencethat people significantly underestimate their riskof involvement in an automobile accident. For ex-ample, a recent survey (45) queried: How likely

do you think it is that you will be involved in anautomobile accident of any kind in the next year?Fewer than a quarter of the survey respondentsselected an answer equal to or greater than theactual societywide average, roughly 1 in 10. Amajority selected odds of 1 in 100 or smaller still.

The second explanation for the judgment thatthe disutility of buckling up outweighs the utili-ty of protection is that the expected value of theprotection for any given trip is extremely low,owing to the low probability of a serious accident,and thus, a driver (or passenger) simply may valueavoidance of discomfort or inconvenience morethan protection. This will be particularly true if,as above, the individual significantly underesti-mates the probability of an accident.

Arnould and Grabowski estimate the annualper-person benefits of buckling up at between $38and $78 (in 1975 dollars) and state that, at prevail-ing wage rates (the assumed opportunity cost oftime), this amount must exceed the time costs ofbuckling up. Hence, they conclude that a rationalweighting of costs and benefits cannot explain thefailure of so many people to wear their seatbelts.The authors acknowledge that “there may be sig-nificant discomfort costs to some individuals towearing seatbelts. ” Nevertheless, they suggest, “itwould seem hard to argue that [this] would sochange the . . . benefit-cost calculus to explain the80- to 90-percent current nonutilization rate ofseatbelts. ”

While it is agreed that discomfort costs couldnot explain the entirety of nonutilization, it couldbe argued that they might explain much of it, par-ticularly if “discomfort” is defined to include thepsychological discomfort of those who fear beingtrapped in their cars by belts. Certainly, manyriders who experience discomfort from seatbeltswould accumulate hundreds of hours of discom-fort if forced to wear them, and one would notneed to value discomfort time highly to concludethat nonuse was rational behavior for these indi-viduals.

Numerous characteristics differentiate beltwearers from nonusers (41). The latter tend tohave less education than the former, rate belts asmore uncomfortable and inconvenient, and aremore likely to be smokers. The victims of serious

—

72 ● Background paper #1: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

crashes are less likely to wear belts than people risk-averse, or perhaps more risk-loving, thannot involved in serious accidents, especially in the their belted counterparts.case of youthful drivers and drivers under the in-fluence of alcohol. Furthermore, unbelted drivers Whatever the explanation, the fact remains thattend to follow the cars in front of them closer than only a small minority of automobile occupantsdo belted drivers. Collectively, all of these char- choose to wear seatbelts. The consequence inacteristics suggest that unbelted drivers are less human destruction is tragic.

—

3 .

Alternatives to Mandatory PassiveRestraint Systems

Alternatives to Mandatory PassiveRestraint Systems

Prior to its rescission, Federal Motor VehicleSafety Standard (FMVSS) 208—the requirementthat all new cars come equipped with passiverestraint systems—represented a policy judgmentthat passive restraints were the best way to ad-dress the failure of the vast majority of Americansto wear their (manual) seatbelts. But passive sys-tems are not the only option; other alternativeshave been suggested and some have been tried.

Advocacy for the passive restraint alternativereflects in part the experienced or predicted failureof the other alternatives. In this chapter, the alter-natives are briefly identified, and the evidence thatrelates to their effectiveness is presented. (Ch. 4considers the nature, effectiveness, and costs ofpassive restraint systems. ) It should be noted thatthe alternatives listed below are by no means mu-tually exclusive:

1. Do nothing—i.e., leave the current manuallap/shoulder belts in place, leave the deci-

00 NOTHINGThis alternative has few proponents in the po-

litical arena. Lacking some bone tossed in thedirection of promoting passenger restraint, the “do

2.

3.

4.

5.

sion of restraint to the individual automobileoccupant, and do nothing to promote the useof manual belts. *Actively promote the use of manual beltsthrough media campaigns and other educa-tional efforts. This is one of the options theReagan administration has favored.Pass legislation requiring the wearing ofseatbelts.Require a “technological fix” that forcesmanual belt use (e.g., the previously triedignition-interlock system).Offer economic incentives to automobile oc-cupants to wear their seatbelts.

● A still more laissez-faire option would be to eliminate the re-quirement that cars have manual belt systems as standard equip-ment, leaving purchase and installation of belts as a buyer option.There are sound theoretical reasons to oppose this possibility (dis-cussed in ch. 6 below), in addition to the political uproar it wouldcreate.

INFORMATION/EDUCATION CAMPAIGN

nothing” alternative must be dismissed as beingof unlikely political viability.

The theoretical appeal of this alternative is con-siderable, and the Reagan administration hasadopted this approach as its policy to promoteuse of passenger restraints. Provision of informa-tion—education—is the conceptually appropriateapproach if one perceives the problem of beltnonuse as resulting from riders’ ignorance of thetrue risks of automobile travel and/or the true ef-fectiveness of wearing belts. Public education ad-dresses the public’s information deficit and still

preserves the freedom of (informed) riders tochoose not to be restrained (see also ch. 6).

Borrowing from related health education ex-perience, one might be tempted to conclude thatmajor publicity efforts could increase the volun-tary use of seatbelts. For example, there isevidence that the antismoking campaign of thepast nearly two decades has had a substantial im-pact on cigarette smoking (49), and specifically,

15

16 ● Background Paper #l: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

that the major broadcast media antismoking cam-paign of 1968-70 significantly reduced cigaretteconsumption (48). Unfortunately, however, theevidence on whether the smoking experience gen-eralizes to the case of seatbelt use is not encourag-ing and is at best ambiguous.

Influenced by evidence such as the following,a number of experts (30,41) have concluded thatinformation/education campaigns are unlikely tosignificantly increase belt usage. In 1968, the Na-tional Safety Council (NSC) received $51.5 millionworth of media public service announcement timeto encourage seatbelt use; similar NSC campaignswere mounted in 1972 and 1973, yet interviewdata indicated no change in reported seatbeltusage (15). Several controlled experiments andquasi-experiments with media promotion of beltuse also have failed to produce increases in use(8,43). The American experience is echoed by ex-perience in Canada, Great Britain, and France:major publicity campaigns either did not increasebelt use at all or, at best, they increased use slight-ly and only in a transitory manner; i.e., cam-paign-induced increases disappeared soon afterconclusion of the campaign (15).

Proponents of the information/education ap-proach claim that it has not been given a fair trial,that more sophisticated understanding of the fac-tors influencing seatbelt usage will permit develop-ment of more effective education and publicitypackages. They also argue that an effective cam-paign requires a long-run approach, with suc-cessive education and information efforts reinforc-ing preceding ones and gradually converting non-belt users into users.

However, even the most optimistic assessmentsuggests that an “all-out” campaign could notboost usage rates to greater than 40 percent (13).Thus, if one’s objective is to reduce preventablemotor vehicle fatalities by as much as possible,this alternative clearly cannot be relied on byitself. *

MANDATORY SEATBELT-USE LAWS

● The difficulty of promoting truly informed decisionmaking—the objective of this alternative—is more substantial than one mightthink at first. As Arnould and Grabowski (3) demonstrate, people’sunderstanding of accident risks is poor, and the statistical or prob-abilistic nature of such risks constitutes a subtle message to com-municate to a comprehending public. This is illustrated by evidencethat the public’s understanding of the health hazards of cigarettesmoking is remarkably unsophisticated despite nearly two decadesof publicity, education, and public discussion of those hazards (26).

This alternative—passing mandatory seatbeltuse laws—has precedents around the world in wellover a dozen countries. In Australia, NewZealand, Canada, Sweden, and West Germany,mandatory belt laws have been in effect forseveral years and prelaw and postlaw belt-usagedata are available. The first such law, in the Stateof Victoria, Australia, produced belt-use rates of70 to 80 percent and accounted for reductions infatality rates of 20 percent in urban crashes and10 percent in rural crashes. In the Provinces ofOntario and Quebec in Canada, “weakened” laws(i.e., exempting shoulder belt use because of thepublic’s concern about discomfort) have producedusage rates in the vicinity of 40 to 50 percent. Fouryears after passage of its belt-use law in 1972, NewZealand was reported to have a compliance rateof from 80 to 90 percent. An 80-percent rate hasbeen reported for Sweden; and in West Germany,

observed rates have ranged from 45 to 80 percent(9).

No U.S. State has adopted a seatbelt law ap-plying to drivers. The traditional American senseof independence :normally would create signifi-cant opposition to such laws—a poll 4 years agofound more than three-quarters of respondentsopposing them (13)—and the antiregulatory cli-mate of the early 1980’s presumably would in-crease the level and intensity of opposition. Asa representative of the Consumer Federation ofAmerica stated (13):

All too often State legislators, faithfully reflectingthe sentiment of the people in the State, see mandatorybelt legislation as an outrageous and unnecessary in-trusion into people’s lives. Whether one agrees withthat view or not, it is a sentiment that is prevalent inthe large majority of our States.

Ch. 3—Altematives to Mandatory Passive Restraint Systems ● 17

This mentality is reflected in the fact that overhalf of the States have recently repealed laws re-quiring motorcyclists to wear helmets, despitesolid evidence that helmet laws save lives (25,51).

Several States—including Tennessee, RhodeIsland, and Michigan—have adopted laws requir-ing that children be restrained in moving auto-mobiles. The laws often include exemptions or ex-ceptions (e.g., permitting babies to be unre-strained when nursing). Observational studiesconclude that belt-usage rates by children haveincreased, though they remain low overall (ap-proximately one-third) (41).

“TECHNOLOGICAL FIX”Also unlikely is the “technological fix”

alternative-unless a system can be designed thatis technically superior to, and much more publiclypalatable than, the ignition-interlock systemfound in 1974 (and some 1975) model cars. Thetechnical problems with that system and theresulting public furor led to the elimination of theFederal requirement of the system.

A sequela of that experience was that FMVSS208 prohibited use of an interlock system in any

ECONOMIC INCENTIVES

The remaining alternative-offering economicincentives to wear manual belts—has not beentried directly, though a few insurers do offerpremium discounts of up to 30 percent on medicalor personal injury protection coverage for carsequipped with passive restraints (air bags or auto-matic seatbelts) (3). It is conceivable that premiumdiscounts or increases in coverage could begranted users of manual belts, but the problemof verifying compliance is not a minor one. Sim-ilarly, insurers could offer positive or negativebenefit incentives. For example, medical paymentscould be increased for belt-wearing accident vic-tims (an approach adopted by at least one in-surer), or payments could be conditioned on beltuse at the time of an accident.

Mandatory belt-use laws are potentially themost effective approach to ensuring passengerrestraint. Experience in other industrialized coun-tries suggests that a mandatory law might resultin usage rates exceeding those achievable withpassive seatbelts, because so many passive beltswould be detached. Nevertheless, in today’s po-litical climate in the United States, mandatoryseatbelt-use laws seem unrealistic. It also shouldbe noted that this assessment of belt laws has ut-terly ignored theing the laws.

question—and cost—of enforc-

passive restraint system developed to satisfy the(then-existing) requirement. Certainly, it is possi-ble that passive belt systems could be developedwhich would make permanent, or even tem-porary, disconnecting difficult or inconvenient,and this might have a marginal impact on passivebelt use. But given the public’s clear oppositionto evident and burdensome “technological fixes, ”manufacturers and the Government seem unlikelyto go this route.

If the compliance problem could be resolved(mechanical solutions are conceivable), the poten-tial for the premium approach is intriguing. Fora car with automatic belts, Nationwide Insurancehas estimated premium savings of roughly $20 peryear. Over the lifetime of a car, this translates intoa present discounted value of $150 (33). Sincemany automatic belts are disconnected, * assur-

● A survey found a disconnect rate of 22 percent among ownersof VW Rabbits equipped with automatic belts. The finding thatowners of Rabbits with manual belts tend to use their belts morefrequently than the average automobile owner suggests that the auto-matic belt disconnect rate under a mandatory passive restraint lawmight be greater than 25 percent. The rate could be considerablygreater. In particular, the automatic-belt Rabbits have had an inter-lock system, so disconnection requires action after the car has started–perhaps a more active form of passive-belt rejection (23),

18 ● Background paper #1: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

ance of compliance with manual belts should beworth at least as much as the initial presence ofautomatic belts. Thus, the lifetime insurance valueof buckling up might prove to be a significantincentive.

Even if the insurance incentive “worked,” itseems unlikely that it would raise effectiverestraint rates high enough to satisfy people whowant to see maximum reduction of the highwaydeath and disability toll. Aside from paternalism,an argument grounded in the existence of negativeexternalities suggests that self-selected compliancerates will be too low from a social point of view(see ch. 6). Furthermore, the economic incentiveapproach would take time to achieve widespreadeffect—compliance technology would have to bedeveloped and installed, insurance companieswould have to be sold on the desirability (and

amount) of discounts, etc. Despite these draw-backs, the incentive approach would seem to war-rant more attention than it has received to date.To my knowledge, it is not now, nor has it been,an option under serious policy consideration.

This chapter’s review of the policy alternativesshould suggest one of the reasons many peopleconcerned with automotive safety have been sosupportive of passive restraint systems: for onereason or another-be it effectiveness or politicalacceptability-each of the alternatives, consideredindividually, has significant drawbacks. There isa school of thought that advocates a mix of severalof these alternatives as a cost-effective means ofachieving effective passenger restraint (9,18).Given the recent demise of FMVSS 208, it seemsprobable that more attention will be directedtoward a multiple-approach strategy.

4Passive Restraint Systems;

The Bag and the Belt

.

4Passive Restraint Systems;

The Bag and the Belt

BACKGROUNDThere are only two major passive restraint

systems: air bags (or cushions) and automaticbelts. Table 7 identifies their relative strengths(many of which are discussed in this chapter), butthe important point is that, when used properly,both passive restraint systems are quite effectivein reducing the risk of death and serious injury.

Technically, the automatic belt system is quitesimple. Most systems in operation consist of asingle shoulder belt (a “two-point system”) thatcrosses the rider automatically upon entry intothe car, and a padded knee panel below the dash-board to provide added protection. Automaticbelts with a lap as well as a shoulder component(a “three-point system”) have also been designed.Some systems offer automatic shoulder belts andmanual lap belts. Close to half a million automaticbelt systems are currently on American roads,almost all of them in VW Rabbits and ChevroletChevettes, offered as part of option packages.

Air bags consist of deflated bags situated in thesteering wheel (driver) and glove compartmentarea (passenger) sides which inflate virtually in-stantaneously when front end or dashboard sen-sors detect crash forces substantial enough to beharmful. Different systems have different infla-tion mechanisms, and all include ancillary equip-

Table 7.—Relative Strengths of the Two PassiveRestraint Systems

Air bag Automatic belt

Less obtrusive Less expensive

Less uncomfortable No chemicals involved

Greater protection in most serious Potentially greater protectionaccident situations (particularly in some accident situa-when used with manual lap tions a

belt)a

Less likely to be disconnected Redeployment followingemergency use lessexpensive

asee the discussion of safety and effectiveness below.

ment (e.g., knee restraints to prevent riders fromsliding under bags, a readiness monitor, and anindicator light) (28). Altogether, over 10,000 air-bag-equipped cars have accumulated over a bil-lion miles on American roads since they first ap-peared in 1972.

This more complex technology has a historywhich dates back 30 years. In 1952, the first ofseveral patents for automatically inflating aircushions was filed. Federal Government interestdates from 1968, when prototype development be-came sufficiently advanced to consider large-scaleapplication in the near future. In 1969, the Na-tional Highway Safety Bureau (NHSB, predeces-sor of the National Highway Traffic Safety Ad-ministration, NHTSA) announced a proposedrulemaking for an “Inflatable Occupant RestraintSystem,” with an initially proposed effective dateof January 1, 1972. This marked the beginningof a longstanding adversarial relationship betweenautomobile manufacturers and the Department ofTransportation (DOT) which saw dozens of de-bates over the effectiveness and cost of passiverestraints and delays in implementation of, andchanges in, a passive restraint requirement.

Delay and change have occurred during eachsuccessive administration, with the Reagan ad-ministration’s re-examination of the issue beingthe latest and most radical. The courts, Congress,public interest groups, and the media have all beenactive participants in the drama. The conflict hasmade strange bedfellows of such diverse interestsas Ralph Nader and the Pacific Legal Foundation(PLF), the latter a public interest group advocating“limited government.” In 1977, PLF filed suit inthe U.S. Court of Appeals to block the passiverestraint rule, claiming that DOT had an “insuf-ficient basis for the air bag decision.” The follow-ing year, Nader and Public Citizen, the consumerrights group, filed suit in the Court of Appeals

21

zz ● Background Paper #1: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

seeking a ruling that DOT’s scheduled 3-yearphase-in of restraints was illegal; plaintiffs wantedall new vehicles to have passive restraints at thesame time. The Court of Appeals consolidated thesuits of PLF and Nader and Public Citizen. *

Part of the regulatory debate picture has beenrepeated proposals by automakers to introduceair bags or automatic belts on their own. Eachof these proposals has come in response to a pro-posed passive restraint rulemaking by the Govern-ment. As each of the proposed rulemakings wasaltered or delayed, often because of agency ac-tions, occasionally because of judicial decisions,the automakers’ plans were themselves altered,invariably in the direction of limiting introduc-tion of passive restraints.

In 1970, for example, General Motors (GM) in-formed NHSB that it would provide air bags onall its cars by 1975, introducing the equipment asan option and then converting it to standardequipment. In 1973, GM informed DOT that itwas reducing its planned production of air-bag-equipped 1974-75 cars from 1 million to 150,000units, blaming both tooling difficulties and theGovernment’s standard-setting process. Half ayear later, a GM spokesperson told the InsuranceInstitute for Highway Safety (IIHS) that the figureof 150,000 was probably too high.

In fact, GM built about 10,000 air-bag-equippedcars, all large models, during the 1974-76 modelyears. The company then canceled production,

● For a detailed chronology of events in air bag development,debate, and rulemaking, see Back,ground Manual) on the OccupantRestraint Issue (13).

claiming insufficient consumer demand, a claimwhich GM officials have acknowledged was basedin part on a failure of the company to promotethe technology. GM’s revising its plans and de-parting from the air-bag-equipped automobilemarket followed an earlier court decision over-turning a proposed NHSB air-bag rule.

A similar picture emerges in the years since themid-1970’s, with promises of air-bag-equippedvehicles repeatedly made and then scaled back orrescinded (17). Recently, GM announced the ter-mination of its inflatable restraint program, call-ing the device economically infeasible (23).

Three years ago, the “Big Three” domestic auto-mobile manufacturers and several foreign pro-ducers reported to NHTSA that they were work-ing on belt and bag systems and intended to in-troduce them, as options, on several modelswithin the next few model years (28). Even priorto the rescission of Federal Motor Vehicle SafetyStandard (FMVSS) 208, several of the announcedintentions had not been realized.

The technology for both bags and belts is de-veloped and available. As is discussed in the sec-tion below, when used properly, these technolo-gies are commonly acknowledged to work, tosave lives, and prevent disabling injuries. What isless clear is whether the American car rider willwear belts, active or passive, and whether the con-sumer’s car-buying propensity will be reduced asa result of passive-restraint-induced car price in-creases. This too is examined below.

SAFETY AND EFFECTIVENESSFollowing a review of the evidence, William reduce in severity over one hundred thousand

Coleman, Secretary of Transportation in the Ford moderate to critical injuries per year” (13).administration, concluded that passive restraints“are a reliable and effective means of substantially The precise quantitative findings can be chal-reducing death and injuries on the Nation’s high- lenged-for example, recent estimates have placedways.” If air bags were installed on all cars, Cole- the life-saving potential of passive restraints in theman estimated, they “would probably save over vicinity of 6,000 to 9,000-but the qualitative con-twelve thousand lives annually and prevent or elusion is beyond dispute. All of the quantitative

Ch. 4–Passive Restraint Systems: The Bag and the Belt ● 23

evidence supports it and the logic is impeccable:“The principle behind an occupant restraint is thatan occupant is much less likely to be injured orkilled in an automobile crash if the crash forcesare applied in a controlled way to the strongestparts of the human body” (28). Furthermore, pro-ponents of passive restraints argue, occupant re-straints are much more likely to be used if theyare automatic (i. e., passive) than if they requireaction on the part of the occupant, as do today’smanual belts.

Passenger restraint systems are not perfect; theyreduce but do not invariably prevent deaths andinjuries. Neither belts nor air bags are particularlyeffective in rear-end collisions, and the twotechnologies have different merits and liabilitiesin other types of crashes. In the instance of an im-pact on the passenger side of a car, for example,belts are effective in keeping the driver from beingpitched to the side. Air bags occasionally inflateon side impact, and the passenger-side bag canthen protect the driver, but inflation of the bagsis uncertain. If the bags do inflate, they shield thedriver from flying glass and metal and therebyafford the driver a form of protection that beltscannot offer.

The technology of passive restraints might posehazards. Concern has been expressed about thetoxicity of sodium azide, a chemical in compoundsused in many air bag systems to inflate the bags.Testimony has been offered indicating that thechemicals are confined sufficiently so that car oc-cupants are not exposed to them, either before orduring deployment of a bag. NHTSA is convincedby the evidence, but is quick to point out thatother, nonchemical inflation mechanisms havebeen developed (13).

Other concerns expressed about bags includefears that inadequately restrained children couldslide beneath and be smothered by the bags, oreven conceivably be thrown backwards into therear window. There are worries that air bagsmight inflate spontaneously without an impact oron minor impact, thereby causing a serious acci-dent (12). Again, the bulk of the evidence is thatthese are not significant problems. The trackrecord of the more than 10,000 air-bag-equippedcars is quite impressive in this regard (13,28).

There may have been isolated incidents, andmight be others in the future, but the life-savingand injury-reduction potential of air bags dwarfsthese adverse outcomes.

The principal problem with automatic belts isthat they can be disconnected. They have to beequipped with a safety release for emergencies.For the person for whom belt-wearing is trulyburdensome, passive belts can be actively disen-gaged either permanently or repeatedly. In effect,the individual converts the passive restraint sys-tem into an active nonrestraint system, and ef-fectiveness goes to zero.

A common worry expressed about belts (man-ual or automatic) is that they may prevent a per-son from being thrown from a car in an accidentin which remaining within the car would proveto be more damaging. This outcome might occurin rare instances, but the opposite outcome ismuch more probable—i.e., an unrestrained oc-cupant is much more likely to be injured or killedby being thrown from (or around) the car thanby being restrained within it.

In the remainder of this section, the evidenceon the effectiveness of belts and bags in reducingdeath and injury is presented and discussed. Thefollowing caveats should be kept in mind in in-terpreting the data:

1.

2.

Air-bag field experience relates virtually ex-clusively to large-size and luxury cars, prin-cipally the GM vehicles produced in the1974-76 model years. Whether this experi-ence generalizes directly to all cars (withalmost all new cars much smaller than these)remains to be seen.

Automatic belt experience in the VW Rab-bit, the source of most relevant data, mightnot typify general experience once all carswere so equipped. Small-car buyers in thelate 1970’s exhibited a greater propensity tobuckle up than did the average driver of theperiod: whereas average drivers buckled uponly 10 to 15 percent of the time, owners ofRabbits with manual (active) belt systemswore their belts 34 percent of the time (28).Thus, in a world of passive seatbelts, therewould be reason to expect a disconnect rate

24 ● Background Paper #1: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

in the general population greater than thatof automatic-belt Rabbit owners, particular-ly since DOT policy could not require thekind of interlock device found in automatic-belt Rabbits (23).

3. Related to the above point, current users ofpassive restraint systems represent a largelyself-selected group. It is possible that theirmotivation, particularly their concern aboutdriving safety, differs enough from that ofthe overall population to invalidate gener-alizations based on their experience. Mitigat-ing this concern are studies which control foraccident severity before assessing the effec-tiveness of the restraint systems. Also, in theVW case, the passive belts were part of a lux-ury option package, one sufficiently expen-sive that it seems unlikely that people wouldbuy the package simply to get the belts. Fur-thermore, evidence from claims data in-dicates almost identical rates of accidentclaims and size of damage awards for Rab-bit owners with and without passive belts(3). Nevertheless, it should be recognizedthat victims of serious crashes are less oftenrestrained than nonvictims (41), so the aver-age effectiveness of restraints in a mandatorysystem may be less than that observed withvoluntary experience. *

Subject to these caveats, the evidence is strongthat both air bags and automatic belts are veryeffective devices for reducing death and seriousinjury. Different studies have employed differentdata bases, estimating techniques, and measuresof health outcomes, making direct comparabili-ty difficult.

The most effective restraint system is the air bagcombined with use of a (manual) lap belt. In 1977,NHTSA estimated this combination to be 66 per-cent effective in reducing fatalities, with the bag

‘A related view is that greater protection from equipment maycause drivers to drive with greater abandon, since they feel “safer.”Peltzman (3s) advanced this view in an article in which he claimedthat autoinobile safety regulations had led to a redistribution, ratherthan reduction, in highway deaths, with occupant deaths falling andpedestrian deaths rising. Peltzman’s work has been criticized on bothempirical and theoretical grounds (38).

alone rated as 40 percent effective. * NHTSA es-timated the effectiveness of seatbelts, when worn,at 50 to 60 percent (3). Other estimates of effec-tiveness range from 25 percent (12) to 79 percent(12) for the air bag and from 28 percent (12) to72 percent (13) for seatbelts. However, with boththe lowest and highest figures, biases may be ex-aggerating the estimates. * * From the entirety ofthe studies, a figure in the vicinity of 50 percentseems reasonable for both passive restraint tech-nologies when used properly.

While the effectiveness ratios are close to eachother, air bags rank distinctly higher as life-savingdevices for one simple reason: they are used whenneeded. Disconnect rates for passive belts mightrun 30 to 40 percent, conceivably much higher,and almost certainly would exceed 20 percent(VW Rabbit experience being estimated at 22percent).

Many of the studies share the finding that pas-sive restraint effectiveness decreases with adecrease in injury severity, particularly as onemoves to the most minor accidents. Automaticbelts and air bags are most effective in protectingagainst fatal and life-threatening accidents, lessso but still highly effective for severe but not life-threatening accidents, less effective for moderateinjuries, and least effective for minor injuries. * * *

NHTSA, for example, estimated that belt ef-fectiveness drops 1 to 5 percentage points as onemoves from life-threatening to non-life-threat-ening but severe injuries, another 1 to 5 points

*In studies of air bag deployments, 16 or 17 percent of occupantswere wearing manual lap belts (24).

**For example, Huelke and O’Day (12) examined rural accidentsand relied on a team of medical and other experts to estimate theprobability of occupant survival with different kinds of restraints.Rural accidents tend to be more severe than average, which mayproduce a downward bias in the authors’ estimaws (3).

● **As discussed in ch. 2, automobile accidents are categorizedaccording to the Abbreviated Injury Scale (AIS), which runs fromminor injuries (AIS 1) to fatalities (AIS 6). Attempts to utilize in-jury data in a comparative framework often involve manipulatingAIS codes (e.g., using the root mean square of the AIS). More recent-ly, in an attempt to combine the effects of the severity and numberof injuries, an Injury Severity Score (1SS) was developed. The 1SShas been shown to be highly correlated with the probability of death,length of hospitalization, and extent of disability (24).

Ch. 4–Passive Restraint Systems: The Bag and the Be/t ● 25

moving to moderate injuries, and 20 to 27 pointsto minor injuries (3). For air bags, NHTSA esti-mated more precipitous drops, by 10 to 20 percen-tage points moving to the severe-but-not-life-threatening category, 8 to 13 points to moderate,and 8 to 22 points for minor. The air bag’s effec-tiveness, without a lap belt worn, is rated at closeto zero for minor injuries, since the bag will rarelydeploy.

Mohan, Zador, and O’Neill (24), provide sup-port for this general finding. In their assessmentof the mean Injury Severity Score (ISS) for thetwo restraint systems and unrestrained occupantsby Vehicle Deformation Index (VDI), * Mohan, etal., found that for high VDISs (considerable vehi-cle deformation), both restraint systems per-formed well: lap/shoulder belts resulted in anaverage ISS 55 percent below that of the no-restraint situation, and air bags scored an ISS 66percent below the no-restraint condition. Both ofthese differences were statistically significant.While the small difference between the tworestraint systems was not statistically significant,it was consistent with laboratory test data in-dicating the superiority of air bags in severe front-al crashes.

Mohan and colleagues found that for lowerVDIs (less vehicle deformation), belts continuedto show a considerably lower mean ISS than norestraints; air bags, however, did not. The authorsexplain this, at least in part, as an artifact of theaccident-designation process: an accident wasrecorded as involving an air-bag- equipped vehi-cle only when an air bag deployed; at low VDIs,deployment might indicate a stronger-than-average crash force for the VDI class. Regardlessof the explanation, however, the data are consist-ent with the finding that relative restraint effec-tiveness decreases with decreasing accident severi-ty, and that air bag effectiveness decreases morerapidly —i.e., an occupant is better off wearinga seatbel in a minor accident but perhaps betteroff being in an air-bag-equipped car in a severecrash. And, again, wearing a belt in an air-bag-equipped vehicle is the safest form of restraint.**

● The VDI is exactly what the name suggests-an index of theseverity of damage to the body of a vehicle.

● *A GM study in 1976 concluded that air bags’ effectiveness inreducing injuries ranged from only 6 to 20 percent. An NHTSA cri-

From the relationship between probability ofdeath and IS, Mohan and colleagues estimatedthe expected number of deaths per 1,000 occu-pants in cars in frontal crashes with VDIsof3 to 5. Their estimates, given in percentage termsabove, constitute very high estimates of belt andbag effectiveness. Nevertheless, it seems worthpresenting these striking numbers: for 1,000unrestrained occupants, the estimated deaths total19.4; for 1,000 lap/shoulder-belt-restrained oc-cupants, deaths equal 5.4; and for 1,000 air-bag-restrained occupants (16 percent also wearing alap belt), the figure is 4.0 The authors cautionthat these estimates apply only to full-size and lux-ury cars. I would caution, further, that they areestimates, based on a general correlation betweenan index and an extreme outcome—death.

The different technologies have comparativeadvantages in the specific injury protection theyconfer. Table 8 presents data on the root meansquare Abbreviated Injury Scale (AIS) ratings, bybody region and restraint system, for occupantsin frontal crashes with VDIsf 3 to 5. The datashow that both air bags and lap/shoulder beltsprovide greater protection for all body regions(with one exception) than no restraint.

Air bags are particularly effective in reducinghead and neck injuries (by 58 percent comparedwith no restraint) and have their greatest advan-

Table 8.—Root Mean Square AIS Ratings forOccupants in Frontai Crashes With VDis 3-5 by

Body Region and Restraint System

Restraint system

Body region Air baga Lap/shoulder belt NoneHead and neck . . . . 0.5 0.7 1.2Face . . . . . . . . . . . . . 0.7 0.8 1.2Chest . . . . . . . . . . . . 0,8 0.8 1.4Abdominal or

pelvic region . . . . 0.3 0.8 0.5Extremities and

pelvic girdle . . . . . 1.1 1.2 1.8approximately 16-percent lap belt use.

SOURCE: D. Mohan, at al., “Air Bags and Lap/Shoulder Belts—A Comparisonof Their Effectiveness in Real World, Frontal Crashes” (Washington,D. C.: Insurance Institute for Highway Safety, 1976).

tique of the study identified two potential sources of serious biasthat led NHTSA to conclude that the data supported an effectivenessestimate of from 30 to 60 percent (28). GM has defended the meth-odology of its study [D. Martin, General Motors, personal com-munication, 1982].

26 ● Background paper #l: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

tage over belts in the abdominal and pelvic re-gions, where belts actually perform worse thanno restraint at all (though the injury severity levelis not great). Belts are less effective than bags inall regions except the chest area, where they con-fer equal protection.

The most serious injuries occur to the extremi-ties and pelvic girdle, with both restraint systemsaffording relatively less protection than in mostother areas. While bags dominate belts in thisanalysis, it must be remembered that in the minoraccidents not recorded in this table (VDIs less than3) in which air bags do not deploy, the unbeltedoccupant of an air-bag-equipped car generally re-ceives no additional protection, while the beltedoccupant does.

Most of the estimates presented above havebeen based on actual accident data; some havederived from laboratory tests and theoretical anal-ysis. In most, though not all, attempts have beenmade to correct for sources of bias, differencesin types of drivers, and so on. Nevertheless, itseems useful to close this consideration by pre-senting some unadjusted on-the-road data, num-bers in the simplicity of which lies clarity.

With regard to passive belts, experience withthe VW Rabbit is considerable. From 1975through early 1981, VW produced 400,000 carswith automatic belts satisfying the requirementsof FMVSS 208. These cars have averaged 0.78deaths per 100 million vehicle-miles, comparedwith a national average of 2.4—a 68 percentreduction in the death rate (14). In an earlier com-parison of accident-death experience in Rabbitswith automatic v. manual belts, there were 51 per-cent fewer deaths per 1,000 car-years with theformer than with the latter (28). Since some ofthe automatic belts were disconnected (an esti-mated 22 percent) and many of the manual beltswere being used (about 34 percent), this differenceprovides an underestimate of, or at least a lowerbound on, the life-saving potential of passive beltsystems. *

● Once again, it is important to keep in mind potential differencesby car size, etc. However, the VW data suggest an incremental beltusage of 44 percentage points (78 percent of automatic belt owners

With regard to air bags, almost all of the ex-perience clusters at the other end of the car line:full-size and luxury cars, primarily the 10,000 to12,000 produced by GM between 1974 and 1976.By 1979 these cars had accumulated over 700million miles of driving on American roads. Sta-tistical expectations for deaths and moderate tocritical injuries would have been 10 and 124, re-spectively, yet experience showed half of thesefigures: 5 occupants of air-bag-equipped cars haddied and 62 had received moderate to critical in-juries. There had been some 200 air-bag deploy-ments and only two known deployment failures,one attributable to a mechanic’s mistakenly dis-connecting the mechanism (28).

Finally, while fatality and injury data constitutethe bottom line, it is interesting to note thatpassive restraints may be beginning to passanother acid test–-i.e., their effect on automobileinsurance. Beginning in the late 1970’s, a few in-surance companies, including Nationwide andAllstate, offered owners of passive-restraint-equipped vehicles reductions of approximately 30percent on the portion of car insurance premiumsapplicable to medical (Medpay) or personal in-jury protection (PIP) coverage. Nationwide hasestimated insurance savings of roughly $20 peryear for a car equipped with automatic belts. Overthe lifetime of the car, this translates into a pres-ent value of $150 (33). A Highway Loss Data In-stitute study found reductions in Medpay and PIPclaims of from 20 to 27 percent when comparingVW Rabbits with and without passive belts (13).These market data support the accident data find-ings that passive restraints are an effective meansof reducing death and disability.

not disconnecting their belts minus 34 percent of manual belt ownerswearing their belts). The national discomect rate under a mandatorysystem would have to exceed 45 percent, given manual belt usagein the vicinity of 11 percent, for incremental belt usage to fall belowthat of the Rabbit (5s – ’11 = 44). Some participants in the passiverestraint debate believe that such a disconnect rate would beexceeded.

Ch. 4—Passive Restraint Systems: The Bag and the Belt ● 27

COSTS OF PASSIVE RESTRAINTSEstimates of the costs of passive restraint

systems are abundantly available, but determina-tion of the likely true costs remains a challenge.Most of the estimates come from the automobilecompanies, and according to one nonindustryanalyst, these companies (33):

. . . have a very strong incentive to engage instrategic estimates of costs. If the estimated costsare high enough, they may well persuade[NHTSA] to rescind the rule [FMVSS 208].

For some observers, both leaked companydocuments and analytical work have raised ques-tions about the validity of manufacturers’ esti-mates.

During hearings leading to the l-year delay inimplementing FMVSS 208, Ford and GM pro-vided estimates of incremental automatic belt costwhich averaged $114, including $88 for manufac-turers’ cost, $20 for markup, and $6 for incremen-tal fuel costs (additional fuel consumed as a resultof installation of the technology). Nordhaus (33)has argued that these estimates are inconsistentwith the companies’ estimates of their investmentprograms. He believes that a reasonable estimatefor the true incremental cost to the consumerwould be $60, instead of the $88 cited by GM.Nordhaus’ figure is not inconsistent with earlierNHTSA and GM estimates based on incremen-tal costs in Rabbits and Chevettes, which run (in1981 dollars), from $42 to $85. GM points out,however, that automatic belt systems for largecars would be more expensive than those designedfor small cars. * In terms of annual costs (i.e.,amortizing and depreciating over the life of thecar), the small-car range becomes $8.25 to $15.45(3) (figures updated to 1981 dollars).

The costs of air bags are still less clear.** Atone point, GM claimed that it would sell air bagsfor an incremental cost of from $290 to $325, as-suming mass production. More recently, the com-pany estimated large-volume consumer costs inexcess of $600 per car. Ford’s estimates have also

● Personal communication, 1982. Manufacturers’ estimates in 1978(updated to 1981 dollars) ranged from $84 to $140, depending oncomfort and convenience features (28).

● ● All of the dollar figures in the remainder of this section are up-dated from their original sources to 1981 prices.

varied. In 1976, Ford estimated a consumer costof close to $400, but a recently leaked internalmemorandum placed the figure at from $425 to$1,150. The same source indicated that for bothFord and GM, manufacturer cost would run only$135 to $140 (2). other cost estimates by manufac-turers range from $150 to $280. DOT developedan estimate of $190 (13,28).

Put together, these figures suggest that in-cremental consumer cost for the basic air bagsystems probably would run from $250 to $425.Annual costs would range from $55 to $115. Inaddition, for those air bags deployed in vehicleswhich were not demolished, costs of reinstalla-tion would become relevant. Such costs wouldcertainly exceed initial installation costs underconditions in which air bags would be mass-pro-duced standard equipment. However, such costswould apply to a very minor proportion of vehi-cles, since only a small percentage of air bagswould ever deploy during the lives of the vehicles.

The variation in air bag cost estimates mayrepresent some strategic gaming, as Nordhaus (33)suggests, but it also reflects several technicalchanges and uncertainties. Of fundamental im-portance is the question of scale. Several analysesemphasize the great economies to be realized fromlarge-scale production. Ford estimated a 200-percent difference in per-unit manufacturer costbetween producing a total volume of 885,000 carsand a volume of only 200,000 cars (2). In 1980,GM estimated a consumer cost of $1,100 basedon a volume of 100,000 units, falling to between$650 and $700 if 400,000 units were produced. *

A separate estimate, based on quotations ofcomponents, found that the cost of a driver bagand inflator module would be 25 percent of itsbase cost (which assumed 13,000 units) if 900,000units were purchased (28). Similarly, a passengerbag and inflator module would fall to 33 percentof base cost at the high volume. Some com-ponents, such as sensors and diagnostic parts, arenot so sensitive to the scale, but the fact remainsthat the cost of an air bag system would be verydependent on the number produced and sold (28).

● D. Martin, General Motors, personal communication, 1982.

28 ● Background Paper #l: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

The costs discussed above are direct costs at-tributable to manufacture and installation of thepassive restraint systems, plus small items of in-direct cost (e.g., additional fuel consumption). Ig-nored in these numbers is the potential directmonetary benefit the owner of a passive-restraint-equipped vehicle might derive from lower in-surance premiums. As noted above, the present

discounted value of this benefit over the life ofa typical car will total is $l50 (33), a considerableoffset to the incremental cost incurred by the con-sumer. Indeed, in the case of the automatic belt,the insurance savings outweigh the incrementalprice of the belt, implying that the purchaser ofa passive-belt-equipped car could end up savingat least $35 over the life of the car (33).

5Cost-Benefit Analyses of;

Passive Restraints

5Cost= Benefit Analyses of

Passive Restraints

Should passive restraints be required on all newcars in the United States? Theoretically, theanswer can be determined by a comprehensivecost-benefit analysis (CBA), one which accountsfor all of the social costs and benefits associatedwith implementation of the rule—pecuniary andnonpecuniary, tangible and intangible. In prac-tice, however, few CBAs even approach such acomplete analysis.

Nonquantifiable costs and benefits and thosewhich are quantifiable but not readily valued arecommonly mentioned and then put aside; oftenthey are ignored altogether. In either case, theyare left out of the final calculus of the CBA whicharrives at a “bottom line, ” an economic assess-ment of the worthiness of the program in ques-tion. As such, the “bottom line” is deficient. Itcan and should be recognized as a useful inputinto an overall assessment of the program, butit should not be viewed as the sole determinantof the program’s desirability (20,50). This caveatshould be kept in mind when reading this chapter.

CBA is in its ascendancy as a tool of policyanalysis, particularly with regard to the issue ofgovernmental regulation. Thus, it is not surpris-ing to see that new CBAs entered the debate overFederal Motor Vehicle Safety Standard (FMVSS)208 prior to the October 1981 rescission. Whatmay seem more surprising, however, is that CBAspertaining to the restraint issue date back morethan a decade.

In 1970, Lave and Weber (20) examined thecosts and benefits of seatbelts from the perspec-tive of the individual and found that benefits ex-ceeded costs if the value of an individual’s life wasat least $10,000, a number more than a full orderof magnitude below the smallest estimates of thevalue of life (or livelihood) (50).

Lave and Weber failed to explain why, withbenefits so much larger than costs, the majorityof occupants choose not to wear belts. Thaler and

Rosen (46) addressed this question in an analysisin which they compared the time costs of buck-ling up with the expected benefit. Thaler andRosen estimated an annual benefit from wearinglap belts of approximately $10 (using a value oflife of $200,000), and they argued that the indi-vidual’s opportunity cost of time involved inbuckling and unbuckling the seatbelt could easi-ly exceed this amount.

Recently, Arnould and Grabowski (3) reex-amined Thaler and Rosen’s analysis (46) andundertook two CBAs: 1) the lap/shoulder beltfrom the perspective of the individual, and 2) fromthe point of view of society as a whole on passiverestraint systems.

In the former, Arnould and Grabowski workwith three different weighting schemes to valueindividuals’ willingness to pay to avoid injuriesof varying severity. They conclude that the ex-pected annual benefits from regular belt use, $38to $78 (in 1975 dollars), must exceed the time op-portunity costs associated with buckling up.

Subject to two qualifications which they dismissas insufficient to reverse their conclusion, ArnouIdand Grabowski argue that belt nonuse is not theresult of rational, informed decisionmaking, asThaler and Rosen had suggested it might be. Tothe contrary, these authors interpret their findingsas supporting the “insensitivity-to-low-probabil-ities” hypothesis, discussed in chapter 2. It wasfound, however, that Arnould and Grabowskidismiss too readily the possibility that peoplevalue the freedom from discomfort (physical orpsychological) produced by belts at more than theexpected benefits. Over hundreds of hours of driv-ing per year, the hourly discomfort cost wouldhave to be extraordinarily low to dismiss thisfactor.

Arnould and Grabowski’s passive restraintanalysis is a competent, if standard, CBA of thesocial desirability of a system of mandated passive

31

32 . Background Paper #l: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

restraints, either automatic belts or air bags. Theanalysis finds that both systems would producesubstantial social benefits, with the benefits of airbags slightly exceeding those of automatic belts.However, the costs of the bags are so much higherthan those of the belts that Arnould and Grabow-ski conclude belts are superior on the basis of netbenefit (or benefit-cost ratio). Indeed, they findthat only under very favorable conditions will airbags result in positive net benefits—i.e., the costsof bags quite likely would exceed the benefits.

Arnould and Grabowski’s study is noteworthyfor its sensitivity analysis, which tests the impactof varying assumptions on the benefit-cost con-clusions. The analysis is thorough in its considera-tion of economic costs and benefits, but it mere-ly mentions the costs of inconvenience and dis-comfort associated with belts—costs which mustbe considered potentially large, given all of theevidence on belt use. And, like almost all CBAs,the analysis ignores the costs of the suffering ex-perienced by the loved ones of automobile acci-dent victims. (This is discussed further below.)

Subject to these limitations, Arnould and Gra-bowski’s analysis estimates that in a steady-statesituation, i.e., after passive restraints were in vir-tually all automobiles (commonly estimated to re-quire about 10 years), net benefits of passive beltscould be as high as $8.5 billion and would notbe likely to be less than $3.4 billion. Net benefitsof air bags, by contrast, could reach $6.6 billionbut could also be as low as –$4.9 billion. Theanalysis also calculates expected annual costs perlife saved, which range from $135,000 to $557,000for belts and $472,000 to $2,159,000 for bags. *Obviously, these cost figures attribute no valueto injuries avoided. Lave (19) concurs with thefinding that passive belts would be more cost ef-fective than air bags, though he does not demon-strate analytically the basis of his conclusion.

Another recent CBA, by Nordhaus (32,33), wasundertaken precisely to feed into the Departmentof Transportation’s (DOT’s) reconsideration ofFMVSS 208. Sponsored by five major automobileinsurance companies, this analysis also examinedand compared the two passive restraint systems,

“All of these figures are Arnould and Grabowski’s estimates (3)inflated to 1981 dollars.

although it was focused on passive belts, the sys-tem generally expected to dominate if FMVSS 208had gone into effect.

The analytical slant taken in Nordhaus’ CBAis somewhat different from that in Arnould andGrabowski’s. Nordhaus concentrates on the (net)cost of delaying or reordering implementation ofFMVSS 208, rather than making an “either-or”comparison of the status quo or a fully imple-mented passive restraint rule. As one of DOT’soptions was to rescind FMVSS 208 altogether,Nordhaus’ analysis of the net cost of this optionis directly comparable to other analyses of the netbenefit of fully implementing a passive restraintrule.

Nordhaus estimates that, in a steady state, theannual net cost of a rescission would equal $2.4billion, assuming that all cars are equipped withautomatic belt systems. This number was derivedfrom National Highway Traffic Safety Adminis-tration (NHTSA) and manufacturer data whichNordhaus believes to be in error, biased againstthe belt system; so he views his estimate as a lowerbound on the net benefits of passive belts. Underthese conservative assumptions, the benefits of thepassive belt system ($3.6 billion) are three timesgreater than the costs ($1.2 billion). Nordhausestimates the total discounted net social benefitsof the passive restraint rule at $33 billion. Alter-natively, $33 billion represents the net cost to soci-ety of a complete rescission of the rule. Under theassumptions that he believes to be more reason-able, this figure rises to $69 billion. Nordhaussummarizes his findings as follows:

[T]he passive restraint rule is, from an econom-ic point of view, as important as any environ-mental, health, or safety rule on the books. If theestimates of the impact on fatalities are accurate,a rescission would be equivalent to repealing alaw that cuts in half the homicide rate. It is equiv-alent to forgoing the medical advances that al-lowed the virtual elimination of death from tuber-culosis over the last quarter century.

Nordhaus’ CBA shares with Arnould and Gra-bowski’s an effective use of sensitivity analysis.Unlike Arnould and Grabowski, Nordhaus findsthat a world of air bags would be preferable toa world of automatic belts. Despite the high costestimate he uses to evaluate bags ($425), Nord-

.

Ch. 5—Cost-Benefit Analyses of Passive Restraints ● 33

haus concludes that an all-bag system would gen-erate total discounted net social benefits of $47billion. * He does not dwell on the comparison,however, since at the time of his study belts ap-peared to be the wave of a 208 future.

As part of his policy analysis, Nordhaus exam-ines the impact on the automobile manufacturersof implementing FMVSS 208. The only major or-ganized opposition to implementation of the rulethrough the years, the manufacturers have ex-pressed concern about the costs of adding passiverestraints as standard equipment and what thiswould do to the demand for their product. Thepotential problem is of particular concern in 1982,with the domestic industry in a depressed condi-tion. Nordhaus presents a case that adverse ef-fects would be minimal. He suggests that it is evenconceivable that the industry would benefit fromimplementation of FMVSS 208. This would oc-cur if consumers recognized the net economic sav-ings involved in buying a passive-restraint-equipped car, given an associated reduction inautomobile insurance costs.

Nordhaus’ is not the first CBA undertaken di-rectly in connection with governmental evalua-tion of a passive restraint rule. In 1974, NHTSAreleased a CBA that demonstrated “the superiorityof passive restraint systems compared to belt sys-tems presently required” (13). The analysis wasrevised in response to criticisms and still came upwith the same conclusion (13). Two years later,another CBA accompanied the announcement ofa public hearing to be held by DOT.

At about the same time, Robertson (39) directeda survey which, though not itself a CBA, pro-duced a finding of direct relevance to CBA: a sam-ple of new car buyers expressed a willingness topay an average of $12 more per month ($144 peryear) in car payments to save 6,000 lives per yearand $17 per month ($204 per year) to save 12,000lives. ** It appears that many new car buyersmight stand prepared to pay considerably more

● Recall that under the assumption he finds more realistic, Nord-haus found a greater net benefit for the belts. However, he doesnot apply a set of “more realistic” assumptions in the air bag case,relying instead on manufacturers’ estimates of cost. Thus, this $47billion figure should be compared with the $33 billion estimate forthe belts.

**The dollar figures have not been adjusted for inflation, so theyunderstate current value.

than the amount passive belt systems would re-quire. Whether answers to a hypothetical ques-tion would translate into equivalent action in themarketplace remains to be seen.

Graham, Henrion, and Morgan (9) have iden-tified half a dozen other CBAs on the occupantrestraint issue and have produced a detailed anal-ysis of their own, one which compares passive re-straint systems with other methods of encourag-ing restraint. Their analysis ranks FMVSS 208 be-low other alternatives in terms of both net bene-fits and benefit-cost ratio, with a compulsory beltusage law having the highest ratio (in large partbecause its measurable costs are so low) and acombined air-bag/mandatory-belt usage law pro-ducing the greatest net benefits (and saving themost lives). All of the alternatives these authorsexamine produce positive net benefits—i.e., eachalternative is preferable to the complete absenceof occupant restraints (and superior to the cur-rent system of merely requiring belts in cars). Inparticular, they estimate that the benefits ofFMVSS 208 would have exceeded the costs by 95percent.

Thus, CBAs have served as inputs, the impor-tance of which is difficult to assess, throughoutthe long debate on a Federal passive restraint rule.Each of the analyses differs from the others in cer-tain important ways: some adopt a human capitalapproach to valuing life (or livelihood), whileothers use willingness-to-pay (50); basic datasources, and hence magnitudes, often vary signif-icantly; restraint alternatives studied differ fromone analysis to the next; some analyses incorpo-rate concerns like the effect of restraint systemson insurance costs, while others ignore them, andso on.

Almost all of the studies can be faulted for theirfailure to treat analytically the inconvenience anddiscomfort costs which, though nonpecuniary,seem to play a significant role in many people’sdecisions about using manual belts. It maybe dif-ficult or unreasonable to place a dollar value onsuch costs directly. However, there are sensitiv-ity analysis techniques that would permit an eval-uation of the potential significance of these costs.For example, one might employ break-even analy-sis to determine how highly people would haveto value the inconvenience and discomfort in

34 ● Background Paper #l: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

order to make passive belts appear to be sociallyundesirable (so).

Despite their idiosyncrasies and their individualand collective flaws, as a body the passive re-straint CBAs present an impressive case that soci-ety would benefit more than it would lose froma compulsory passive restraint rule. The findingsare reasonably consistent and robust. In general,the studies rank air bags ahead of passive beltsas life-saving devices, in large part reflecting theability (and desire) of many passive-belt vehicleowners to disconnect their belts. With somenotable exceptions, the analyses rank belts higherthan bags on cost-effectiveness and cost-benefitgrounds, primarily reflecting the much lower costof the passive belt option. While all of the studiesfind passive belts to produce positive net benefits,several note that the positive net benefit of air bagsis sensitive to assumptions.

Despite the uniformity of these studies’ findings,the analyses have articulate critics who challengebasic assumptions of the models. In particular,while industry critics argue that belt cost estimatesin the analyses are too low, most of the criticismis focused on estimates of passive seatbelt use. Theautomobile industry seems convinced that incre-mental belt-usage rates would be extremely low,assuming passive belts that could be easily discon-nected (a condition that they believe would bedemanded by the public). Thus, the industryviews passive belts as increasing vehicle costswithout significantly increasing effective passengerrestraint. Current (manual) belt users would there-by be “punished” by an unnecessary additionalcharge, while confirmed nonusers would have tobear the same additional burden but would realizeno additional protection (23).

In closing this glimpse at occupant restraintCBAs, several caveats should be mentioned. The

first, illustrated by the work of Graham, Henrion,and Morgan (9), is that a finding that a manda-tory passive restraint rule would be cost-beneficialdoes not necessarily mean that it would be themost cost-beneficial approach to saving livesthrough occupant restraint. Other alternativesshould be compared in order to seek the approachthat would maximize net social benefits.

Use of the phrase “net social benefits” suggestsan important aspect, and limitation, of using CBAin a policy framework: as was noted at the outsetof this discussion, reasonable costs and benefitsare not the only, nor necessarily the most impor-tant, variables in policy decisionmaking. Above,studies were faulted for their failure to valuediscomfort and inconvenience; but other unmeas-ured variables may be of much greater conse-quence. For example, CBAs generally do not ade-quately address the issue of who benefits and wholoses—not everyone realizes a net gain from im-plementation of a passive restraint rule-and itis this distinction that has made a (slow) horserace out of what appears to be a socially desirableobjective (22). The distributional issue—winnersand losers—constitutes one of the themes of thenext and concluding chapters.

Finally, it should be noted that none of theCBAs attempts to directly value avoidance of thepain and suffering of accident victims and theirloved ones. * As a result, the CBA's findings ofpositive net economic benefit support the noneco-nomic desire to minimize human suffering.

● This is considered indirectly in willingness-to-pay valuations,though these generally cover only the victims themselves and notthe suffering of people close to them. Furthermore, willingness-to-pay estimates suffer from people’s inability to fully comprehend andcontemplate their own deaths or serious disabilities (50).

6

Philosophical and Ethical Issue;in the Passive Restraint Debate

Philosophical and Ethical Issuesin the Passive Restraint Debate

Historically, the principal argument againstFederal Motor Vehicle Safety Standard 208 hashad a common two-pronged antiregulation struc-ture: 1) mandating passive restraints representsone more example of governmental intrusion intoindividuals’ rights to make their own decisions andrun their own lives; it is a flagrant example ofpaternalism; and 2) mandatory passive restraintswill increase the cost of car ownership, * un-necessarily raising expenses for people who cur-rently use their manual belts and people whoprefer not to be restrained (many of whom mightdisconnect automatic belts).

The “individual freedom” argument is not with-out merit, nor is it without very close precedent.The debate over delay or rescission of the passiverestraint rule echoed the debate which has beenrepeated in recent years in one State capitol tothe next as a majority of States have repealed theirmotorcycle helmet laws (51). Nevertheless, severalfactors mitigate the force of the freedom theme.These include economic and equity considera-tions, public opinion, and precedent.

Regarding the last of these, there are numerousprecedents for Government’s mandating auto safe-ty features. Some involve driving behavior, suchas speed limits and drunk-driving laws; and many,like the passive restraint rule, involve physical at-tributes of cars themselves—to name a few, shat-terproof glass, energy-absorbing steering-wheelcolumns, padded dashboards, and today’s manuallap/shoulder belts. Precedent does not make thenext related case—i.e., passive restraints—neces-sarily “right,” but it does serve as a supportingargument.

Public opinion is another consideration—in ademocracy, ultimately perhaps the most impor-tant one. While survey results concerning passive

*As was noted in the preceding chapter, if insurance rates for pas-sive-restraint-equipped cars fall, the net cost of owning a car withan automatic belt might decrease, even though the showroom stickerprice would rise.

restraints have varied, it is decidedly not the casethat the public opposes passive restraints. For ex-ample, a 1977 Gallup poll found that of the 83percent of respondents who had an opinion, a ma-jority favored installation of air bags on all cars.This theoretical support for air bags does not dis-appear when respondents are reminded that theywill have to pay extra to have this extra protec-tion (13). * The will of the majority should notinvariably dominate the rights of the minority,but in this case, given the limited input of publicopinion into the decision, at least there is no ques-tion of the tyranny of the majority.

Conceptually, the most compelling responsesto the “individual freedom” view have a theo-retical base grounded in economic theory. Perhapsthe most important of these are the negative ex-ternalities associated with the absence of adequateoccupant restraint. Loosely, negative externalitiesare negative consequences visited upon one ormore persons as a result of some other person’sindependent decisions or actions. In the occupantrestraint case, negative externalities are of twotypes: economic and noneconomic.

Concerning the former, the nature of automo-bile and health insurance is such that the insuredcosts of automobile accidents are spread over largenumbers of car and health insurance policyholders(and taxpayers in the instance of public healthinsurance). Thus, if an unrestrained victim of anautomobile accident experiences greater inju~and hence higher medical bills than would havebeen the case had he or she been restrained, othersshare in the economic liability created by themedical treatment. In effect, one person’s inde-pendent decision not to wear a seatbelt has ad-

● Once again, the question is whether (or how much of) suchsurvey support would translate into market decisions. The auto-mobile manufacturers seem convinced that in practice consumerswill not be willing to pay the high cost air bags would entail (23).It should be noted, however, that the industry’s limited experiencein this regard involves a radically different set of circumstances thanwould obtain in a world of mandated air bags.

37

38 ● Background Paper#1:Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

verse consequences for other people’s pocket-books. Pecuniary negative externalities such asthis constitute a technical market failure and serveas a theoretical rationale for governmental inter-vention. *

The noneconomic negative externalities concernthe health risks imposed on people other than theinitial car buyer, and refer specifically to air bags,not passive belts. If the initial buyer purchases acar equipped with an air bag, all future frontseatoccupants-passengers, children, second and thirdowners-will receive that protection. If the initialbuyer does not purchase an air-bag-equipped ve-hicle, future passengers will be subjected to greaterrisk. This argument depends further on some spe-cific economic conditions, which will be returnedto, but it is sobering to note the following statisticsfrom a study of 137 crashes (with 172 people in-jured) in Baltimore County, Md. (4):

approximately half of the people injured wereoccupants of vehicles no longer owned bytheir original purchasers;nearly 60 percent did not own the vehiclesin which they were injured;three-quarters of the passengers were not re-lated to the vehicle owners; andalmost two-thirds of the victims were underthe age of 30, and only one-third of 16- to29-year-olds injured owned the vehicles inwhich they were injured.

In effect, these data challenge, or at least ques-tion, the argument that new-car purchasers shouldbe free to determine whether to invest in safetydevices like air bags since it is their own protec-tion that is in question. The health and safety ofmany others are also involved. In a well-function-ing market, one might find this potential externa-lity acceptable. After all, people can and shouldchoose not to ride in cars in which they do notfeel adequately protected. And, more to the point,future owners of used cars can have desired safe-

● Some of these pecuniary externalities could be overcome throughthe market. For example, insurers could charge lower (actuariallyfair) personal injury protection (PIP) and medical (Medpay) pre-miums for owners of passive-restraint-equipped vehicles, as a cou-ple of insurers are currently doing. However, other pecuniary neg-ative externalties could not be addressed by the market. Consider,for example, the accident-related medical costs of Medicare andMedicaid beneficiaries.

ty equipment added when they purchase theircars. But can they, in the case of air bags?

There may be firms that will retrofit a car tohave air bag protection, but the cost is quite highbecause of the small scale of the business and thesizable economies of scale that pertain to air bagmanufacture and installation. Thus, the prospec-tive buyer of an air bag system is penalized eco-nomically by the small scale of active demand forair bags. Alternatively, all of those people whowould want to purchase air bags at mass-pro-duced prices would benefit from a requirement(or industry decision) to install bags in much ofthe new-car fleet. *

Clearly, there is one obvious group of poten-tial economic victims of a passive restraint law:current regular users of manual lap/shoulderbelts. They are already receiving all of the pro-tection that would be afforded them under a man-datory passive restraint rule, yet under such a rulethey would have to pay more for their new cars.In effect, they would be penalized for other driv-ers’ poor judgment. If Nordhaus’ (33) assessmentis correct, the net cost of automobile ownershipwould drop under a passive restraint rule, bene-fiting current belt users as well as nonusers. Butone might argue that current users should alreadybe receiving an insurance break, something in-surers do not offer—in large part presumably be-cause of the problem of verification of belt use(see ch. 3).

Thus, the “individual freedom” argument haspluses and minuses. Certainly there is a public in-terest in passive restraints which competes withthe private interest in free individual choice; thisis not clearly a case of unwarranted governmen-tal intrusion, pure and simple. Which of the in-terests one rates as dominant probably reflectsone’s basic political philosophy as much as theinherent merits of the case.** In this regard, it is

● A ready response to this problem is that people can use thealready required lap/shoulder belts and receive nearly the sameamount of protection. However, the above argument applies to theapparently sizable segment of the population that chooses not towear belts, perhaps for reasons of discomfort, but would like topurchase air-bag protection at a reasonable cost, though not at cur-rent small-scale high market prices.

● *For a recent debate on this theme, see Perkins (36) and Bakerand Teret (5).

Ch. 6–Philosophical and Ethical Issues in the Passive Restraint Debate ● 39

interesting to read the assessment of George Will consenting adults to behave in ways disastrous(52), noted conservative political analyst, com- to themselves. Besides, too many children pas-menting on the view “that Government has no sengers are sacrificed on that altar. And a largebusiness requiring drivers to buy and use inex- part of the bill for the irrationality of individualpensive devices that might save them from self- drivers is paid by society.

destruction:”

There is a pitiless abstractness, and disrespectfor life, in such dogmatic respect for the right of

7.Conclusion

7Conclusion

There is a basic philosophical issue in the debateon passive restraints —one which has risen toprominence in an antigovernment, antiregulationera. Nevertheless, it is important to recognize thatthe principal deterrent to implementation of a pas-sive restraint rule has been opposition on the partof automobile manufacturers—grounded, it ap-pears, in concern about the economic implicationsof passive restraints for the industry, and not theissue of individual freedom. * Nordhaus (33) hasconcluded that the ramifications probably wouldnot be as dire as industry statements and docu-ments seem to suggest, at least with regard to thedemand for automobiles and manufacturer anddealer revenues. However, there is a long historyof manufacturer opposition, with repeated pro-posals for the voluntary phasing in of passive re-straints generally going unrealized (13,17).

The impact of higher car prices on the demandfor automobiles—the issue addressed by Nord-haus—is only one of the manufacturers’ concerns.Manufacturers have also expressed concern over

*This point is made quite clearly in General Motors’ (GM’s sub-mission to the Department of Transportation for the 1981 hearingson passive restraint options (23):

We believe that [this issue of mandatory passive restraints] hingessquarely on what the American public is willing to accept in terms ofeconomic or behavioral burden. Inflatable restraints are too costly tobe acceptable to most . . . automatic belts that are coercive will be re-jected by the public. Noncoercive (separate) automatic belts will notsignificantly increase belt use over that achievable with manual belts,and thus will not be cost effective.

Because we believe that the rule will not work, we recommend thatit be rescinded .

Should the Government determine that additional regulation is nec-essary, a mandatory belt use law will be more effective, sooner, andat a fraction of the cost of a passive restraint rule.

The mandatory belt use law clearly violates the individual freedomperspective at least as much as would passive restraints. Thus, GM’sconcerns seem firmly grounded in pragmatic economic considera-tions.

On that pragmatic level, GM’s points are worthy of serious at-tention. GM is arguing, in essence, that the CBAs overestimate theamount of incremental belt usage that would be achieved under apassive restraint rule, and that the low levels GM anticipates wouldflip the cost-effectiveness balance over. In this regard, note thatArnould and Grabowski (3) found that an incremental belt usagerate of only about 20 percentage points is all that would be neededto make passive belts cost effective. Of course, while they mightbe cost effective at that level, their effectiveness would be small com-pared with the injury problem and the potential of belts to reduce it.

the effect of passive restraints on product liabil-ity claims, worrying “that ‘endless lawsuits’ wouldallege the failure of automatic restraints to pro-vide adequate protection” (28). Furthermore, giv-en a climate of “exaggerated public expectations”about the effectiveness of passive restraints, theyhave anticipated lawsuits as a result of an occu-pant injury even when restraints worked as theywere supposed to. Thus, the availability and costof liability insurance, manufacturers have feared,could loom as serious problems.

But some evidence suggests just the opposite.For a variety of reasons, insurance might be abun-dantly available. And there is reason to believethat passive restraints might decrease, rather thanincrease, the number and size of liability claims.For one thing, automatic restraints should de-crease the number of deaths and serious injuriesresulting from crashes attributable to manufactur-ing defects and design problems such as stallingengines, malfunctioning brakes, and tire blowouts(28).

In considering the support for and oppositionto mandatory passive restraints, one should rec-ognize that most of the organized interests fall intothe support camp. Insurers have expressed theirsupport frequently and strongly. The AmericanMutual Insurance Alliance, a group representingover 95 percent of all of the automobile insurancewritten in the United States, has gone on recordas unequivocally favoring mandatory passive re-straints (13). And, as discussed above, a fewmajor insurance companies have already put theircorporate mouths where their money is, offeringpremium reductions to owners of passive-re-straint-equipped vehicles.

There is an irony, however, in the likely out-come of implementation of a passive restraint rule:a major but unorganized interest group, automo-bile consumers, has expressed an apparent pref-erence for air bags over passive belts; yet the Na-tional Highway Traffic Safety Administration(NHTSA) has estimated that if passive restraintswere required, 99 percent of the new-car fleet

43

—.

44 . Background Paper #l: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

would have passive belts, with only 1 percent ofcars coming equipped with air bags. The manufac-turers must have concluded that whatever con-sumers’ abstract preferences might be, in buyingpractice most consumers would not choose to paythe extra cost required for air bags (23).

There is further irony here, one that has a dis-tinct Catch-22 flavor: air bags are clearly effec-tive in significantly reducing motor vehicle injuriesand fatalities—yet because the automakers areconvinced that consumers would not be willingto pay the high cost for the air bags, they plannedto produce primarily passive belts had FederalMotor Vehicle Safety Standard (FMVSS) 208 re-mained in effect; yet because they believed thatmost passive belts would be disconnected, theyargued forcefully, and successfully, that passivebelts were not cost effective, and hence thatFMVSS 208 should be rescinded.

Mandatory passive restraints would have rep-resented only one addition to a lengthy list oftechnological safety features required on auto-mobiles by the Federal Government. But passiverestraints are of special interest and importancefor two reasons: 1) because of the long delays andoften acrimonious debate over implementation ofFMVSS 208, and now over its rescission; and2) due to the realistic potential of these technolo-gies to make an extraordinary dent in the deathtoll of automobile accidents. Estimates vary, butthe consensus suggests that passive restraintscould prevent a minimum of 6,000 and perhapsas many as 12,000 highway deaths a year. Giventhe current total of roughly 27,000 frontseat oc-cupant deaths a year, a savings of this magnitudewould constitute a truly major public healthvictory.

Less dramatic than lives saved, but more nu-merous, would be injuries avoided. Here the es-timates also vary widely, depending in large parton the severity of the injuries considered. Arnouldand Grabowski (3) estimate that from 20,000 to40,000 severe injuries (Abbreviated Injury Scale3 to 5) would be prevented by passive restraintseach year; Nordhaus (33) estimates that 120,000moderate to critical injuries would be avoided.However they are grouped and counted, injuryreductions would be impressive.

With the demise of the passive restraint rule,and over a decade of working toward its realiza-tion, NHTSA has a special obligation to seek al-ternative strategies to reduce the motor vehicleaccident toll. The chore remains the same as be-fore—to find a cost-effective, politically accept-able means of providing effective vehicle occu-pant restraint. The air bag represents a technologythat is effective and that would apparently be ac-ceptable to the public, were it not for its great cost.At the other extreme lies the mandatory belt-uselaw—an approach that might be quite effectiveand inexpensive, yet appears to be politicallyunacceptable. In between lies a myriad of alter-natives that are probably more acceptable and lesseffective (e.g., a renewed public information cam-paign). Recent analyses suggest that a mix of ap-proaches should be explored.

From the point of view of OTA’s study Tech-nology and Handicapped People, it is unfortunatethat automobile accident data do not permit acareful assessment of the number and severity ofhandicaps resulting from automobile accidents.Data on disabilities resulting from accidents arenot particularly good, and disabilities do not in-variably become permanent (or longstanding)handicaps. Nevertheless, despite these data defi-ciencies, simply considering the number andnature of serious motor vehicle accident injuriessuggests the truly extraordinary potential of oc-cupant restraints to reduce disabilities and hand-icaps.

In the context of a case study of the technolog-ical prevention of handicaps, the importance ofthis reduction is twofold. From a social/fiscalpoint of view, the resource savings attributableto this prevention technology appear to be sub-stantial. The costs of technologies for handicappedpeople, both individual and collective, are con-siderable, as a glimpse at the main OTA reportimmediately suggests. The prevention of handi-caps requiring use of these technologies impliesa savings of economic resources by reducing theneed for the technologies. * Thus, any policy

● This is a benefit of the prevention of technology that should beincluded in a cost-benefit analysis (CBA). The CBAs on passive re-straints consider many such resource savings-e. g., medical expensesavoided—but undoubtedly miss the reduction in the need for manyof the collective technologies for the handicapped. To this extent,the passive restraint CBAs understate the net benefits of mandatingpassive restraints.

Ch. 7—Conclusion ● 45

deliberations on resource allocation to address theproblems of the handicapped should give seriousconsideration to prevention efforts as viablemeans of reducing the total societal costs ofhandicaps.

More to the point, the true costs of handicapsvastly exceed those that can be readily measuredin dollars and cents. The physical and psycholog-ical suffering that accompany handicaps are costs

for which no technology can ever fully compen-sate. But some technologies, like motor vehiclepassenger restraints, can prevent this suffering.The full value of this benefit, included in noneof the passive restraint CBAs, must be enormous(1). Thus, regardless of its purely fiscal implica-tions, the prevention approach warrants especiallycareful attention.

References

References

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

Abt, C., “The Social Costs of Cancer,” Social In-&ators 2:175, 1975.“Air Bag Cost Put at $100 in Memo, ” Ann ArborNews, Aug. 13, 1981, p. A-1.Arnould, R., and Grabowski, H., “Auto SafetyRegulation: An Analysis of Market Failure,” Bellj. Econ. 12:27, 1981.Baker, S., “Who Bought the Cars in Which Peo-ple Are Injured? An Exploratory Study,” At. 1.Pub. Health 69:76, 1979.Baker, S., and Teret, S., “Freedom and Protection:A Balancing of Interests,” Am. ~. Pub. Health71:295, 1981.Carsten, O., “Use of the Nationwide PersonalTransportation Study To Calculate Exposure,”The HSRI Research Review II 11(6):1, 1981.Comptroller General, General Accounting Office,“Effects, Benefits, and Costs of Federal SafetyStandards for the Protection of Car Occupants, ”report to the Committee on Commerce, U.S. Sen-ate, Washington, D. C., 1976.Fleisher, G., An Experiment in the Use of Broad-cast Media in Highway Safety: Systematic Anal-ysis of the Effect of Mass Media Communicationsin Highway Safety (Los Angeles: University ofSouthern California, Department of Industrial andSystems Engineering, 1972).Graham, J., Henrion, M., and Morgan, M., “AnAnalysis of Federal Policy Toward AutomobileSafety Belts and Air Bags” (Pittsburgh: Carnegie-Mellon University, Department of Engineering andPublic Policy and School of Urban and Public Af-kaks, 1981).Haddon, W., submission of documents to W.Coleman, Jr., Secretary, Department of Transpor-tation, Washington, D. C., Sept. 17, 1976.Hartunian, N., Smart, C., and Thompson, M.,“The Incidence and Economic Costs of Cancer,Motor Vehicle Injuries, Coronary Heart Disease,and Stroke: A Comparative Analysis, ” Am. J.Pub. Health 70:1249, 1980.Huelke, D., and O’Day, J., “Passive Restraints: AScientist’s View, “ in The scientific Basis of Healthand Safety Regulations, R. Crandall and L. Lave(eds.) (Washington, D. C.: Brookings Institution,1981).Insurance Institute for Highway Safety, Back-ground Manual on the Occupant Restraint Issue(Washington, D. C.: IIHS, 1978).Insurance Institute for Highway Safety, “Commit-tee Leader Backs Automatic Restraints,” H@hwayLoss Reduction Status Report 16(7):1, 1981.

15.

16.

17.

18.

19.

20.

21.

22,

23,

24,

25,

26,

27,

28,

29.

Insurance Institute for Highway Safety, “Promo-tional Campaigns To Increase Seat Belt Use: A Re-view” (Washington, D. C.: IIHS, July 1981).Joksch, H., and Wuerdemann, H., “Estimating theEffects of Crash Phase Injury Countermeasures—11: The Fatality Trend and Its Modification byCountermeasures,” Accident Analysis & Preven-tion 5:1, 1973.Kelley, B., “GM and the Air Bag: A Decade of De-lay,” Bus. & Society Rev. (35):54, 1980.Kuechenmeister, T., and Yanik, A., “Summary ofMarket Opinion Research Study: An Analysis ofthe Factors Affecting Seat Belt Use” (Detroit, Mich:Market Opinion Research, 1978).Lave, L., The Strategy of Social Regulation: lkci-sion Frameworks for Policy (Washington, D. C.:Brookings Institution, 1981).Lave, L., and Weber, W., “A Benefit-Cost Anal-ysis of Auto Safety Features, ” Applied Eon.2:265, 1970.Levine, D., and Campbell, B., “Effectiveness ofLap Seat Belts and the Energy-Absorbing SteeringSystem in the Reduction of Injuries” (Chapel Hill,N. C.: University of North Carolina Highway Safe-ty Research Center, 1971).Luft, H., “Benefit-Cost Analysis and Public PolicyImplementation: From Normative to PositiveAnalysis, “ Pub. pol. 24:437, 1976.Martin, D., “Comments of General Motors Cor-poration Regarding Passive Restraints (FMVSS208),” docket 74-14; notice 22, Department ofTransportation, Washington, D. C., 1981.Mohan, D., Zador, P., O’Neill, B., et al., “AirBags and Lap/Shoulder Belts—A Comparison ofTheir Effectiveness in Real World, FrontalCrashes” (Washington, D. C.: Insurance Institutefor Highway Safety, 1976).Muller, A., “Evaluation of the Costs and Benefitsof Motorcycle Helmet Laws, ” Am. J, Pub. Health70:586, 1980.Myers, M., Iscoe, C., Jennings, C., et al., StaffReport on the Cigarette Advertising Investigation(Washington, D. C.: Federal Trade Commission,1981).National Highway Traffic Safety Administration,H&hway Sa&ty Facts 3 (Washington, D. C.: De-partment of Transportation, 1978).National Highway Traffic Safety Administration,Occupant Protection Program Progress ReportNo. 2 (Washington, D. C.: Department of Trans-portation, April 1979).National Highway Traffic Safety Administration,

49

— .

N ● Background paper #l: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40,

41<

Report on Traffic Accidents and Injuries for 1979(Washington, D. C.: Department of Transporta-tion, 1981).National Highway Traffic Safety Administration,Task Force Report on Saidy Belt Usage Laws(revised) (Washington, D. C.: Department ofTransportation, 1978).National Safety Council, Accidimt Facts (Chicago:NSC, 1980).Nordhaus, W. “Comments of William Nordhauson Notice of Proposed Rulemaking on FederalMotor Vehicle Safety Standards: Occupant CrashProtection, ” Department of Transportation,Washington, D. C., March 1981.Nordhaus, W. “Comments of William Nordhauson Notice of Proposed Rulemaking on FederalMotor Vehicle Safety Standards: Occupant CrashProtection,” docket No. 74-14, notice 22, Wash-ington, D. C., Department of Transportation, May26, 1981.Partyka, S., “Effects of Traffic Accident Injurieson the Work Force Estimated From NCSS andNASS Data” (Washington, D. C.: National Centerfor Statistics and Analysis, NHTSA, June 1981).Peltzman, S., “The Effects of Automobile SafetyRegulation,” J. PoZit Eon. 83:677, 1975.Perkins, R., “Perspective on the Public Good,”Am. J. Pub. Health 71:294, 1981.Robertson, L., “Automobile Safety Regulationsand Death Reductions in the United States, ” Am.J. Pub. Health 71:818, 1981.Robertson, L., “Car Crashes: Perceived Vulnera-bility and Willingness to Pay for Crash Protec-tion,” J. Community Health 3:136, 1977.Robertson, L., “A Critical Analysis of Peltzman’s‘The Effects of Automobile Safety Regulation’, ” J.Ekon. Issues 3:587, 1977.Robertson, L., “Estimates of Motor Vehicle SeatBelt Effectiveness and Use: Implications for Occu-pant Crash Protection,” Am. J. Pub. Health66:859, 1976.Robertson, L., “Public Perception and Behavior

42.

4 3 .

4 4 .

4 5 .

46,

4 7 .

4 8 .

4 9 .

50.

51,

52.

in Relation to Vehicle Passenger Restraints, ” pre-sented at the International Workshop on the Anal-ysis of Actual v. Perceived Risk, National Acad-emy of Sciences, Washington, D. C., June 1-3,1981.Robertson, L., “State and Federal New-Car Safe-ty Regulation: Effects on Fatality Rates,” AccidentAnalysis & Prevention 9:151, 1977.Robertson, L., Kelly, A. B., O’Neill, B., “A Con-trolled Study of the Effect of Television Messageson Safety Belt Use, ” Am. J. Pub. Health 64:1071,1974.Slovic, P., Fischoff, B., and Liechtenstein, S., “Ac-cident Probabilities and Seat Belt Usage: A Psy-chological Perspective” (Eugene, Oreg.: DecisionResearch, 1977’).Teknekron Research Inc., “1979 Survey of PublicPerceptions on Highway Safety” (Washington,D. C.: Department of Transportation, 1979).Thaler, R., and Rosen, S., “The Value of Savinga Life: Evidence From the Labor Market, ” inHousehold Production and Consumption, Studiesin Income and Wealth No. 40, N. Terleckyj (cd. )(New York: National Bureau of EconomyResearch, 197s).Tofany, V., “Life Is Best at SS,” Traffic Q. 35:5,1981.Warner, K., “Clearing the Airwaves: The CigaretteAd Ban Revisited,” Po]. Analy. 5:43s, 1979.Warner, K., “Cigarette Smoking in the 1970’s: TheImpact of the Antismoking Campaign on Con-sumption,” Science 211:729, 1981.Warner, K., and Luce, B., Cost-Benefit and Cost-Efkctiveness Analysis in Health Care: Principles,Practice, and Potential (Ann Arbor, Mich.: HealthAdministration Press, 1982).Watson, G., Zador, P., and Wilks, A., “TheRepeal of Helmet Use Laws and Increased Motor-cyclist Mortality in the United States, 1975-1978,”Am. J. Pub. Health 70:579, 1980.Will, G., “Driving Without Restraint,” The Wash-ington Post, Apr. 14, 1977.

U.S. GQVHWMENI PRINI’ING OFFICE : 1982 0 - 98-607 : QL 3