8217

8/14/2019 8217

http://slidepdf.com/reader/full/8217 1/47

Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

September 1982

NTIS order #PB83-159822

8/14/2019 8217


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 m aterial is usu ally not of immed iate 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

8/14/2019 8217


Library of Congress Catalog Card Number 82-600546

For sale by the Sup erintend ent of Docum ents,U.S. Government Printing Office, Washington, D.C. 20402

8/14/2019 8217


Foreword

Technology exerts a p owerful 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 hand icapp ed persons is oneof the prime social and economic goals of public policy.

The Federal Government is deeply involved in p rograms that affect the develop-men t and use of techn ologies for disabilities. Congress and other institutions h ave beenincreasingly interested in how well programs that directly develop technologies and sup-port th eir u se have been performing. The Senate Committee on Labor and Hu man Re-sources requested the Office of Technology Assessment (OTA) to condu ct a stud y of technologies 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 pu blic policy. The main rep ort of the study Technology and Handicapped People was released in May 1982.

This case is backgrou nd pap er #1 of the stud y. A nu mber of case stud ies will bepu blished as part of the assessment, and each will be issued separately. The case stud ieswere commissioned by OTA both to provide information on specific technologies andto gain lessons that could be app lied to the broad er policy aspects of techn ology anddisability.

Drafts of each case stud y w ere reviewed by OTA staff; by members of the ad visorypanel to the overall assessment, chaired by Dr. Daisy Tagliacozzo; by members of theHealth Program Adv isory Committee, chaired by Dr. Frederick Robbins; and by nu -merou s other experts in m edicine, disability p olicy, Governm ent, economics, publicinterest and consumer rights, and rehabilitation engineering. We are grateful for theirassistance. However, respon sibility for the case stud ies remains with th e auth ors.

#

..

JOHN H. GIBBONS

Iii

8/14/2019 8217


Advisory Panel on Technology and Handicapped People

Daisy Tagliacozzo, Panel Chair Department of Sociology, University of Massachusetts, Harbor Campus

Miriam K. BazelonWashington, D.C.

Tom BeauchampKennedy Institute-Center for BioethicsGeorgetown University

Monroe Berkowitz Bureau of Economic Research Rut gers University

Henrik BlumUniversity of California, Berkeley

Frank BoweW oodmere, N. Y.

Jim Gallagher Martha Porter Graham Cent er University of N orth Carolina, Chapel Hill

Melvin GlasserCommittee for National Health Insurance

Ralf HotchkissOakland Calif.

John KimberlyThe Wharton SchoolUniversity of Pennsylvania

Robert Leopold Department of Psychiatry Hospital of the University of Pennsylvania

LeRoy Levitt Mount Sinai Hospital

A. Malachi Mixon, III Invacare Corp.

Jacquelin Perry Rancho Los Amigos Hospital

Barbara W. Sklar Mount Zion Hospital

William StasonVeterans Adm inistration and Harvard School of

Public Health

Gregg VanderheidenTrace Research a nd Dev elopment Cent er University of W isconsin

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

Disabled Persons

8/14/2019 8217


OTA Staff for Background Paper #1

H. David Banta, Assistant Director, OTA Health 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 Cw alina, 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

8/14/2019 8217


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

2, AUTOM OBILE SAFETY PROBLEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

H ealth Con sequ ences of Accid ents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7Record of Fed era l Safety Stand ards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8Manual Seatbelts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3. ALTERNATIVES TO MAN DATORY PASSIVE RESTRAIN T SYSTEMS . . . . . . . . . . 15

Do Nothing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Information/ Education Campaign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Mandatory Seatbelt-Use Laws . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16“Techn ologica l F ix’’.... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17Econom ic Incentiv es . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .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 . . . . . . . . . . . . . . . . . . . . . . . 316. PHILOSOPHICAL AND ETHICAL ISSUES IN THE PASSIVE

RESTRAIN T 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

Distr ibution of Injuries by Injury Sever ity . . . . . . . . . . . . . . . . . . . . .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 Occupan ts in Frontal Crashesby Body Region and Restra int System . . . . . . . . . . . . . . . . . . . . . . . .

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

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

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

Page7 7 8

910

10

21

2 5

vii

8/14/2019 8217


1 .

8/14/2019 8217


1

Introduction;

The OTA stud y Techn ology and Handicapped

People concentrated on an examination of technologies developed for and used by peoplewith handicaps. An alternative, or complemen-tary focus, wou ld be on technologies d esigned toprevent h and icaps. There are literally hu nd redsof prevention technologies that constitute familiaraspects of 20th century American life. Many pre-vention technologies have derived from scientificresearch; the vaccines that prevent cripp ling child-hood diseases are among these. Others representa more p rosaic app lication of comm on sense toeveryday problems; an obvious example is theubiquitous rubber bathtub mat. Some prevention

technologies require the installation of elaboratesafety equipment, such as that found in a nuclearpower plant, others require only the installationof a simple idea in peop le’s mind s—e.g., the ad-vice not to drink and drive.

Regardless of their diverse technical character-istics, all preven tion technologies can be view edin a common persp ective vis-a-vis the class of technologies designed to treat, restore, rehabili-tate, and palliate hand icapp ed ind ividu als. Pre-vention technologies complement these “after-the-fact” technologies in the b attle to ameliorate the

consequences of han dicaps. But p revention tech-nologies also compete with treatment technol-ogies, for a prim ary goal of the former is to ob-viate the n eed for th e latter. Thu s, physicians canexhort their patients not to smoke, or years laterthey can attemp t 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, w hat w ill pr evention cost? In th e coldestof analytical persp ectives, the pu rely econom iccosts of preventing handicaps can be compared

*NOTE: The writing of this background p aper on p assive re-straints in automob iles was completed in M ay 1982. As the pap erwent to p ress in Septemb er 1982, the situation regarding passiverestraints had b een further altered by court decisions but h ad notyet been finally resolved.

with the p urely economic costs of dealing w ith

preventable handicaps after they are realized. Butthe cha llenge of social resou rce allocation d eci-sionmaking calls for a more complex and some-what “warmer” cost-benefit calculus, one whichblend s the economic and hu manitarian concerns.In the effort to minimize the ad verse consequencesof handicaps, analysts must develop measures (orat least concepts) of social cost which incorporateboth p ecun iary and n onp ecun iary costs, wh ichadd th e costs of suffering to the costs of materials.

The purp ose of this backgrou nd pap er 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 introdu cing the p revention p erspectivethrou gh a case stud y of a single prevention tech-nology. It is hop ed that th is will enrich p olicy-mak ers’ deliberations, imp licit or explicit, on th esocial costs and benefits of alternative strategiesfor dealing w ith the problems of handicaps.

This case study examines issues in the debateon w hether p assive 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 am end ed, d ecreed th at all n ewcars would h ave to have a passive restraint systemcapable of meeting a 30-mph crash performancerequirement by September 1, 1983 (1984 modelyear), with p hase-in beginning with th e largest1982 model cars by September 1, 1981.

On April 9, 1981, the National Highway Traf-fic Safety Administration (NHTSA) announceda d elay of 1 year in imp lementation of FMVSS208, and new hearings w ere held in Aug ust 1981

to consider w hether the (delayed) rule should bepu t into effect or one of three alternatives shouldbe adop ted. Two of the alternatives involved areordering of implementation d ates for the varioussizes of cars; the third involved elimination of thepassive restraint requirement.

3

8/14/2019 8217


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 d ecision have been filed, andcongressional action with a similar intent has beenthreatened . Thus, though th e passive restraint re-quirement h as been eliminated by ad ministrative

fiat, the issue is not en tirely d ead.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 m ajor political issue of th e early 1980’s—gov-ernmental regulation.

The preceding paragraph should su ggest themultidimensional significance of the outcome of the Federal rulemaking, bu t the raison d ’etre of passive 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 of them frontseat occupants of vehicles) and inflictdisabling injuries (ranging from temporary andminor to permanent and serious) on an add itional2 million p eople. Automob ile acciden ts are th eleading cause of death and accident-producedhandicaps among young people. Some analystshave estimated that passive restraint systems couldpreven t up to half of the d eaths of frontseat oc-cupa nts an d over 100,000 moderate to critical in-

ju ries each year.The remaining chapters of this pap er explore

the issues and evidence in the passive restraintsystem debate. Chapter 2 considers the nature andextent of the autom obile safety problem, examin-ing accident, death, and injury data an d review-ing the record of automobile safety standards anddev ices, includ ing th e curren t (“active”) seatbeltsystem. Chapter 3 offers a glimpse at app roachesto impr oving the u se of passenger restraints bymeans other than mandating passive restraints.Attention then tu rns in chapter 4 to a description

of the two passive restraint systems—air bags andautom atic belts—and a presentation of data ontheir estimated costs, safety, and effectiveness inreducing deaths and disabilities. The fifth chapterreviews and compares cost-benefit analyses of these systems. Chapter 6 identifies and discussesthe ph ilosophical and ethical issues related to man-

dating passive restraint systems. Concludingthoughts are p resented in chapter 7.

In closing this introduction, it seems worthw hileto reflect momentarily on a peripheral but impor-tant theme—20th century technology and its im-

pact on health (a theme which is well illustratedby th e case of the autom obile). Certainly, overall,technological developments in this century haveenhan ced th e quality of life and add ed 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-du ced health hazard s ind irectly (for examp le, theaffluence bred of modern industrialization has im-proved our diets and thereby rendered us moreresistant to d isease). But 20th century technologyhas not invariably reduced health hazards. Rather,the history of new technology is one of continualredefinition of the typ es and sources of risks tohealth, with new hazard s replacing old, and thegains generally exceeding the losses: the overalltrend in d isability and m ortality has been d own-ward.

The automobile serves as a prime example of the comp lexity of mod ern t echnology’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 wou ld occur ever more rap idly and in-expensively. Accomp anying these health ben efitsof motor vehicles, however, have been the signifi-cant health costs of street and highway travel, andthe deaths and injuries which reflect the size, struc-tur e, and velocity of the vehicles, as well as char-acteristics of the roads and of the operators of thevehicles. The d isprop ortionate imp act of motorvehicle accidents on the young is particularlytragic, as thou sand s of lives are cut sh ort in th eir

prime and healthy bodies are committed to d ec-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.

8/14/2019 8217


2Automobile Safety Problem

8/14/2019 8217


2m

Automobile 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 grou p, 28,900, were occup ants of passen-ger cars; 6,700 were in trucks, and 3,700 were onmotorcycles. An ad ditional 2 m illion peop le re-ceived disabling injuries. All told, there were 18.1million accidents involving 29.7 million vehicles,the large majority of which resulted primarily inprop erty dam age or nondisabling injuries. Onein tw elve vehicles registered in the U nited Stateswas involved in an acciden t, and a similar ratio

characterized the fraction of the p opu lation in-volved in an accident. Additional consequencesincluded the following:

q 3.5 million h ospital bed-days beyond initialemergency care;

q 35,700 person -years of work effort lost; andq an estimated $ 3 5 . 8 b i l l i o n i n e c o n o m i c

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 are

under the age of 25, an age group constituting justunder 23 percent of licensed drivers. Table 1show s the most trag ic consequence of this ph e-nom enon: the m otor vehicle death ra te for 15- to24-year-olds is tw ice the n ational average an d fivetimes that of youn ger children . Table 2 dem on-strates further that there is a strongly unequal sexdistribu tion of motor vehicle fatalities, with maledr ivers’ 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 N ational Highw ay Traffic Safety Adm inistration (29). Itshould be n oted that many accident-related data are estimates andthat nu merous d ata inconsistencies are found in the literature on motor vehicle accidents.

**This figu re includes lost wages, medical expenses, insu rance ad-ministration costs, and property d amage. It does not include police,fire, and court expen ses, the value of lost cargo on comm ercial ve-hicles, etc. (31). A recent in dep end ent stu dy estim ated th e costs of motor vehicle deaths and injuries at $20 billion (11).

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

Age Number of fatalities Fatality rate a

<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 rate a

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 12.0 55 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 of 20) 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 th atattributable to polio at its w orst (13). *

Unfortun ately for the pu rposes of this stud y,there are no good national data on the types and

q Age-specific motor veh icle death rates are in part a fun ctionof 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 passen ger-miles than w omen . Asa result, fatality rates per 100 million passenger-miles by age differfrom the sim ple age-specific fatality rates. The youn g males’ ratesremain the most socially alarming, but th e highest rates per 100million passenger-miles belong to the very elderly (over 75) (6).

7

8/14/2019 8217


8 q 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 d istribution of injuries in a recent NHTSA sample by injuryseverity. The d ata ind icate that over 70 percentof all injuries are scored as minor on the Ab-breviated Injur y Scale (AIS), a common index of injury severity. Under 12 percent of injuries arein categories (AIS 3 to 5) in which survival is prob-able or possible and in w hich serious hand icapscould be a result. Given the total num ber 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 su ggest the substan -tial probable burden.

Despite the “bad news” contained in the abovemotor vehicle accident dat a, 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 grow thin 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 surviving1 . . . . . . . . . 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 quitedr amatic. For examp le, the d eath ra te 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 r ate w as in th e 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-men ts, includ ing the national 55-mph speed limitimplemented in 1974, the increasing price of energy in the m id to late 1970’s, and a utom obilesafety features (47). We now tur n to a look at t herecord of Federal governmental regulation of automobile safety.

RECORD OF FEDERAL SAFETY STANDARDSAs substan tial as the acciden t toll is today, there

is evidence that it wou ld hav e been considera blygreater 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 stand ard s.

Robertson observed a total death rate of 5.5person s p er 100 million v ehicle-miles for cars n otsubject to safety regu lations, but a rate of only3.4 for cars meeting the Feder al safety stand ard s.A differential characterized all classes of victims

(includ ing ped estrians, motorcyclists, and ped al-cyclists), but occupants of automobiles meetingthe standards realized the greatest benefit, witha d eath r ate of 1.5 per 100 million vehicle-milescompared with 2.9 for occupants of vehicles notsubject to the stand ard s.

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

q 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 n onfatal injuries go han d-in-hand.

8/14/2019 8217


8/14/2019 8217


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

MANUAL SEATBELTSFront-lap seatbelts were first installed in all cars

as stand ard equip ment 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 th e United States.Thus, almost all cars on the road today areequipped w ith 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 p ublicity slogan pu t it years ago.

Yet, according to recent observations of manualbelt use, only 11 percent of drivers w ear 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 m id-1970’s ma y havebeen attributable in part to the ignition-interlock systems that w ere installed on 1974 and som e 1975model cars. The interlock systems provoked aloud an d an gry pu blic response as drivers foundthemselves una ble to start their cars wh en theyplaced cargo on the passenger seat (e.g., groceriesor the family dog) that exceeded th e weight min-imum which activated the system. The congres-sional response was to p rohibit the Departm entof Transportation from requiring the system on

later cars (41).Table 5 shows how belt usage varied in 1977-78

by 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 ind icates howbelt usage varied by sex, region of the country,and car size.

The result of such low usage rates is that ma n-ual belts in cars cur rently red uce the fatality andserious injury rates by less than 10 percent. Thus,a tremend ous p otential for saving lives and pre-

venting injuries is going unrealized, despite theread y accessibility of the technology an d 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

—2.4 4.2

14.4 12.7 12.2 12.2 12.5

8.9 10.9 16.9 17.1

14.4 13.6 13.3 13.7

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 invar iably cited in p olls arediscomfort and inconvenience (28), although asurvey sp onsored by Gen eral 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 d iscomfort r eflects belts’pressuring or abrading hips, chests, and necks andcreating an un pleasant sense of confinement orrestriction of movement. For some people, incon-venience refers simply to the minor effort involved

in bu ckling a w ell-fun ctioning b elt system, whilefor others it relates to d ifficulties in r etracting thebelts or latching or releasing th em.

8/14/2019 8217


Ch. 2—Automobile Safety Problem q 11

In a stud y of its own , NHTSA concluded thatmany of the complaints about belts were wellfounded. A representative sample of Detroit-areadrivers identified mod erate or serious p roblemswith comfort or conven ience in a ll 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 au tomatic belt systems, NH TSA 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 w ear a belt presum ably re-flects a jud gment th at the d isutility 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 th e high potential cost of not protecting themselves and indeed m ay notcomprehen d th e meaning of the tiny p robabili-ty; instead, they respond primarily to the unlike-lihood of the event.

Arnou ld and Grabowski discuss a stud y (44)

which illustrates the principle: two groups of ex-perimental subjects were given d ata on th e 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 of a fatal accident and a l-in-3 chance of at least onedisabling injury), while the other group receivedthe same informa tion calculated on a p er-trip basis(where the odds of the accident outcomes areminuscule). Compared with the latter group , thegrou p given the lifetime figures respond ed by in-dicating a much greater increase in expected seat-belt usage and a greater disposition tow ard seat-belt laws.

Arnould and Grabowski also present evidencethat p eople significantly u nd erestimate their risk of involvement in an automobile accident. For ex-ample, a recent survey (45) queried: How likely

do you th ink 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 answ er equal to or greater than theactual societywide av erage, roughly 1 in 10. Amajority selected od ds of 1 in 100 or sm aller 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 th e low p robability of a serious accident,and th us, 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 th e probability of an accident.

Arnould an d Grabowski estimate the annu alper-person benefits of buckling up at between $38and $78 (in 1975 dollars) and state tha t, at p revail-ing wage rates (the assumed opportu nity cost of time), this amou nt m ust exceed the time costs of buckling up . Hence, they conclud e that a rationalweighting of costs and benefits cannot explain thefailur e of so many p eople to wear their seatbelts.The authors acknow ledge that “there may be sig-nificant d iscomfort costs to some ind ividuals towearing seatbelts. ” Nevertheless, they suggest, “itwou ld seem hard to argu e that [this] wou ld sochange the . . . benefit-cost calculus to explain the80- to 90-percent cu rrent n onu tilization rate of seatbelts. ”

While it is agreed th at d iscomfort costs couldnot explain th e entirety of nonu tilization, it couldbe argued that they might explain much of it, par-ticularly if “discomfort” is defined to include th epsychological discomfort of those w ho fear beingtrapped in their cars by belts. Certainly, manyriders who experience discomfort from seatbeltswou ld accum ulate hund reds of hours of discom-fort if forced to wear them, and one would notneed to va lue discomfort time highly to conclud ethat nonuse was rational behavior for these indi-viduals.

Numerous characteristics differentiate beltwearers from nonusers (41). The latter tend tohave less edu cation th an th e former, rate belts asmore uncomfortable and inconvenient, and aremor e likely to be smokers. The victims of serious

8/14/2019 8217


—

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

crashes are less likely to wear b elts than p eople risk-averse, or perhaps more risk-loving, thannot involved in serious accidents, especially in the their belted counterparts.case of youthful d rivers and drivers und er the in-fluen ce of alcohol. Fur therm ore, unbelted d rivers Whatever the explanation, the fact remains thattend to follow the cars in front of them closer than only a small minority of autom obile occup antsdo belted drivers. Collectively, all of these char- choose to wear seatbelts. The consequence in

acteristics suggest that unbelted drivers are less hu man d estruction is tragic.

8/14/2019 8217


3 .

Alternatives to Mandatory PassiveRestraint Systems

8/14/2019 8217


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 p assive restraints were the best way to ad-dress th e failure of the vast ma jority of Americansto w ear their (man ual) seatbelts. But passive sys-tems are not the only op tion; other alternativeshave been suggested and some have been tried.

Advocacy for the passive restraint alternativereflects in part the experienced or pred icted 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 of passive 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 d eci-

00 NOTHING

This alternative has few proponents in the po-litical arena. Lacking some bone tossed in thedirection of prom oting passenger restraint, the “do

2.

3.

4.

5.

sion of restraint to the individual automobileoccupant, and d o nothing to promote the useof manual belts. *Actively promote the use of manual beltsthrough m edia camp aigns and other edu ca-tional efforts. This is one of th e options theReagan administration has favored.Pass legislation requiring the wearing of seatbelts.Require a “technological fix” that forcesman ual belt u se (e.g., the p reviously triedignition-interlock system).

Offer economic incentives to automobile oc-cupan ts to wear their seatbelts.

q A still more laissez-faire option would be to eliminate the re-quiremen t that cars have manual b elt systems as standard equ ip-ment, leaving purchase and installation of belts as a buyer option.There are sound theoretical reasons to oppose th is possibility (dis-cussed in ch. 6 below), in addition to the political uproar it wou ldcreate.

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 p romoteuse of pa ssenger restraints. Provision of informa-tion—education—is the conceptually appropriate

approach if one perceives the problem of beltnon use as resulting from rid ers’ 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).

Borrow ing from related health ed ucation ex-perience, one might be tempted to conclude that

major publicity efforts could increase the volun-tary use of seatbelts. For example, there isevidence that the antismoking campaign of thepast nearly two d ecades has had a substantial im-pact on cigarette smoking (49), and specifically,

15

8/14/2019 8217


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

that the major broadcast media antismoking cam-pa ign of 1968-70 significant ly red uced cigaretteconsump tion (48). Unfortunately, how ever, 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 un likely 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 camp aignswere m oun ted 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 p rod uce increases in u se(8,43). The American experience is echoed by ex-perience in Canad a, Great Britain, and France:major publicity campaigns either did not increasebelt use at a ll 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 mor e sophisticated u nd erstand ing of the fac-tors influencing seatbelt usage w ill perm it develop-ment of more effective edu cation an d pu blicitypackages. They also argue t hat 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.

How ever, even the m ost optimistic assessmen tsuggests that an “all-out” campaign could notboost usage rates to greater th an 40 percent (13).Thus, if one’s objective is to red uce pr eventablemotor vehicle fatalities by as m uch as p ossible,this alternative clearly cannot be relied on byitself. *

MANDATORY SEATBELT-USE LAWS

q The difficulty of promoting truly informed decisionmaking—

the objective of this alternative—is more sub stantial than one mightthink at first. As Arnould and Grabowski (3) demonstrate, people’sunderstanding of accident risks is poor, and th e 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 w ell

over a dozen countries. In Australia, NewZealand, Canada, Sweden, and West Germany,mandatory belt laws have been in effect forseveral years and p relaw and postlaw belt-usageda ta are av ailable. The first such law , in the Stateof Victoria, Au stralia, pr odu ced belt-use rates of 70 to 80 percent an d account ed for red uctions infatality r ates of 20 percent in urban crashes and10 percent in rur al crashes. In the Provinces of Ontario 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 p assage 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 d rivers. 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 presu mably w ould in-crease the level and intensity of opp osition. Asa representative of the Consu mer Federation of America stated (13):

All too often State legislators, faithfully reflecting

the sentiment of the p eople in the State, see mandatorybelt legislation as an outrageous and u nn ecessary in-trusion in to people’s lives. Whether one agrees withthat view or not, it is a sentiment that is prevalent inthe large majority of our States.

8/14/2019 8217


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

This mentality is reflected in the fact that overhalf of the States have recently rep ealed laws r e-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 mov ing auto-mobiles. The law s often include exemptions or ex-ceptions (e.g., permitting babies to be unre-strained when nursing). Observational studiesconclud e that belt-usage rates by children ha veincreased, thou gh th ey 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 sup erior to, and m uch more pu bliclypalatable than, the ignition-interlock systemfoun d in 1974 (and some 1975) mod el 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 INCENTIVESThe rema ining alternative-offering econom ic

incentives to wear manual belts—has not beentried directly, though a few insurers do offerpremium d iscounts of up to 30 percent on medicalor p ersonal injur y p rotection 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 manu al belts, but th e problemof verifying comp liance is not a minor one. Sim-ilarly, insurers could offer p ositive or negativebenefit incentives. For examp le, medical paymentscould be increased for belt-wearing accid ent vic-tims (an approach adopted by at least one in-surer), or payments could be conditioned on beltuse at th e 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 mand atory law m ight resultin u sage rates exceeding th ose achievable withpassive seatbelts, because so many passive belts

wou ld be detached. Nevertheless, in today’s po-litical climate in the United States, mandatoryseatbelt-use law s seem u nrealistic. It also shouldbe noted th at this assessment of belt laws has u t-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 b e developedwhich would make permanent, or even tem-porary, disconnecting difficult or inconvenient,and this might have a marginal impact on passivebelt use. But g iven th e pu blic’s clear oppositionto evident and burdensome “technological fixes, ”manufacturers and the Government seem u nlikelyto go this route.

If the comp liance problem could be resolved(mechanical solutions are conceivable), the poten-tial for the premium approach is intriguing. Fora car with au tomatic belts, Nation wid e Insu rancehas estimated premium savings of roughly $20 peryear. Over th e lifetime of a car, this translates intoa present discounted value of $150 (33). Sincemany automatic belts are disconnected, * assur-

q A survey found a disconnect rate of 22 percent among ownersof VW Rabbits equipped with automatic belts. The findin g thatowners of Rabbits with m anual b elts tend to use their belts more

frequently than th e average automobile owner suggests that the auto-matic belt disconnect rate under a mandatory passive restraint lawmight b e greater than 25 percent. The rate could be considerablygreater. In particular, the automatic-belt Rabbit s have had an inter-lock system, so disconnection requi res action after the car has started–perhaps a more active form of passive-belt rejection (23),

8/14/2019 8217


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

ance of compliance with manual belts should bewor th at least as mu ch as the initial presence of autom atic belts. Thus, the lifetime insurance valueof buckling up might prove to be a significantincentive.

Even if the insuran ce incentive “worked ,” it

seems unlikely that it would raise effectiverestraint rates high enough to satisfy people whowant to see maximum redu ction of the highwaydeath and disability toll. Aside from paternalism,an argument grounded in the existence of negativeexternalities su ggests that self-selected comp liancerates w ill be too low from a social point of view(see ch. 6). Furthermore, the economic incentiveapproach would take time to achieve widespreadeffect—compliance techn ology wou ld h ave 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 m ore attention than it has received to d ate.To my knowledge, it is not now, nor has it been,an option u nd er serious policy consideration.

This chap ter’s review of th e policy alternatives

should su ggest one of the reasons many peopleconcerned with automotive safety have been sosup portive 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 though t that ad vocates a mix of severalof these alternatives as a cost-effective means of achieving effective passenger restraint (9,18).Given the recent demise of FMVSS 208, it seemsprobable that more attention will be directedtoward a multiple-approach strategy.

8/14/2019 8217


4Passive Restraint Systems;

The Bag and the Belt

8/14/2019 8217


.

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 p roperly,both passive restraint systems are qu ite effectivein reducing the risk of death and serious injury.

Techn ically, the au tomatic belt system is quitesimple. Most systems in operation consist of asingle shoulder belt (a “two-point system”) that

crosses the rider au tomatically up on entry intothe car, and a padd ed knee panel below the dash-board to p rovide add ed p rotection. Automaticbelts with a lap as w ell as a shoulder compon ent(a “three-point system”) have also been designed.Some systems offer autom atic shou lder belts andmanual lap belts. Close to half a million autom aticbelt systems are currently on American roads,almost all of them in VW Rabbits and ChevroletChevettes, offered as par t of option p ackages.

Air bags consist of deflated bags situated in thesteering wheel (driver) and glove compartment

area (passenger) sides w hich inflate virtu ally in-stantaneously wh en front end or dashboard sen-sors detect crash forces substantial enough to beharmful. Different systems have different infla-tion mechanisms, and all includ e 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 pr event rid ers fromsliding u nd er bags, a readiness monitor, and anindicator light) (28). Altogether, over 10,000 air-bag-equipp ed cars have accumulated over a bil-lion m iles on Am erican road s since they first ap-pea red in 1972.

This more complex technology h as a historywhich dates back 30 years. In 1952, the first of several patents for automatically inflating aircushions was filed. Federal Government interest

dates from 1968, when prototype development be-came su fficiently advanced to consider large-scaleapp lication in the n ear future. In 1969, the Na-tional High wa y Safety Bur eau (N HSB, pred eces-sor of the N ational Highw ay Traffic Safety Ad-ministration, NHTSA) announced a proposedru lemaking for an “Inflatable Occup ant RestraintSystem,” with an initially proposed effective dateof January 1, 1972. This marked the beginningof a longstanding adversarial relationship betweenautomobile manufacturers and the Department of Transportation (DOT) wh ich saw dozen s of de-bates over th e effectiveness and cost of passiverestraints and d elays in imp lementation of, andchanges in, a passive restraint requiremen t.

Delay and change have occurred during eachsuccessive adm inistration, with th e Reagan ad -ministration’s re-examination of the issue beingthe latest and most rad ical. The courts, Congress,public interest groups, and the med ia have all beenactive participan ts in the d ram a. The conflict hasmade strange bedfellows of such diverse interestsas Ralph N ad er and the Pacific Legal Found ation(PLF), the latter a p ublic interest group advocating“limited government.” In 1977, PLF filed suit in

the U.S. Court of Ap peals to block the p assiverestraint rule, claiming that DOT had an “insuf-ficient basis for the air bag d ecision.” The follow-ing year, Nader and Public Citizen, the consumerrights group , filed su it in the Court of App eals

21

8/14/2019 8217


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

seeking a ruling that DOT’s scheduled 3-yearphase-in of restraints w as illegal; plaintiffs wan tedall new vehicles to have passive restraints at thesame time. The Court of Appeals consolidated thesuits of PLF and N ader an d Public Citizen. *

Part of the regulatory debate picture has been

repeated proposals by automakers to introduceair bags or automatic belts on their own. Eachof these proposals has come in response to a pr o-posed passive restraint rulemaking by the Govern-ment. As each of the p roposed ru lemakings wasaltered or delayed, often because of agency ac-tions, occasionally because of judicial decisions,the automakers’ plans were themselves altered,invariably in the d irection of limiting introd uc-tion of passive restraints.

In 1970, for example, General Motors (GM) in-formed NH SB that it would provide air bags onall its cars by 1975, introd ucing the equ ipmen t asan option and then converting it to standardequip ment. In 1973, GM informed DOT that itwas red ucing its planned prod uction of air-bag-equipped 1974-75 cars from 1 million to 150,000un its, blaming both tooling d ifficulties and theGovernmen t’s stand ard-setting p rocess. Half ayear later, a GM spokesperson told the Insur anceInstitute for H ighway 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,

q For a detailed chronology of events in air bag developm ent,debate, and rulemak ing, see Back,ground Manual) on the Occupant

Restraint Issue (13).

claiming insufficient consumer demand, a claimwhich GM officials have acknowledged was basedin part on a failure of the company to p romotethe technology. GM’s revising its p lans and de-parting from the air-bag-equipped automobilemarket followed an earlier court d ecision over-turn ing a prop osed N HSB 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 orrescind ed (17). Recently, GM an nou nced th e ter-mination of its inflatable restraint p rogram , call-ing the device economically infeasible (23).

Three years ago, the “Big Three” dom estic auto-mobile manufacturers and several foreign pro-du cers reported to NH TSA that they were w ork-ing on belt and bag systems and intend ed to in-troduce them, as options, on several modelswith in the n ext few m odel years (28). Even pr iorto 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, w hen used prop erly, 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-indu ced car price in-creases. This too is examined below.

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

Coleman, Secretary of Transp ortation in th e Ford moderate to critical injuries per year” (13).administration, concluded that passive restraints“are a reliable and effective means of substan tially The precise quantitative findings can be chal-redu cing death an d injuries on the Nation’s high- lenged-for examp le, recent estimates have p lacedwa ys.” If air bags w ere installed on all cars, Cole- the life-saving potential of passive restraints in theman estimated, they “w ould p robably save over vicinity of 6,000 to 9,000-but the qualitative con-twelve thousand lives annually and prevent or elusion is beyond disp ute. All of the quan titative

8/14/2019 8217


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

evidence sup por ts it and th e logic is imp eccable:“The principle behind an occupant restraint is thatan occup ant is mu ch less likely to be injured orkilled in an autom obile crash if the crash forcesare app lied in a controlled w ay to the strongestparts of the human body” (28). Furthermore, pro-ponents of passive restraints argue, occupant re-straints are mu ch 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 system s are not p erfect; theyredu ce but d o not invariably prevent d eaths andinjuries. Neither belts nor air bags are particularlyeffective in rear-end collisions, and the twotechnologies have d ifferent m erits and liabilitiesin other types of crashes. In the instance of an im-pact on th e passenger side of a car, for examp le,belts are effective in keeping the d river from beingpitched to the side. Air bags occasionally inflateon side imp act, and the passenger-side bag canthen protect the driver, but inflation of the bagsis uncertain. If the bags d o inflate, they sh ield thedr iver from flying glass and m etal and therebyafford the d river 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 compou ndsused in man y air bag systems to inflate the bags.Testimony h as been offered ind icating that th e

chemicals are confined sufficiently so that car oc-cupan ts are not exposed to th em, either before orduring deployment of a bag. NHTSA is convincedby the evidence, but is quick to point out thatother, nonchemical inflation mechanisms h avebeen developed (13).

Other concerns expressed about bags includefears that inadequately restrained children couldslide beneath an d be smoth ered by the bags, oreven conceivably be thrown backwards into therear window. There are worries that air bagsmight inflate spontaneously without an impact or

on m inor imp act, thereby causing a serious acci-den t (12). Again, the bu lk of the eviden ce is thatthese are not significant problems. The track record of th e more th an 10,000 air-bag-equipp edcars 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 th ey can be disconnected. They have to be

equipped with a safety release for emergencies.For the person for whom belt-wearing is trulybur den some, passive belts can be actively disen-gaged either permanently or repeatedly. In effect,the individual converts the passive restraint sys-tem into an active nonr estraint system, and ef-fectiveness goes to zero.

A common worry expressed about belts (man-ual or autom atic) is that they may p revent 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 occur

in rare instances, but the opposite outcome ismu ch more p robable—i.e., an u nrestrained oc-cupant is much more likely to be injured or killedby being thrown from (or around) the car thanby being restrained w ithin it.

In the remaind er of this section, the evidenceon th e effectiveness of belts and bags in red ucingdeath an d injury is presented an d discussed. Thefollowing caveats should be kept in m ind in in-terpreting the data:

1.

2.

Air-bag field experience relates virtually ex-clusively to large-size and luxur y cars, prin-

cipally the GM vehicles produced in the1974-76 mod el years. Whether th is experi-ence generalizes directly to all cars (withalmost all new cars much smaller than these)remains to be seen.

Autom atic belt experience in the VW Rab-bit, the source of most relevant data, mightnot typify general experience once all carswere so equ ipp ed. Small-car bu yers in thelate 1970’s exhibited a greater p rop ensity tobuckle up than d id the average driver of theperiod: whereas average drivers buckled up

only 10 to 15 percent of the time, own ers of Rabbits w ith m anu al (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

8/14/2019 8217


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

in the general population greater than thatof autom atic-belt Rabbit own ers, particular-ly since DOT policy could not requ ire thekind of interlock device foun d in au tomatic-belt Rabbits (23).

3. Related to the above point, current users of passive restraint systems represent a largelyself-selected g rou p. It is possible that th eirmotivation, particularly their concern aboutdr iving safety, differs enough from that of the overall popu lation 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 par t 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 p assive belts(3). Nevertheless, it should be recognizedthat victims of serious crashes are less oftenrestrained th an n onvictims (41), so the aver-age effectiveness of restraints in a mandatorysystem may be less than that observed withvoluntary experience. *

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

The most effective restraint system is the air bagcombined w ith use of a (manu al) lap belt. In 1977,NHTSA estimated this combination to be 66 per-cent effective in red ucing fatalities, with th e bag

‘A related view is that greater protection from equ ipment m aycause 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 redistribu tion, ratherthan reduction, in highway deaths, with occupant d eaths falling andpedestrian deaths rising. Peltzman’s work h as been criticized on b othempirical 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 estim ates 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. How ever, with boththe lowest an d h ighest figur es, biases may be ex-aggerating the estimates. * * From the entirety of the studies, a figure in the vicinity of 50 percentseems reasonable for both p assive restraint tech-nologies when used properly.

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

Many of the stud ies share the find ing that p as-sive restraint effectiveness decreases with adecrease in injury severity, particularly as onemoves to the most minor accidents. Automaticbelts and air bags are m ost effective in p rotectingagainst fatal and life-threatening accidents, lessso but still highly effective for severe but not life-threaten ing accidents, less effective for m odera teinjuries, and least effective for minor injuries. * * *

NH TSA, for example, estimated that belt ef-fectiveness dr ops 1 to 5 percentage points as onemoves from life-threatening to non-life-threat-ening bu t severe injuries, another 1 to 5 point s

*In stud ies of air bag dep loyments, 16 or 17 percent of occupantswere wearing manu al lap b elts (24).

**For example, Huelke and O’Day (12) examined rural accidentsand relied on a team of med ical and other experts to estimate theprobability of occupant survival with different kinds of restraints.Rural accidents tend to be m ore severe than average, which m ayproduce a downward bias in the authors’ es t imaws (3).

q **As discussed in ch. 2, automob ile acciden ts are categorizedaccording to the Abbreviated Injury Scale (AIS),which runs from

minor injuries (AIS 1) to fatalities (AIS 6). Attempts to utilize injury data in a comparative framework often in volve 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 numb erof injuries, an Injury Severity Score (1SS) was developed. The 1S S

has been shown to b e highly correlated with the probability of death,length of hosp italization, and extent of disab ility (24).

8/14/2019 8217


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

moving to moderate injuries, and 20 to 27 pointsto m inor in jur ies (3). For air bag s, NH TSA esti-mated more precipitous drops, by 10 to 20 percen-tage points moving to the severe-but-not-life-threaten ing category, 8 to 13 points to m oder ate,and 8 to 22 points for minor. The air bag’s effec-tiveness, without a lap belt worn, is rated at close

to zero for minor injuries, since the bag w ill 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 occupan tsby Vehicle Deforma tion Index (VDI), * Moh an, etal., found that for h igh VDISs (considera ble 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 66

percent below t he no-restraint cond ition. Both of these 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 occupan t is better off wearinga seatbel in a minor accident but perhaps betteroff being in an air-bag-equipp ed car in a severecrash. And, again, wearing a belt in an air-bag-

equip ped vehicle is the safest form of restraint.**q The VDI is exactly wh at the nam e suggests-an ind ex of the

severity of damage to the body of a veh icle.q *A GM study in 1976 concluded that air bags’ effectiveness in

reducing injuries ranged from only 6 to 20 percent. An NHTSA cri-

From the relationship between probability of death and IS, Mohan an d colleagues estimatedthe expected n um ber of deaths per 1,000 occu-pants in cars in frontal crashes with VDIsof 3 to 5. Their estimates, given in p ercentage termsabove, constitute very h igh estimates of belt andbag effectiveness. Nevertheless, it seems worth

presenting these striking numbers: for 1,000unrestrained occupan ts, the estimated deaths total19.4; for 1,000 lap/ shou lder-belt-restrain ed 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 ap ply only to full-size and lux-ury cars. I would caution, furt her, that they are estimates, based on a general correlation betweenan index and an extreme outcome—death.

The different t echnologies have comparativeadvantages in the specific injury protection they

confer. Table 8 presents data on th e root meansquar e Abbr eviated 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 n o restraint.

Air bags are particularly effective in reducinghead and neck injuries (by 58 percent comparedwith n o restraint) and have their greatest adv an-

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 bag a Lap/shoulder belt N o n eHead 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 stud y 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 h as defen ded th e meth-odology of its stud y [D. Martin, General Motors, personal com-mu nication, 1982].

8/14/2019 8217


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

tage over belts in th e abdom inal and p elvic re-gions, where belts actually perform w orse thanno restraint at all (though the injury severity levelis not great). Belts are less effective than bags inall regions except the chest area, w here th ey con-fer equ al protection.

The most serious injuries occur to the extrem i-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 acciden t d ata; some hav ederived 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 u seful to close this consideration by p re-senting some u nadjusted on-the-road d ata, 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 death s p er 1,000 car-years with theformer th an with the latter (28). Since some of the au tomatic belts were d isconnected (an esti-mated 22 percent) and many of the manu al beltswere being used (about 34 percent), this differenceprovides an underestimate of, or at least a lowerbound on, the life-saving potential of passive beltsystems. *

q Once again, it is important to keepin 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 th e other en d of the car line:full-size and luxur y cars, pr imarily the 10,000 to12,000 produced by GM between 1974 and 1976.By 1979 these cars had accumulated over 700million m iles of driving on American roa ds. Sta-tistical expectations for deaths and moderate tocritical injuries would have been 10 and 124, re-spectively, yet experience show ed half of thesefigures: 5 occupants of air-bag-equipped cars haddied and 62 had received moderate to critical injuries. There had been some 200 air-bag deploy-ments and only two know n d eployment 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 passanoth er acid test–-i.e., their effect on autom obile

insurance. 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 premiumsapp licable to medical (Medp ay) or personal in-

jury protection (PIP) coverage. Nationw ide h asestimated insu rance savings of roughly $20 peryear for a car equipped with automatic belts. Overthe lifetime of the car, this translates into a p res-ent value of $150 (33). A Highway Loss Data In-stitute study found reductions in Medpay an d PIPclaims of from 20 to 27 percent when comparingVW Rabbits with and without passive belts (13).These market d ata supp ort the accident data find-ings that passive restraints are an effective meansof redu cing d eath and disability.

not disconnecting their b elts minus 34 percent of m anual belt own erswearing their belts). The national discomect rate under a mand atorysystem would have to exceed 45 percent, given man ual belt u sagein the vicinity of 11 percent, for incremental belt usage to fall belowthat of the Rab bit (5s – ’11 = 44). Some participants in the passiverestraint deb ate believe that such a disconnect rate would beexceeded.

8/14/2019 8217


Ch. 4—Passive Restraint Systems: The Bag and the Belt q 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 automobile

companies, 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 persuad e[NH TSA] to rescind the ru le [FMVSS 208].

For some observers, both leaked companydocuments and analytical work have raised ques-tions abou t the v alidity of manu factur ers’ esti-mates.

During hearings leading to the l-year delay inimplementing FMVSS 208, Ford and GM pro-

vided estimates of incremental automatic belt costwh ich av eraged $114, includ ing $88 for man ufac-turers’ cost, $20 for markup, and $6 for incremen-tal fuel costs (additional fuel consumed as a resultof installation of the technology). Nor dh aus (33)has argued that these estimates are inconsistentwith the companies’ estimates of their investmentpr ogram s. He believes that a reasonab le estimatefor the true incremental cost to the consumerwould be $60, instead of the $88 cited by GM.Nordhaus’ figure is not inconsistent with earlierNH TSA and GM estimates based on incremen-tal costs in Rabbits and Chevettes, which ru n (in

1981 dollars), from $42 to $85. GM points out,however, that automatic belt systems for largecars would be more expensive than those designedfor sma ll cars. * In terms of annual costs (i.e.,amortizing an d dep reciating 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 po int, GM claimed th at it wou ld 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 in

excess of $600 per car . Ford’s estimates h ave a lsoq Personal commu nication, 1982. Manuf acturers’ estimates in 1978

(updated to 1981 dollars) ranged from $84 to $140, depen din g oncomfort and convenience features (28).

q q All of the dollar figures in th e 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 internalmemorand um placed the figure at from $425 to$1,150. The same sou rce indicated tha t for bothFord an d GM, manu facturer cost wou ld 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.Ann ual costs w ould range from $55 to $115. Inaddition, for those air bags deployed in vehicleswh ich w ere not d emolished, costs of reinstalla-tion wou ld become relevant. Such costs wouldcertainly exceed initial installation costs underconditions in which air bags would be mass-pro-du ced stand ard equ ipment. However, such costswou ld app ly to a very minor p roportion of vehi-cles, since only a small percentage of air bagswould ever d eploy du ring the lives of the vehicles.

The variation in air bag cost estimates mayrepresen t some strategic gaming, as Nordhaus (33)suggests, but it also reflects several technicalchanges and u ncertainties. Of fun dam ental 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 volu me of 100,000 un its, falling to betw een$650 and $700 if 400,000 units were produced. *

A separate estimate, based on quotations of compon ents, foun d th at the cost of a driver bagand inflator module would be 25 percent of itsbase cost (which assu med 13,000 units) if 900,000un its were p ur chased (28). Similarly, a passengerbag and inflator modu le 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 verydepend ent on the num ber produced and sold (28).

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

8/14/2019 8217


28 q Bac kground Paper #l: Mandatory Passive Restraint Systems in Automobiles: Issues and Evidence

The costs d iscussed above are d irect costs at-tributable to manufacture and installation of thepassive restraint systems, plu s small items of in-direct cost (e.g., add itional fuel consump tion). Ig-nored in these numbers is the potential directmonetary benefit the owner of a passive-restraint-equipped vehicle might derive from lower in-surance premium s. As noted above, the pr esent

discounted value of this benefit over the life of a 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, imp lying that th e pu rchaser of a p assive-belt-equipp ed car could end up saving

at least $35 over the life of the car (33).

8/14/2019 8217


5Cost-Benefit Analyses of;

Passive Restraints

8/14/2019 8217


5Cost= Benefit Analyses of

Passive Restraints

Should passive restraints be required on all newcars in the United States? Theoretically, theanswer can be d etermined by a comprehensivecost-benefit analysis (CBA), one which accountsfor all of the social costs and benefits associatedwith implementation of the rule—pecuniary andnonp ecun iary, tangible and intangible. In prac-tice, however, few CBAs even ap proach su ch acomplete analysis.

Non qu antifiable costs and benefits and thosewh ich are qu antifiable but not readily valued ar ecommonly mentioned and then put aside; oftenthey ar e ignored altogether. In either case, theyare left out of the final calculus of the CBA whicharrives at a “bottom line, ” an econom ic assess-ment of the worthiness of the program in qu es-tion. As such, the “bottom line” is deficient. Itcan and shou ld be recognized as a useful inpu tinto 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 w hen reading this chapter.

CBA is in its ascendancy as a tool of policyanalysis, particularly with regard to the issue of governmental 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 th e restraint issue date back morethan a d ecade.

In 1970, Lave and Weber (20) examined thecosts and benefits of seatbelts from the perspec-tive of the ind ividu al and found that benefits ex-ceeded costs if the value of an ind ividual’s life wasat least $10,000, a nu mber mor e than a full order

of magnitu de below th e smallest estimates of thevalu e of life (or livelihood) (50).

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

Rosen (46) add ressed th is question in an an alysisin 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 of life of $200,000), and they argued that the indi-vidual’s opportunity cost of time involved inbuckling and un buckling the seatbelt could easi-ly exceed th is amoun t.

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

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

Subject to tw o qualifications w hich they d ismiss

as insufficient to reverse their conclusion, ArnouIdand Grabowski argue that belt nonu se is not theresult of rational, informed decisionm aking, asThaler and Rosen had suggested it might be. Tothe contrary, these authors interpret their findingsas supporting the “insensitivity-to-low-probabil-ities” hyp othesis, discussed in chapter 2. It w asfound , 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 hou rly discomfort cost would

have 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 mand ated p assive

31

8/14/2019 8217


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

restraints, either autom atic belts or air b ags. Theanalysis finds that both systems w ould p rodu cesubstantial social benefits, with the benefits of airbags slightly exceeding those of au tomatic belts.However, the costs of the bags are so much higherthan those of the belts that Arnould an d Grabow-

ski conclud e belts are superior on the basis of netbenefit (or benefit-cost rat io). Indeed, th ey findthat only under very favorable conditions will airbags result in positive net benefits—i.e., the costsof bags qu ite likely wou ld exceed the ben efits.

Arnould and Grabowski’s study is noteworthyfor its sensitivity ana lysis, which tests the impa ctof 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 t he costs of inconvenience and dis-comfort associated w ith belts—costs w hich mu stbe considered p otentially large, given all of theeviden ce on belt use. And , like almost all CBAs,the an alysis ignores th e costs of the suffering ex-perienced by the loved ones of automobile acci-den t victims. (This is d iscussed further below.)

Sub ject to these limitations, Arnould and Gra-bow ski’s analysis estimates th at in a steady-statesituation, i.e., after passive restraints w ere 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 attr ibute no valueto injuries avoided. Lave (19) concurs with thefind ing that p assive belts wou ld be m ore cost ef-fective than air bags, though he d oes not dem on-strate analytically the basis of his conclusion.

Another recent CBA, by Nord haus (32,33), wasundertaken precisely to feed into the Departmentof Transportation’s (DOT’s) reconsideration of

FMVSS 208. Sponsored by five major automobileinsurance companies, this analysis also examinedand comp ared the two p assive restraint systems,

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

althoug h it w as focused on p assive belts, the sys-tem generally expected to dominate if FMVSS 208had gone into effect.

The analytical slant taken in N ordh aus’ CBAis somewh at d ifferent from that in Arn ould andGrabowski’s. Nord hau s concentrates on the (net)cost of delaying or reordering implementation of FMVSS 208, rather than making an “either-or”comparison of the status quo or a fully imple-mented passive restraint ru le. 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, theannu al net cost of a rescission wou ld equ al $2.4billion, assuming that all cars are equipped with

automatic belt systems. This number was derivedfrom National Highway Traffic Safety Adminis-tration (NHTSA) and man ufacturer data w hichNordhaus believes to be in error, biased againstthe belt system; so he views h is estimate as a low erboun d on the net benefits of passive belts. Und erthese conservative assumptions, the benefits of thepassive belt system ($3.6 billion) are th ree timesgreater than the costs ($1.2 billion). Nordhausestimates the total d iscoun ted n et social benefitsof the p assive restraint r ule 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-menta l, health, or safety ru le 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 m edical advances that al-lowed the virtual elimination of death from tuber-culosis over the last quarter centu ry.

Nord haus’ CBA shares with Arnould and Gra-bowski’s an effective use of sensitivity analysis.Unlike Arnould and Grabowski, Nord haus findsthat a w orld of air bags wou ld be preferable toa world of automatic belts. Despite the high costestimate he u ses to evaluate ba gs ($425), Nord -

8/14/2019 8217


.

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

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

As par t of his policy analysis, Nor dh aus exam-

ines the impact on the automobile manufacturersof implementing FMVSS 208. The only major or-ganized opp osition to implementation of the ru lethrough the years, the manufacturers have ex-pressed concern about the costs of adding passiverestraints as standard equipm ent and w hat thiswould do to the d emand for their product. Thepotential problem is of particular concern in 1982,with the d omestic indu stry in a d epressed condi-tion. Nord hau s presents a case that ad verse ef-fects wou ld be minimal. He suggests that it is evenconceivable that the indu stry wou ld benefit fromimplementation of FMVSS 208. This would oc-cur if consumers recognized the net economic sav-ings involved in buying a passive-restraint-equipp ed car, given an associated red uction inautomobile insurance costs.

Nordhaus’ is not the first CBA undertaken di-rectly in connection with governmental evalua-tion of a p assive restra int ru le. 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 sam e conclusion (13). Two years later,

another CBA accompan ied the an noun cement of a public hearing to be held by DOT.

At abou t the same time, Robertson (39) directeda survey which, though not itself a CBA, pro-du ced a finding of d irect 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 pay men ts to save 6,000 lives per yearand $17 per m onth ($204 per y ear) to save 12,000lives. ** It appears that many new car buyersmight stand prepared to pay considerably more

q Recall that und er the assumption he finds more realistic, Nord-

haus foun d a greater net benefit for the belts. However, he doesnot app ly a set of “more realistic” assump tions in th e air bag case,relying instead on man ufacturers’ 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 theyund erstate current value.

than the am ount p assive belt systems wou ld re-quire. Whether answ ers to a hypothetical ques-tion wou ld translate into equivalent action in themark etplace remains to be seen.

Graham, Henr ion, and Morgan (9) have iden-tified half a d ozen other CBAs on the occup ant

restraint issue and h ave produ ced a detailed anal-ysis of their own, one which compares passive re-straint systems with other methods of encourag-ing r estraint. Their analysis ran ks FMVSS 208 be-low other alternatives in terms of both net bene-fits and benefit-cost ratio, with a comp ulsory beltusage law having the highest ratio (in large partbecause its measurable costs are so low) and acombined air-bag/ mand atory-belt usage law pr o-du cing th e greatest net benefits (and saving themost lives). All of the alternatives these authorsexamine p rod uce p ositive n et benefits—i.e., eachalternative is preferable to the complete absenceof occupan t restraints (and sup erior to the cur-rent system of m erely requiring belts in cars). Inparticular, they estimate that the benefits of FMVSS 208 would have exceeded the costs by 95percent.

Thus, CBAs have served as inputs, the impor-tance of which is difficult to assess, throu ghou tthe long debate on a Federal passive restraint rule.Each of the analyses d iffers from the oth ers in cer-tain important ways: some ad opt a hu man capitalapproach to valuing life (or livelihood), whileothers use willingness-to-pay (50); basic data

sources, 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 syst emson insurance costs, while others ignore them, andso on.

Almost all of the stud ies 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 abou t using m anu al belts. It maybe d if-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 p otential 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

8/14/2019 8217


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

ord er to m ake 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 from

a compu lsory passive restraint ru le. The findingsare reasonably consistent and robust. In general,the stud ies 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 p assive belt option. While all of the stud iesfind 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 an alyses have articulate critics w ho challengebasic assum ptions of the mod els. In p articular,while ind ustry critics argue tha t belt cost estimatesin the analyses are too low, most of the criticismis focused on estimates of p assive seatbelt use. Theautom obile indu stry seems convinced th at 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 costswithou t significantly increasing effective passengerrestraint. Current (manual) belt users would there-by be “pu nished” by an unnecessary add itionalcharge, wh ile confirmed n onu sers wou ld have tobear the same add itional burden but w ould realizeno a dd itional p rotection (23).

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

first, illustrated by the work of Graham, Henrion,and Morgan (9), is that a find ing that a man da-tory p assive restraint ru le wou ld 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 app roach

that w ould maximize net social benefits.Use of the phrase “net social benefits” suggests

an important aspect, and limitation, of using CBAin a policy framewor k: as was noted at the outsetof this discussion, reasonable costs and benefitsare not th e only, nor necessarily the m ost impor-tant, variables in p olicy decisionm aking. 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 who

loses—not everyon e realizes a n et gain from im-plementation of a passive restraint rule-and itis this distinction that has m ade 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 attemp ts to d irectly value avoida nce of thepain an d su ffering of accident victims and theirloved on es. * As a result, the CBA's findings of positive net economic benefit support the noneco-nomic desire to minimize human suffering.

q This is considered indirectly in willingn ess-to-pay valuations,though these generally cover only the victims themselves and n otthe suffering of people close to them. Furthermore, willingness-to-pay estimates suffer from people’s inability to fully comprehend andcontemp late their own d eaths or serious disab ilities (50).

8/14/2019 8217


6

Philosophical and Ethical Issue;in the Passive Restraint Debate

8/14/2019 8217


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 of paternalism; 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 au tomatic 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 on e State capitol tothe next as a m ajority of States have r epealed theirmotorcycle h elmet law s (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 d riving behav ior, suchas speed limits and drun k-driving laws; and many,like the passive restraint rule, involve physical at-tributes of cars themselves—to nam e a few, shat-terproof glass, energy-absorbing steering-wheelcolumns, padd ed d ashboards, and today’s manu allap/ shoulder belts. Precedent does not make thenext related case—i.e., passive restraints—neces-sarily “right,” but it d oes serve as a sup portingargument.

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 hav e varied, it is d ecided ly not the casethat the public opposes passive restraints. For ex-amp le, a 1977 Gallup poll found that of the 83percent of respondents who had an opinion, a ma-

jority favored installation o f air bags on all cars.This theoretical sup por t for air bags does not d is-appear when respondents are reminded that theywill have to pay extra to hav e this extra p rotec-tion (13). * The will of the majority should notinvariably dominate th e rights of the minority,but in th is case, given the limited inp ut of pu blicopinion 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 tw otypes: economic and noneconomic.

Concerning the former, the nature of automo-bile and health insur ance is such that the insuredcosts of autom obile acciden ts are sp read over largenumbers of car and health insurance policyholders(and taxp ayers in the instan ce of pu blic healthinsurance). Thus, if an unrestrained victim of anautomobile accident experiences greater inju~and hence higher medical bills than wou ld havebeen the case had he or she been restrained, othersshare in the economic liability created by themedical treatment. In effect, one person’s inde-pend ent decision n ot to wear a seatbelt has ad -

q Once again, the question is whether (or how much of) suchsurvey support would translate into market decisions. Th e a u t o -

mobile m anufacturers seem convinced that in p ractice consumerswill not be w illing to pay the high cost air bags would en tail (23).It should be noted, however, that the indu stry’s limited experiencein this regard involves a radically different set of circumstances thanwould obtain in a world of mandated air bags.

37

8/14/2019 8217


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

verse consequences for other peop le’s p ocket-books. Pecun iary n egative 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 p assive belts. If the initial buyer p ur chases acar equipp ed 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, futu re passengers will be subjected to greaterrisk. This argument depends further on some spe-cific econom ic cond itions, which w ill be return edto, but it is sobering to note the following statisticsfrom a stud y of 137 crashes (with 172 peop le in-

jured) in Baltimore County, Md. (4):

app roximately half of the people injured wereoccup ants of vehicles no longer owned bytheir original purchasers;nearly 60 percent d id not ow n th e vehiclesin wh ich they w ere injured;three-quarters of the passengers were not re-lated to th e vehicle own ers; andalmost two-thirds of the victims were underthe age of 30, and only one-third of 16- to29-year-olds injured own ed the v ehicles inwhich they were injured.

In effect, these data challenge, or at least qu es-tion, the argu ment that new-car p urchasers shouldbe free to determine whether to invest in safetydevices like air bags since it is their own pr otec-tion that is in question. The health and safety of many 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 d o notfeel adequately protected. And, more to the point,future owners of used cars can have desired safe-

q Some of these pecuniary externalities could be overcome throughthe mark et. For example, insure rs 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 in surers 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 add ed wh en they purchase theircars. But can they, in the case of air bags?

There may b e firms that will retrofit a car tohave air bag p rotection, but th e cost is quite highbecause of the small scale of the business and thesizable economies of scale that p ertain to air b agmanufacture and installation. Thus, the prospec-tive buyer of an air bag system is pen alized eco-nomically by the small scale of active demand forair bags. Alternatively, all of those people w howou ld w ant to pu rchase air bags at mass-pro-duced prices would benefit from a requirement(or indu stry d ecision) to install bags in mu ch of the new-car fleet. *

Clearly, there is one obvious grou p of poten -tial economic victims of a passive restraint law:current regular users of manu al lap/ shoulderbelts. They are already receiving all of the p ro-

tection that w ould be afforded them und er a man-datory passive restraint rule, yet under such a rulethey wou ld have to pay m ore for their new cars.In effect, they wou ld be pen alized for other dr iv-ers’ poor judgment. If Nordhaus’ (33) assessmentis correct, the net cost of automobile ownershipwou ld drop un der a passive restraint rule, bene-fiting current belt users as well as nonusers. Butone might argue that current u sers should alreadybe receiving an insu rance 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 ” argum ent haspluses and minuses. Certainly there is a public in-terest in passive restraints which competes withthe p rivate interest in free individu al choice; thisis not clearly a case of unwarranted governmen-tal intrusion, pur e and simp le. Which of the in-terests one rates as dom inant p robably reflectsone’s basic political philosophy as mu ch as theinherent merits of the case.** In this regard , it is

q A ready response to this problem is that people can use thealready required lap/shoulder b elts 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, butwould like topurchase air-bag protection at a reasonable cost, though not at cur-rent small-scale high market prices.

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

8/14/2019 8217


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

interesting to read the assessment of George Will consenting adults to behave in ways d isastrous(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 individual

pensive devices that might save them from self- drivers is paid by society.

destruction:”

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

8/14/2019 8217


7.

Conclusion

8/14/2019 8217


7Conclusion

There is a basic philosophical issue in the d ebateon 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 ind ividu al freedom . * N ord hau s (33) hasconcluded that the ramifications probably wouldnot be as dire as ind ustry statements and docu-ments seem to suggest, at least w ith regard to th edemand for automobiles and manufacturer and

dealer revenu es. Ho wev er, there is a long h istoryof manufacturer opposition, with repeated pro-posals for the volun tary ph asing in of passive re-straints generally going unrealized (13,17).

The imp act of higher car pr ices on the dem andfor 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 Dep artment of Transportation for the 1981 hearingson passive restraint options (23):

We believe that [this issue of mandatory passive restraints] hinges

squarely on what the American public is willing to accept in terms of economic or behavioral bur den. 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 mand atory belt use law will be more effective, sooner, andat a fraction of the cost of a passive restrain t rule.

The mandatory belt use law clearly violates the in dividual freedomperspective at least as much as would passive restraints. Thus, GM’sconcerns seem firmly groun ded in p ragmatic economic considera-tions.

On that pragmatic level, GM’s points are w orthy of serious at-tention. GM is arguing, in essence, that the CBAs overestimate theamoun t of incremental belt usage that would b e achieved un der apassive restraint rule, and that the low levels GM anticipates wouldflip the cost-effectiveness balance over. In this regard, note thatArnould and Grabowsk i (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 th ey mightbe cost effective at that level, their effectiveness would be small com-pared with th e injury problem and the potential of belts to reduce it.

the effect of passive restraints on product liabil-ity claims, worrying “that ‘end less 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 wh en restraints worked as theywere su pp osed to. Thu s, the availability and costof liability insurance, manufacturers have feared,could loom as serious p roblems.

But some evidence suggests just the opposite.For a variety of reasons, insurance might be abun-dan tly available. And th ere is reason to b elievethat passive restraints might decrease, rather thanincrease, the n um ber 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 sup port for and oppositionto mandatory passive restraints, one should rec-ognize that m ost of the organized interests fall intothe sup port camp . Insurers have expressed their

support 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 u nequivocally favoring m andatory 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 imp lementation of a passive restraint ru le:

a major but un organized interest group, automo-bile consumers, has expressed an apparent pref-erence for air bags ov er p assive belts; yet the N a-tional Highway Traffic Safety Administration(NH TSA) has estimated th at if passive restraintswere required, 99 percent of the new-car fleet

43

8/14/2019 8217


—.

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

would have passive belts, with only 1 percent of cars coming equipped with air bags. The manufac-turers must have concluded that whatever con-sum ers’ abstract preferences might be, in bu yingpractice 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 au tomakers areconvinced that consumers w ould n ot 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 that

FMVSS 208 shou ld be rescinded.

Mand atory passive restraints would h ave rep-resented only one addition to a lengthy list of technological safety features required on au to-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 of FMVSS 208, and now over its rescission; and2) du e to the r ealistic poten tial of these techn olo-gies to make an extraordinary d ent in the d eathtoll of autom obile acciden ts. Estimates vary, bu tthe consensus suggests that passive restraintscould prevent a minimum of 6,000 and perhapsas man y as 12,000 highw ay d eaths a year. Giventhe current total of roughly 27,000 frontseat oc-cupant d eaths a year, a savings of this magnitu dewould constitute a truly major public healthvictory.

Less dram atic than lives saved, but more nu -merous, wou ld be injuries avoided. Here the es-timates also vary widely, depending in large parton the severity of the injuries considered. Arnou ldand Grabow ski (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 120,000mod erate to critical injuries wou ld be avoid ed.However they are grouped and counted, injuryredu ctions would be impressive.

With the d emise of the p assive restraint ru le,and over a decade of working toward its realiza-tion, NHTSA has a special obligation to seek al-ternative strategies to redu ce the m otor 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 th at wou ld ap par ently be ac-ceptable to the p ublic, were it not for its great cost.At the other extreme lies the mandatory belt-uselaw—an ap proach th at might b e quite effectiveand inexpensive, yet appears to be politicallyun acceptable. In betw een 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 m ix of ap-proaches should be explored.

From the p oint of view of OTA’s stud y Tech-nology and Handicapped People, it is unfortunatethat autom obile accident d ata do not permit acareful assessment of the number and severity of handicaps 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 red uce disabilities and hand -

icaps.In the context of a case stud y of the technolog-

ical prevention of han dicaps, the imp ortance of this redu ction is tw ofold. From a social/ fiscalpoint of view, the resource savings attributableto this prevention technology app ear to be sub-stantial. The costs of technologies for h and icapp edpeop le, both ind ividual an d collective, are con-siderable, as a glimpse at the m ain 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

q This is a benefit of the p revention 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 u ndoubtedly miss th e reduction in the need for manyof the collective technologies for the hand icapped . To this extent,the passive restraint CBAs un derstate the net b enefits of mandatingpassive restraints.

8/14/2019 8217


Ch. 7—Conclusion q 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 of handicaps.

More to the point, the true costs of hand icapsvastly exceed those that can be read ily 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, includ ed in non eof the passive restraint CBAs, must be enorm ous(1). Thus, regardless of its purely fiscal implica-tions, the prevention approach warrants especiallycareful attention.

8/14/2019 8217


References

8/14/2019 8217


8/14/2019 8217


Documents

8217