November 17, 2008 David Kelly Acting Administrator National Highway Traffic Safety Administration 1200 New Jersey Avenue, SE, West Building Washington, DC 20590 Notice of Proposed Rulemaking; 49 CFR Part 571 Federal Motor Vehicle Safety Standards, Motorcycle Brake Systems; Docket No. NHTSA-2008-0150 Dear Mr. Kelly: On September 17, 2008, the National Highway Traffic Safety Administration (NHTSA) announced a proposal to amend Federal Motor Vehicle Safety Standard (FMVSS) No. 122, Motorcycle Brake Systems. The amendments would strengthen the requirements and test procedures that cover many aspects of motorcycle brake systems, including antilock braking systems (ABS) voluntarily installed by manufacturers. The Insurance Institute for Highway Safety (IIHS) supports these changes and submits with this comment our recent research showing that ABS is improving motorcycle safety. The proposed changes to FMVSS 122 will create minimum performance standards that have been tested successfully elsewhere in the world and will help keep motorcycles with unsafe brakes from being sold in the United States. The proposed ABS tests in particular will help ensure that if ABS is installed on motorcycles, the system will provide operators with adequate stopping distances and stability. IIHS and its affiliated Highway Loss Data Institute (HLDI) are studying the safety effects of motorcycle ABS as its availability increases. Results from our initial analysis of a variety of models show that ABS is having a positive effect (IIHS, 2008). Both the rate of fatal motorcycle crashes and the frequency of crashes for which insurance collision claims are filed are lower among motorcycles with ABS compared with the same motorcycles without ABS. The attached IIHS study (Teoh, 2008) shows 6.7 fatal crashes per 10,000 registered vehicle years among motorcycles not equipped with ABS during 2005-06. The corresponding rate for the same models equipped with optional ABS was 4.1 fatal crashes per 10,000 registered vehicle years — 38 percent lower. The attached HLDI (2008) study shows that the estimated effect of ABS was a 21 percent decrease in overall collision losses, primarily because collision claim frequencies for motorcycles with ABS were 19 percent lower than for motorcycles without ABS. The importance of equipping motorcycles with ABS increases as motorcycling continues to grow in popularity. Motorcycle sales more than tripled between 1997 and 2005, and motorcyclist deaths have more than doubled since 1997. Due largely to safety features that are increasingly available as standard equipment on passenger vehicles, there were fewer passenger vehicle occupant deaths in 2007 than in any year since NHTSA began collecting these data in 1975. On the other hand, more motorcyclists died in crashes in 2007 than in any other year during the same time period.

David Kelly November 17, 2008 Page 2 The Institute supports the proposed strengthening of FMVSS 122 to establish stronger minimum performance standards for motorcycle brake systems. Given the beneficial effects of ABS, as documented in our research, we urge NHTSA to consider further changes to FMVSS 122 to encourage or require ABS on all motorcycles. Sincerely, Joseph M. Nolan, M.S. Senior Vice President, VRC Operations cc: Docket Clerk, Docket No. NHTSA-2008-0150 References and Attachments Highway Loss Data Institute. 2008. Motorcycle antilock braking system (ABS). HLDI Bulletin 25(1). Arlington, VA. Insurance Institute for Highway Safety. 2008. Antilock brakes on motorcycles reduce both crash frequencies and deaths. Status Report 42(9):1-3. Arlington, VA. Teoh, E.R. 2008. Effectiveness of antilock braking systems in reducing fatal motorcycle crashes. Arlington, VA: Insurance Institute for Highway Safety.

IF THEY NEED TO

STOP

Vol. 43, No. 9, Oct. 22, 2008

on a dime, these riders will enjoyan advantage most others don’tbecause of the antilock brakes on

their motorcycle. Two new stud-ies indicate crash reduc-

tions associatedwith anti-

SR43-9 TO PDF:SR 43-9 10/16/08 11:37 AM Page 1

2

4

6

withoutantilocks

withantilocks

-10

0

claimfrequency

claimseverity

2 Status Report, Vol. 43, No. 9, Oct. 22, 2008

locks. Both the frequency of crashes forwhich insurance claims are filed and therate of fatal motorcycle crashes go downamong bikes with antilock brakes.

The importance of equipping bikes withantilocks increases as motorcycling prolifer-ates. Motorcycle sales more than tripledfrom 1997 to 2005. Deaths of motorcyclistshave more than doubled since 1997, withsome kinds of bikes having much higherdeath rates than others (see Status Report,Sept. 11, 2007; on the web at iihs.org). About5,000 motorcyclists died in crashes last year.

The new study of fatal motorcycle crash-es was conducted by Institute researchers,while the analysis of insurance claims is byresearchers at the affiliated Highway LossData Institute (HLDI). Adrian Lund is presi-dent of both organizations.

“Even though adding antilocks won’tmake motorcycling as safe as going by car,it’s something manufacturers can do to re-duce the risk of traveling on 2 wheels insteadof 4,” Lund says. “It’s a way to reduce thechances of overturning a bike and crashing,so it can save lives among people who choosemotorcycles for their basic transportation,to save on gasoline, or just for fun.”

When antilocks are needed: Stopping amotorcycle is trickier than stopping a car.For one thing, front and rear wheels typicallyhave separate brake controls. Both under-braking and overbraking the front and rearwheels contribute to crashes (see StatusReport, June 21, 1979). In an emergency, arider faces a split-second choice to brakehard, which can lock the wheels and causea motorcycle to overturn, or to hold backon the brakes and risk running headlonginto the emergency.

This is when antilocks can help. They reduce brake pressure when they detect impending lockup and

about to be struck from behind, for example.But the new studies indicate that antilocksreduce crashes overall and save lives.

Crash reduction benefit: The HLDI studycompares insurance losses under collisioncoverage for 12 motorcycle models withoptional antilock brakes versus the samemodels without this option. The research-ers evaluated the effects of antilock brakeson both the frequency of insurance claimsthat are filed for crash damage and theaverage cost of the damage, after ac-counting for rider age and gender,motorcycle age, and other factors

that influence the like-lihood of a crash.

Regression analysis revealed 21 percentlower insurance losses for motorcycles withantilocks, primarily because the claim fre-quency was 19 percent lower than for bikeswithout antilocks. These findings are basedon a dataset of 72,000 insured years of 2003-

increase the pressure again when traction isrestored. Brake pressure is evaluated multi-ple times per second, so riders may fullybrake without fear of locking the wheels.

Antilocks won’t prevent every motorcy-cle crash. They won’t help a rider who’s

FATAL CRASHESPER 10,000 MOTORCYCLE REGISTRATIONS,2001-06 MODELS DURING 2005-06

INSURANCE CLAIMSPERCENT CHANGE IN COLLISION LOSSES FOR 2003-07 MODEL MOTORCYCLES WITH ANTILOCKS

SR43-9 TO PDF:SR 43-9 10/16/08 11:37 AM Page 2

Status Report, Vol. 43, No. 9, Oct. 22, 2008 3

07 model Honda, Suzuki, Triumph, and Yam-aha bikes (an insured year is 1 motorcycleinsured for 1 year or 2 insured for 6 monthseach, etc.). BMW models aren’t includedbecause it’s impossible to determine fromvehicle identification numbers which oneshave optional antilocks and which don’t.Harley-Davidsons aren’t included becauseantilocks were added after the study years.

Antilock brakes “appear to reduce colli-sion claims,” says Matthew Moore, HLDIvice president and lead author of the study,“but they don’t affect the severity of thecrashes for which claims are filed. The costof these claims doesn’t go down.”

Lives are being saved: In a complemen-tary study, Institute researchers examinedrates of fatal crashes of motorcycle modelswith and without antilocks. Eight modelswere studied, a subset of the 12 included inthe HLDI analysis. The other 4 models wereexcluded because of sample size limitations.

A main finding is that there were 6.6 fatalcrashes per 10,000 registered motorcycleswithout antilocks during 2005-06. The corre-sponding rate for the same bike modelsequipped with optional antilocks is 4.1, or38 percent lower. Institute statistician EricTeoh, author of the study, says the findings

are statistically signifi-cant at the 90 percentconfidence level.

2008 MODELS WITH ANTILOCKSAntilocks are optional except as noted; models in bold are included in one or both of the newstudies of the effectiveness of antilock brakes

BMW K1200GT (std)BMW K1200LT (std)BMW R1200RT (std)BMW F800S/F800STBMW G650 XchallengeBMW G650 XCountryBMW G650 XMotoBMW HP2 Megamoto/HP2BMW K1200R/K1200R SportBMW K1200SBMW R1200R/R1200SBMW R900RT

Can-Am SpyderHarley-Davidson Electra Glide ClassicHarley-Davidson Electra Glide StandardHarley-Davidson Night RodHarley-Davidson Night Rod SpecialHarley-Davidson Road Glide/Road KingHarley-Davidson Road King ClassicHarley-Davidson Screaming Eagle Electra GlideHarley-Davidson Screaming Eagle Road KingHarley-Davidson Street GlideHarley-Davidson Ultra Classic Electra GlideHarley-Davidson V-Rod

Honda Gold WingHonda Interceptor 800Honda ReflexHonda Silver WingHonda ST1300

Kawasaki Concours 14Moto Guzzi Norge 1200

Suzuki Bandit 1250SSuzuki B-KingSuzuki Burgman 650 ExecutiveSuzuki SV650/SV650S/SV650SAFSuzuki V-Strom 650

Triumph Sprint STTriumph TigerYamaha FJR1300 (std)Yamaha FJR1300 Electric Shift (std)

Antilocks on cars versus motorcycles:Passenger cars began to be equipped withantilock brakes during the 1970s, after stud-ies conducted on the test track indicatedthey reduce stopping distances. However,this promise didn’t pan out in real-worldcrashes (see Status Report, Jan. 29, 1994).Antilocks didn’t reduce relevant collisions.

“It isn’t surprising that antilock brakesare more beneficial on motorcycles thanthey are on cars because the 2-wheelers areso much less stable, and it’s this instabilitythat contributes to so many crashes,” Lundpoints out. “By reducing wheel lockup dur-ing braking, antilocks keep a lot of motorcy-cles from overturning.”

Antilock brakes are recent additions tomotorcycles. They’re available almost exclu-sively as optional equipment (see list), whichmeans shoppers have to find models onwhich the option is offered and then payextra for it. Antilocks were on only 18 per-cent of the motorcycles included in the newstudies of effectiveness.

For a copy of “Antilock braking systemsfor motorcycles and insurance collision loss-es” by M. Moore and Y. Yan or “Effectivenessof antilock braking systems in reducing fatalmotorcycle crashes” by E. Teoh, write: Publi-cations, Insurance Institute for HighwaySafety, 1005 N. Glebe Rd., Arlington, VA 22201,

or email publica-tions@iihs.org.

SR43-9 TO PDF:SR 43-9 10/16/08 11:37 AM Page 3

CHILD SEAT USEAMONG KIDS INCRASHES GOES UPUse of child safety seats has surged since1999 among restrained children youngerthan 9 riding in insured vehicles. Restrainttypes also have changed. These are the mainfindings of new research from the decade-long Partners for Child Passenger Safetystudy of the Children’s Hospital of Philadel-phia (CHOP) and State Farm, with supportfrom the Association of International Auto-mobile Manufacturers. It’s based on 1998-2007 insurance claims and phone surveydata on more than 875,000 kids in crashes.Overall safety seat use among restrainedchildren 8 and younger rose to 80 percent in2007 from 51 percent 8 years earlier.

Virtually 100 percent of restrained chil-dren 3 and younger in crashes have been insafety seats since 1999. Safety seat use ismuch lower among older children. Progresshas been made, but there’s room to improve.

In 1999 only 15 percent of restrained 4-8year-olds in the CHOP study were in anappropriate restraint — a harness restraintor booster. By 2007 appropriate restraintuse in this group had quadrupled to 63 per-cent. The rest of restrained 4-8 year-olds rodein adult belts alone. Typically, such beltsdon’t begin to fit properly until kids grow toabout 4 feet, 9 inches tall.

“Along with the increase in the number ofkids riding in child safety seats, we can alsosee changes in the types of restraints they areusing now versus 10 years ago,” says KristyArbogast, director of engineering at CHOP’sCenter for Injury Research and Prevention,where the study was conducted. She saysmore restrained 4 and 5 year-olds ride inboosters now instead of harness restraints.

Only 31 percent of appropriately re-strained 4-5 year-olds rode in harness re-straints during 2007. Highback boosters areslightly more popular now than backlessones among restrained 4-5 year-olds. Butbackless boosters are used nearly 3 times asoften as highbacks for 6-8 year-olds.

MORE STATES BAN DRIVERS’ TEXTINGCalifornia and Alaska are the latest US states to ban text messaging by drivers of all ages,not just teenagers. The two states join Connecticut, the District of Columbia, Louisiana,Minnesota, New Jersey, and Washington in banning texting by all drivers amid concernthat such distractions increase crash risk. Alaska’s law took effect Sept. 1. California’s banbegins Jan. 1, 2009. Both states make the use of an electronic device to write, send, orread text messages a primary offense, meaning that police officers can pull over drivers

solely for violating the bans. Texting while driving also is a primary offensein Connecticut, the District of Columbia, Minnesota, and New Jersey.

Bans in Louisiana and Washington are secondary, so motor-ists must be violating another traffic law in order to be

stopped by police for texting. Nine states have textmessaging bans that apply only to novice drivers.For details on state bans on text messaging by driv-ers, go to www.iihs.org/laws/cellphonelaws.aspx.

There’s lots of anecdotal evidence tying textingwhile driving to crashes, but not much data fromreal-world collisions. Studies have linked cellphoneuse with crash risk (see Status Report, July 16, 2005,and March 22, 1997; on the web at iihs.org).

In one of the first published studies on textingand driving, the Transport Research Labora-

tory in the United Kingdom (on the web attrl.co.uk) found that texting degrades per-

formance in a driving simulator. Re-searchers found that composing a text

message affected driving more thanreading one. The 17 drivers in the

study — all were 17-24 years old —had slower reaction times, were

more likely to drift outof their virtual

lanes, and weremore likely to

reduce theirspeeds while

they weretexting.

4 Status Report, Vol. 43, No. 9, Oct. 22, 2008

SR43-9 TO PDF:SR 43-9 10/16/08 11:37 AM Page 4

Status Report, Vol. 43, No. 9, Oct. 22, 2008 5

Among the study’s other findings arethat 60 percent of crashes involving chil-dren occur within 10 minutes of home, and84 percent take place within 20 minutes ofhome. Only 14 percent of crashes are onroads where posted speed limits are 55 mphor higher, but these crashes result in the

Previous CHOP research shows boosterslower crash injury risk by 59 percent for 4-7year-olds compared with belts alone. Boost-ers elevate children so lap and shoulderbelts are properly positioned. Earlier thismonth the Institute released evaluations of41 booster models, finding that several

Of the states in the study, booster seatuse among 4-8 year-olds was lowest in Ohio(18 percent) and Texas (20 percent). Notsurprisingly, these states don’t have boosterseat laws. On the other hand, 2 of the 5states with the highest use of boosters,Pennsylvania (72 percent) and Illinois (62

don’t improve belt fit (see Status Report,Oct. 1, 2008; on the web at iihs.org).

Arbogast attributes the increase in boost-er use among older kids to education of par-ents and caregivers plus state laws requiringolder kids to ride in safety seats. Laws in 43states and the District of Columbia includebooster provisions (on the web at iihs.org/laws/restraintoverview.aspx).

“More parents than ever now realize thatkids need the help of a booster seat to makesure the belt fits properly across the bonyparts of their lap and shoulder rather thanacross the soft belly or the neck, which aremore prone to injury,” Arbogast says.

percent), do require child restraints orboosters for children through age 7.

CHOP researchers found that parentsaren’t widely using lower anchors and teth-ers for children, or LATCH, which are sup-posed to make it easier to attach infant andchild restraints securely to vehicle seats(see Status Report, Jan. 16, 1999; on the webat iihs.org). LATCH has been required innew vehicles and on child restraints since2002. However, only 43 percent of all chil-dren buckled into restraints in vehiclesequipped with LATCH in 2007 were riding inseats attached to the lower anchors, theCHOP study reports.

highest rates of injury. Nearly half of allcrashes involving children occur on roadswith posted speed limits of 25 to 44 mph.

Although the American Academy of Pedi-atrics recommends that children youngerthan 13 ride in the back seats of vehicles,about 30 percent of all 8-12 year-olds ride inthe front. Positioning children in back seatsreduces the risk of fatal injuries in crashesby about one-third among kids 12 andyounger (see Status Report, June 27, 1997; onthe web at iihs.org).

For a copy of the September 2008 “Part-ners for Child Passenger Safety: fact andtrend report” go to www.chop.edu/carseat.

SR43-9 TO PDF:SR 43-9 10/16/08 11:37 AM Page 5

6 Status Report, Vol. 43, No. 9, Oct. 22, 2008

FASTER, HEAVIER GOLF CARTS GET THUMBS DOWN FROM

FEDERAL REGULATORS The federal government says it isn’t willing to trade highway safety for fuel econ-omy in denying 4 petitions that sought to increase the maximum gross vehicleweight for low-speed vehicles and also launch a class of medium-speed vehi-cles. The September decisions by the National Highway TrafficSafety Administration (NHTSA) mean that the weight andtop speed of these golf cart-like vehicles, which don’t haveto meet all the safety rules that apply to cars, remaincapped at 3,000 pounds and 25 mph.

“While NHTSA agrees with the importance of environ-mental issues, the agency believes that it is neither necessary nor appro-priate to significantly increase the risk of deaths and serious injuries tosave fuel,” the agency said in denying petitions from EnvironmentalMotors, Porteon Electric Vehicles Inc., and Mirox Corporation. The com-panies asked the agency to create a class of medium-speed vehicles withspeed capabilities of up to 35 mph, arguing they’d fill a need for fuel-effi-cient vehicles for use in fast urban traffic.

They’re known as neighborhood electric vehicles, street-legal golfcarts, and minitrucks, among other names (see Status Report,April 6, 2002; on the web at iihs.org). These electric or gaso-line-powered low-speed vehicles are designed to haul peopleand cargo on private land, such as retirement communities,farms, amusement parks, and construction sites, but they’reoften driven on public streets.

Forty-six states regulate their use, with most limiting theirspeed to no greater than 25 mph (on the web at iihs.org/laws) on public roadways with speed limits of no more than35 mph. They’re exempt from most federal safety standardsthat apply to cars, and they aren’t required to meet any cri-teria for vehicle crashworthiness, so they’d be out of theirleague in crashes with other vehicles going 35 mph.

Electronic Transportation Applications had sought toincrease the maximum allowable gross vehicle weight forelectric-powered low-speed vehicles to 4,000 pounds.

In denying Electronic Transportation Applications’ petition,NHTSA said, “We believe that vehicles over 3,000 pounds arecapable of complying with the full requirements” of the federal motorvehicle safety standards. Increasing the allowable gross vehicle weight, NHTSA said,“would encourage the use of [low-speed vehicles] in circumstances where it could pose anunreasonable risk to safety.” The agency noted that some of the smallest passenger cars —Honda Insight and Toyota Echo, for example — have gross vehicle weights of about 3,000pounds or less and still comply with safety standards. (Read both decisions at http://edock-et.access.gpo.gov/2008/pdf/E8-22736.pdf and 22737.pdf.)

Federal crash databases don’t include a specific category for low-speed vehicles so it’shard to track their crashes. News reports frequently chronicle deaths and injuries that resultwhen these vehicles collide with larger passenger vehicles.

SR43-9 TO PDF:SR 43-9 10/16/08 11:37 AM Page 6

NEW ‘UNDERSTANDING CRASHES’ VIDEOWhy do some car crashes produce only minor injuries? How can a singlecrash of a car into a wall involve 3 separate collisions? Award-winning science educator Griff Jones visits the Institute’s Vehicle Research Center to answer these and other questions in a 24-minute video that’s a follow-upto a previous Institute production, “Understanding car crashes: it’s basicphysics” (2000). In the new video, Jones examines the laws of nature thatdetermine what happens to the human body in a crash. Order “Understandingcar crashes: when physics meets biology” ($35) online at iihs.org/videos.

SR43-9 TO PDF:SR 43-9 10/16/08 11:37 AM Page 7

NON-PROFIT ORG.U.S. POSTAGE

PAIDPERMIT NO. 252ARLINGTON, VA

21st Century InsuranceAAA Mid-Atlantic Insurance GroupAffirmative InsuranceAIG Agency AutoAIG DirectAlfa InsuranceAlfa Alliance Insurance CorporationAllstate Insurance GroupAmerican Family Mutual InsuranceAmerican National Property and CasualtyAmeriprise Auto & HomeAmerisure InsuranceAmica Mutual Insurance CompanyAuto Club GroupAuto Club South Insurance CompanyBituminous Insurance CompaniesBristol West InsuranceBrotherhood MutualCalifornia State Automobile AssociationCapital Insurance GroupChubb Group of Insurance CompaniesConcord Group Insurance CompaniesCotton States InsuranceCOUNTRY FinancialCountrywide Insurance GroupErie Insurance GroupEsurance

Contents may be republished with attribution. This publication is printed on recycled paper.

Farm Bureau Financial ServicesFarm Bureau Mutual Insurance Company of IdahoFarmers Insurance Group of CompaniesFarmers Mutual of NebraskaFirst Acceptance CorporationFlorida Farm Bureau Insurance CompaniesFrankenmuth InsuranceThe GEICO GroupGainscoGeneral Casualty Insurance CompaniesGeorgia Farm Bureau InsuranceGMAC InsuranceGrange InsuranceThe Hartford Hanover Insurance GroupHigh Point Insurance GroupHomeowners of AmericaICW GroupIndiana Farm Bureau InsuranceKemper, a Unitrin BusinessKentucky Farm Bureau InsuranceLiberty MutualThe Main Street America GroupMarkel CorporationMercury Insurance GroupMetLife Auto & HomeMichigan Insurance CompanyMiddleOakMMG InsuranceMutual of Enumclaw Insurance CompanyNationwide InsuranceN.C. Farm Bureau Mutual Insurance CompanyNodak Mutual Insurance

Norfolk & Dedham GroupOhio Casualty GroupOklahoma Farm Bureau Mutual Insurance CompanyOregon Mutual InsurancePalisades InsurancePekin InsurancePEMCO InsuranceThe Progressive CorporationResponse InsuranceRockingham Group Safeco InsuranceSamsung Fire & Marine Insurance CompanyS.C. Farm Bureau Mutual Insurance CompanySECURA InsuranceShelter InsuranceSompo Japan Insurance Company of AmericaState Auto Insurance CompaniesState FarmTennessee Farmers Mutual Insurance CompanyTokio Marine NichidoThe Travelers CompaniesUnitrinUSAAVirginia Farm Bureau Mutual InsuranceWest Bend Mutual Insurance CompanyZurich North America

FUNDING ASSOCIATIONSAmerican Insurance AssociationNational Association of Mutual Insurance CompaniesProperty Casualty Insurers Association of America

The Insurance Institute for Highway Safety is a nonprofit scientific and educational organization dedicated to reducing deaths, injuries, and property damage from crashes on the nation’s highways. The Institute is wholly supported by auto insurers:

1005 N. Glebe Rd., Arlington, VA 22201 Phone 703/247-1500 Fax 247-1588Internet: www.iihs.orgVol. 43, No. 9, Oct. 22, 2008

Antilock brakes on motorcycles reduceboth crash frequencies and deaths ...........1

2008 motorcycles with antilocks .....................3

Text messaging bans now cover all drivers in7 states and the District of Columbia .............4

Child safety seat use in crashes has increasedduring the past decade ....................................4

Souped-up golf carts will remain subject tolimitations on weight and speed, the federalgovernment decides ........................................6

New Institute video explores what happensto the human body in a car crash ...................7

SR43-9 TO PDF:SR 43-9 10/16/08 11:37 AM Page 8

Antilock braking systems (ABS) on motorcycles are designed toallow riders to make urgent, yet controlled stops. Motorcyclesare fundamentally unbalanced; they are kept stable at very lowspeeds by a rider holding the handlebar and maintaining bal-ance. At higher speeds, stability comes almost exclusively fromthe gyroscopic effect of the wheels. While at speed, if one ofthe wheels stops rotating for a fraction of a second, the result isimmediate instability. The effect is more pronounced if it hap-pens to the front wheel, where a fall is almost inevitable —especially while cornering or leaning the motorcycle. ABS hasindependent braking sensors for each wheel. If the systemdetects a locked wheel, it releases the brake to allow that tireto retain grip before reapplying the brake. ABS then modulatesbraking pressure to achieve optimum braking.

This Highway Loss Data Institute (HLDI) bulletin compared thecollision losses of 12 motorcycle models available with option-al ABS. Model years of the motorcycles studied ranged from2003 to 2007. Significant reductions in collision claim fre-quencies and overall losses were found for motorcyclesequipped with ABS. No significant reductions were found forclaim severity.

For motorcycles to be included in the study, their vehicle iden-tification numbers (VINs) had to have an ABS indicator. Thisallowed for very tight control over the study population. Twelvemotorcycles met this criterion. It should be noted that therewere motorcycles available with ABS that were not included inthe study because their VINs did not have an ABS indicator.Data were collected by make and series, rated driver age andgender, and vehicle age and density. Vehicle age was definedas the difference between calendar year and model year meas-ured in years. Previous HLDI studies have shown that all ofthese factors have a significant impact on insurance losses. Forinsurance purposes, a rated driver is assigned to each vehicleon a policy. Information on the actual driver at the time of aloss is not available in the HLDI database. For the present studydata were stratified by rated driver age group (<25, 25-39, 40-64, 65+, or unknown) and gender (male, female, or unknown).The dataset also was stratified by make/series and vehicle den-sity (<100, 100-499, and 500+ vehicles per square mile). Forexample, a 1-year-old Honda Gold Wing, ABS equipped, witha 40-64 year old male as the rated driver, and garaged in anarea with a vehicle density of 100-499 vehicles per square mileconstituted one unit of observation. The distribution of motor-cycle collision exposure for the six independent variables islisted in Appendix A. It is important to note that rated driverfactors and vehicle density were included to control for theirpotential impact on losses and not to produce estimates for

those variables. The estimated parameters for those variablesmay not generalize from this subset to the much larger motor-cycle population.

Regression analysis was used to quantify the effect of ABS onmotorcycle collision losses while controlling for other covari-ates. Claim frequency was modeled using a Poisson distribu-tion, whereas claim severity was modeled using a Gamma dis-tribution. Both models used a logarithmic link function.Estimates for collision overall losses were derived from theclaim frequency and severity models. Reference categories forthe categorical independent variables were assigned to the val-ues with the highest exposure. The reference categories were asfollows: make/series = Honda Gold Wing, ABS = without ABS,rated driver age range = 25-39, vehicle density = 100-499 vehi-cles per square mile, and rated driver gender = male. Losses foreach unit of observation were weighted by the exposure in thelinear regression. The key independent variable in the model,ABS, was treated as categorical. Models were constructed thatexamined the interaction of the rated driver factors and vehicledensity with the presences or absence of ABS. None of thoseinteractions were found to be significant.

Summary results of the regression analysis of motorcycle colli-sion claim frequencies using the Poisson distribution are listedin Table 1. Results for all independent variables in the model,including ABS, had p-values less then 0.05, indicating theireffects on claim frequencies were statistically significant.

Detailed results of the regression analysis using claim frequen-cy as the dependent variable are listed in Table 2. The tableshows estimates and significance levels for the individual val-ues of the categorical variables. To make results more illustra-tive, a column was added that contains the exponents of theestimates. The exponent of the intercept equals 0.0000826claims per day, or about 3 claims per 100 insured vehicleyears. The intercept outlines losses for the reference (baseline)categories: the estimate corresponds to the claim frequency fora new Honda Gold Wing without ABS, garaged in a mediumvehicle density area, and driven by a male age 25-39. Theremaining estimates are in the form of multiples, or ratios rela-tive to the reference categories. For example, the estimate cor-responding to female gender equals 0.81, so female rated driv-ers had estimated claim frequencies 19 percent lower thanthose for male rated drivers.

The estimate corresponding to motorcycle ABS (-0.22) washighly significant (p=0.002). The estimate corresponded to a 19percent reduction in claim frequencies for motorcyclesequipped with ABS.

VOL. 25, NO. 1 APRIL 2008

Highway Loss Data Institute BulletinMotorcycle Antilock Braking System (ABS)

TABLE 1 SUMMARY RESULTS OF LINEAR REGRESSION ANALYSISOF COLLISION CLAIM FREQUENCIESDEGREES

OF FREEDOM CHI-SQUARE P-VALUE

Vehicle Age 1 196.270 < 0.0001

Rated Driver Age 4 57.040 < 0.0001

Vehicle Density 2 18.150 0.0001

Vehicle Make/Series 11 292.300 < 0.0001

Rated Driver Gender 2 8.060 0.0178

ABS 1 9.580 0.0020

INTERCEPT -9.4014 3.015 0.0826 12948.8 <0.0001VEHICLE AGE -0.2628 0.769 0.0192 187.54 <0.0001RATED DRIVER AGEUnknown 0.1791 1.196 0.0925 3.75 0.052914-24 0.5839 1.793 0.094 38.58 <0.000125-39 0 1.000 040-64 -0.1433 0.866 0.0649 4.88 0.027265+ -0.0747 0.928 0.1026 0.53 0.4663

VEHICLE DENSITY0-99 -0.1332 0.875 0.0582 5.23 0.0222100-499 0 1.000 0500+ 0.122 1.130 0.0514 5.62 0.0177

RATED DRIVER GENDERFemale -0.205 0.815 0.0845 5.88 0.0153Male 0 1.000 0Unknown 0.0431 1.044 0.0527 0.67 0.4135

VEHICLE MAKE/SERIESHonda Gold Wing 0 1.000 0Honda Interceptor 800 0.8988 2.457 0.1077 69.6 <0.0001Honda Reflex 0.5596 1.750 0.1093 26.2 <0.0001Honda Silver Wing 0.7658 2.151 0.1025 55.81 <0.0001Honda ST1300 0.1745 1.191 0.1128 2.39 0.1218Suzuki Bandit 1250 1.4525 4.274 0.2295 40.06 <0.0001Suzuki Burgman 650 0.7284 2.072 0.0954 58.29 <0.0001Suzuki SV650 1.0251 2.787 0.0747 188.54 <0.0001Suzuki V-Strom 650 -0.084 0.919 0.1352 0.39 0.5345Triumph Sprint ST 1.077 2.936 0.1357 63 <0.0001Triumph Tiger 0.3632 1.438 0.2177 2.78 0.0953Yamaha FJR1300 0.3276 1.388 0.0949 11.92 0.0006

ABSABS Model -0.2151 0.806 0.0709 9.19 0.0024Non-ABS Model 0 1.000 0

EXPONENT STANDARD CHI-PARAMETER ESTIMATE (ESTIMATE) ERROR SQUARE P-VALUE

TABLE 2 DETAILED RESULTS OF LINEAR REGRESSION ANALYSISOF COLLISION CLAIM FREQUENCIES

Motorcycle collision claim frequencies increased with increases in vehicle density. Claim frequencies in high vehicledensity areas were estimated to be 13 percent higher (p=0.02) than those in medium vehicle density areas, whereas claimfrequencies in low vehicle density areas were estimated to be 13 percent lower (p=0.02). Claim frequencies were esti-mated to decrease 23 percent (p<0.0001) for each 1-year increase in vehicle age.

Individual make/series motorcycles were included in the model, and estimates of their effect on collision claim frequen-cies were found to be significant. As previously mentioned, the reference category for the make/series variable was theHonda Gold Wing. Significant predictions for make/series ranged from 1.39 for the Yamaha FJR 1300 to 4.27 for theSuzuki Bandit 1250. All make/series estimates were significant at the p=0.001 level except for the Honda ST1300, SuzukiV-Strom 650, and Triumph Tiger. The lack of significance for these individual values of the make/series variable indicatethat estimated claim frequencies for these motorcycles were similar to the Honda Gold Wing (reference category).

Driver age was highly significant in predicting motorcycle collision claim frequency. Estimated claim frequencies for rateddrivers 24 and younger were 79 percent higher (p<0.0001) than those for rated drivers ages 25-39 (reference category),whereas estimated claim frequencies for rated drivers ages 40-64 were 13 percent lower (p=0.03). The estimated 7 per-cent decrease in claim frequency for rated drivers 65 and older was not significant. Driver gender also significantly pre-dicted collision claim frequencies. Estimated claim frequencies for female rated riders were 19 percent lower (p=0.02)than those for male rated riders.

Summary results of the regression analysis of motorcycle collision claim severities using the Gamma distribution are list-ed in Table 3. Of the six variables included in the analysis, only vehicle age and make/series had p-values less than 0.05.Neither the rated driver nor the driving environment affects the claim size.

TABLE 3 SUMMARY RESULTS OF LINEAR REGRESSION ANALYSISOF COLLISION CLAIM SEVERITIESDEGREES

OF FREEDOM CHI-SQUARE P-VALUE

Vehicle Age 1 12.400 0.0004

Rated Driver Age 4 3.420 0.4907

Vehicle Density 2 2.490 0.2880

Vehicle Make/Series 11 316.960 < 0.0001

Rated Driver Gender 2 0.190 0.9075

ABS 1 0.070 0.7953

Detailed results of the regression analysis using motorcycle collision claim severity as the dependent variable are listedin Table 4. The structure of the table, as well as the variables and reference categories, are the same as those used forclaim frequency in Table 2. The variables and reference categories that were used for claim frequency were used for claimseverity. The exponent of the intercept equals $9,089. The intercept outlines losses for the reference (baseline) categories:the estimate corresponds to the claim severity for a new Honda Gold Wing without ABS, garaged in a medium vehicledensity area, and driven by a male age 25-39.

The estimate corresponding to the ABS effect, a 2 percent decrease in claim severity, was highly nonsignificant (p=0.8),indicating ABS does not affect claim severity. As previously mentioned, vehicle age and make/series were significant pre-dictors of claim severity. Not surprisingly, as motorcycles age their claim severities decrease. The model estimated a 6 per-cent decrease (p=0.0004) in claim severity per 1-year increase in vehicle age. Estimated claim severities for the 11make/series motorcycles, compared with those for the Honda Gold Wing (reference category), ranged from 21 percentlower for the Honda ST1300 to 75 percent lower for the Honda Reflex. All of the make/series estimates were significantat the p<0.05 level.

INTERCEPT 9.1148 9088.817 0.0849 11515.1 <0.0001VEHICLE AGE -0.0608 0.941 0.0172 12.57 0.0004RATED DRIVER AGEUnknown 0.0819 1.085 0.0838 0.95 0.328614-24 0.1743 1.190 0.1029 2.87 0.090525-39 0 1.000 040-64 0.0624 1.064 0.0586 1.14 0.286665+ 0.0415 1.042 0.0866 0.23 0.6317

VEHICLE DENSITY0-99 -0.0342 0.966 0.0563 0.37 0.5432100-499 0 1.000 0500+ 0.0553 1.057 0.0514 1.16 0.2822

RATED DRIVER GENDERFemale 0.0319 1.032 0.0728 0.19 0.6613Male 0 1.000 0Unknown 0.0078 1.008 0.0501 0.02 0.8766

VEHICLE MAKE/SERIESHonda Gold Wing 0 1.000 0Honda Interceptor 800 -0.5373 0.584 0.0928 33.54 <0.0001Honda Reflex -1.3808 0.251 0.0916 227.27 <0.0001Honda Silver Wing -1.1479 0.317 0.0922 154.93 <0.0001Honda ST1300 -0.2339 0.791 0.0989 5.59 0.018Suzuki Bandit 1250 -0.8037 0.448 0.1774 20.53 <0.0001Suzuki Burgman 650 -0.9265 0.396 0.0908 104.19 <0.0001Suzuki SV650 -0.9031 0.405 0.0801 127.02 <0.0001Suzuki V-Strom 650 -0.989 0.372 0.1179 70.33 <0.0001Triumph Sprint ST -0.5166 0.597 0.1154 20.03 <0.0001Triumph Tiger -0.5712 0.565 0.1755 10.59 0.0011Yamaha FJR1300 -0.4631 0.629 0.0863 28.81 <0.0001

ABSABS Model -0.0155 0.985 0.0597 0.07 0.795Non-ABS Model 0 1.000 0

EXPONENT STANDARD CHI-PARAMETER ESTIMATE (ESTIMATE) ERROR SQUARE P-VALUE

TABLE 4 DETAILED RESULTS OF LINEAR REGRESSION ANALYSISOF COLLISION CLAIM SEVERITIES

Table 5 summarizes the effects of the independent variables on motorcycle collision overall losses, derived from the claimfrequency and severity models. Overall losses can be calculated by simple multiplication because the estimates for theeffect of ABS on claim frequency and severity were in the form of ratios relative to the reference (baseline) categories. Thestandard error for overall losses can be calculated by taking the square root of the sum of the squared standard errors forclaim frequency and severity. Based on the value of the estimate and the associated standard error, the level of statisticalsignificance (p-value) can be obtained from a probability distribution table.

The estimated effect of ABS was a 21 percent decrease in collision overall losses that was significant (p=0.01). This is astrong indication that ABS is effective in reducing collision overall losses for motorcycles.

Motorcycle collision overall losses were predicted to increase with increased vehicle density. Estimated overall losses inhigh vehicle density areas were 19 percent higher (p=0.01) than those medium vehicle density in areas (reference cate-gory), whereas estimated overall losses in low vehicle density areas were 15 percent lower (p=0.04). Estimated overalllosses for the 11 make/series motorcycles, compared with those for the Honda Gold Wing (reference category), rangedfrom 66 percent lower for the Suzuki V-Strom 650 to 91 percent higher for the Suzuki Bandit 1250. Only about half ofthe make/series estimates were statistically different from the reference category.

Driver age was a significant predictor of motorcycle collision overall losses. Estimated overall losses for rated drivers 24and younger were 113 percent higher (p<0.0001) than those for rated drivers ages 25-39 (reference category). Rated driv-ers ages 40-64 and 65 and older had slightly lower estimated overall losses, but these estimates did not reach statisticalsignificance. Vehicle age also was significantly predictive of collision overall losses. Overall losses decreased by an esti-mated 28 percent (p<0.0001) for each 1-year increase in vehicle age.

INTERCEPT -9.4014 0.0826 9.1148 0.0849 -0.2866 0.1185 0.7508 0.0155VEHICLE AGE -0.2628 0.0192 -0.0608 0.0172 -0.3236 0.0258 0.7235 <0.0001RATED DRIVER AGEUnknown 0.1791 0.0925 0.0819 0.0838 0.2610 0.1248 1.2982 0.03651. 14-24 0.5839 0.094 0.1743 0.1029 0.7582 0.1394 2.1344 <0.000125-39 0 0 0 0 0 0 1.000040-64 -0.1433 0.0649 0.0624 0.0586 -0.0809 0.0874 0.9223 0.354965+ -0.0747 0.1026 0.0415 0.0866 -0.0332 0.1343 0.9673 0.8047

DENSITY0-99 -0.1332 0.0582 -0.0342 0.0563 -0.1674 0.0810 0.8459 0.0387100-499 0 0 0 0 0 0 1.0000500+ 0.122 0.0514 0.0553 0.0514 0.1773 0.0727 1.1940 0.0147

RATED DRIVER GENDERFemale -0.205 0.0845 0.0319 0.0728 -0.1731 0.1115 0.8411 0.1207Male 0 0 0 0 0 0 1.0000Unknown 0.0431 0.0527 0.0078 0.0501 0.0509 0.0727 1.0522 0.4839

VEHICLE MAKE/SERIESHonda Gold Wing 0 0 0 0 0 0 1.0000Honda Interceptor 800 0.8988 0.1077 -0.5373 0.0928 0.3615 0.1422 1.4355 0.0110Honda Reflex 0.5596 0.1093 -1.3808 0.0916 -0.8212 0.1426 0.4399 <0.0001Honda Silver Wing 0.7658 0.1025 -1.1479 0.0922 -0.3821 0.1379 0.6824 0.0056Honda ST1300 0.1745 0.1128 -0.2339 0.0989 -0.0594 0.1500 0.9423 0.6921Suzuki Bandit 1250 1.4525 0.2295 -0.8037 0.1774 0.6488 0.2901 1.9132 0.0253Suzuki Burgman 650 0.7284 0.0954 -0.9265 0.0908 -0.1981 0.1317 0.8203 0.1325Suzuki SV650 1.0251 0.0747 -0.9031 0.0801 0.1220 0.1095 1.1298 0.2653Suzuki V-Strom 650 -0.084 0.1352 -0.989 0.1179 -1.0730 0.1794 0.3420 <0.0001Triumph Sprint ST 1.077 0.1357 -0.5166 0.1154 0.5604 0.1781 1.7514 0.0017Triumph Tiger 0.3632 0.2177 -0.5712 0.1755 -0.2080 0.2796 0.8122 0.4570Yamaha FJR1300 0.3276 0.0949 -0.4631 0.0863 -0.1355 0.1283 0.8733 0.2908

ABSABS Model -0.2151 0.0709 -0.0155 0.0597 -0.2306 0.0927 0.7941 0.0128Non-ABS Model 0 0 0 0 0 0 1.0000

FREQUENCY SEVERITY OVERALL LOSSESSTANDARD STANDARD STANDARD EXPONENT

PARAMETER ESTIMATE ERROR ESTIMATE ERROR ESTIMATE ERROR (ESTIMATE) P-VALUE

TABLE 5 RESULTS FOR COLLISION OVERALL LOSSES DERIVED FROM CLAIM FREQUENCY AND SEVERITY MODELS

VEHICLE AGE0 291 0%1 12,177 17%2 20,388 28%3 18,147 25%4 12,748 18%5 8,724 12%

RATED DRIVER AGEUnknown 4,886 7%14-24 1,696 2%25-39 11,274 16%40-64 47,687 66%65+ 6,933 10%

VEHICLE DENSITY0-99 21,281 29%100-499 29,331 40%500+ 21,863 30%

RATED DRIVER GENDERFemale 7,034 10%Male 43,147 59%Unknown 22,294 31%

HIGHWAY LOSSDATA INSTITUTE1005 North Glebe RoadArlington, VA 22201

The Highway Loss Data Institute is a nonprofit public service organization that gathers, processes, and publishes insurance data on the human and economic lossesassociated with owning and operating motor vehicles.

COPYRIGHTED DOCUMENT, DISTRIBUTION RESTRICTED © 2008 by the Highway Loss Data Institute, 1005 N. Glebe Road, Arlington, VA 22201. All rightsreserved. Distribution of this report is restricted. No part of this publication may be reproduced, or stored in a retrieval system, or transmitted, in any form or byany means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the copyright owner. Possession of this publicationdoes not confer the right to print, reprint, publish, copy, sell, file, or use this material in any manner without the written permission of the copyright owner.Permission is hereby granted to companies that are supporters of the Highway Loss Data Institute to reprint, copy, or otherwise use this material for their own businesspurposes, provided that the copyright notice is clearly visible on the material.

APPENDIX A DISTRIBUTION OF EXPOSURE FOR INDEPENDENT VARIABLES

EXPOSURE PERCENT OF TOTAL

EXPOSURE PERCENT EXPOSURE PERCENTVEHICLE MAKE/SERIES WITHOUT ABS OF SERIES WITH ABS OF SERIES

Honda Gold Wing (Touring) 28,904 82% 6,366 18%Honda Interceptor 800 (Sport) 1,925 76% 623 24%Honda Reflex (Scooter) 3,036 87% 443 13%Honda Silver Wing (Scooter) 2,882 85% 506 15%Honda ST1300 (Sport) 3,118 68% 1,468 32%Suzuki Bandit 1250 (Standard) 136 80% 33 20%Suzuki Burgman 650 (Scooter) 3,219 92% 288 8%Suzuki SV650 (Unclad Sport) 8,468 100% 11 0%Suzuki V-Strom 650 (Dual Purpose) 2,986 97% 90 3%Triumph Sprint ST (Sport) 908 82% 198 18%Triumph Tiger (Dual Purpose) 695 94% 42 6%Yamaha FJR1300 (Sport) 3,610 59% 2,520 41%Total 59,886 83% 12,589 17%

Effectiveness of Antilock Braking Systems in Reducing Fatal Motorcycle Crashes

Eric R. Teoh

October 2008

ABSTRACT

The effect of antilock braking systems (ABS) on motorcyclist fatal crash risk in 2005-06 was

studied by comparing fatal crash rates per registrations of motorcycles with and without ABS. Study

motorcycles included those for which ABS was optional equipment and could be identified as present by

the model name. Fatal motorcycle crashes per 10,000 registered vehicle years were 38 percent lower for

ABS models than for their non-ABS versions.

INTRODUCTION

Annual motorcyclist deaths in the United States have more than doubled, from 2,077 in 1997 to

4,697 in 2006 (Insurance Institute for Highway Safety, 2008), and motorcycle registrations have increased

by about two-thirds, from 5,167,693 in 2000 (earliest year for which data are available) to 8,642,243 in

2006, according to data from R.L. Polk and Company. Many factors contribute to motorcycle crashes,

but improper braking was identified as a major pre-impact factor in a study of motorcycle crash causation

(Hurt et al., 1981) and again, 20 years later, in the Motorcycle Accident In-Depth Study (MAIDS)

(Association of European Motorcycle Manufacturers, 2004).

Operating the brakes on most motorcycles is much more complicated than on four-wheel

vehicles. Most motorcycles have separate controls for the front and rear brakes, with the front brake

usually controlled by a lever on the right handlebar and the rear brake controlled by a pedal operated by

the rider’s right foot. During braking, a rider must decide how much force to apply to each control. As

with other types of vehicles, much more deceleration can be obtained from braking the front wheel than

from braking the rear wheel.

Motorcycles are inherently less stable than four-wheel vehicles and rely on riders’ skills to remain

upright during extreme maneuvers such as hard braking. Braking too hard and locking a wheel creates an

unstable situation. Locking the front wheel is particularly dangerous, with loss of control being almost

certain. A locked rear wheel is more controllable but still can lead to loss of control if the rider

simultaneously tries to steer the motorcycle, as in an emergency avoidance maneuver. However, in an

emergency requiring full stopping power, riders concerned about wheel lock may be reluctant to apply

full force to the brakes, particularly to the front brake. Both Hurt et al. (1981) and MAIDS (Association

of European Motorcycle Manufacturers, 2004) had examples of both loss of control due to wheel lock and

failure to adequately brake.

Although proper braking practices can be taught, rider training courses have not been shown to be

effective in reducing motorcycle crashes (Mayhew and Simpson, 1996) or have provided mixed results at

best (Billheimer, 1998). In an effort to address the issue of under-braking (especially the front wheel),

manufacturers have developed braking systems that essentially link the front and rear brake controls.

These systems, collectively referred to here as combined braking systems (CBS), apply braking force to

1

both wheels when either control is engaged. The degree to which braking force is applied to the front

wheel, for example, when the pedal for the rear brake is pressed varies by design, but the concept is the

same. CBS has been shown to reduce stopping distances of experienced riders on closed test tracks

(Green, 2006) and would be expected to be beneficial in situations in which a rider under-brakes (or does

not brake) the front wheel to avoid locking it or causing the motorcycle to pitch forward. Even with CBS,

however, it still is possible to lock a wheel during hard braking, often with catastrophic consequences.

ABS has been developed to help riders solve this dilemma. The system monitors wheel speed

and reduces brake pressure when impending wheel-lock is detected. Brake pressure is increased when

traction is restored, and the system evaluates and adjusts brake pressure many times per second. These

systems allow riders to apply brakes fully in an emergency without fear of wheel-lock. ABS was first

developed for commercial aircraft in 1929 (Maslen, 2008) and was first implemented in production

automobiles with the 1971 Chrysler Imperial (Douglas and Schafer, 1971). BMW was the first

manufacturer to implement ABS on a motorcycle with its K100RS Special model in 1988 (Tuttle, 2001).

ABS and CBS are not necessarily related; either or both can be implemented on a motorcycle.

ABS has not significantly reduced crash risk for passenger vehicles (Farmer et al., 1997; Farmer,

2001), but there is reason to expect ABS will be more helpful to motorcycles because of the instability

that occurs with any wheel-lock. Studies conducted on closed test tracks have demonstrated that ABS can

reduce the motorcycle stopping distances (Green, 2006; Vavryn and Winkelbauer, 2004). It is clear that

reducing wheel-lock is crucial in maintaining stability during hard braking. These results suggest that

ABS has the potential to reduce motorcycle crashes in real-world situations. Serious motorcycle crashes

identified from insurance claims were analyzed in a small study to determine, by crash reconstruction,

how certain crashes would be affected by ABS (Allianz Center for Technology, 2005). About half of the

200 crashes studied were deemed to be relevant to ABS, and the majority of those involved another

vehicle violating a motorcyclist’s right-of-way. Crash reconstruction analyses showed that between 8 and

17 percent of these crashes could have been avoided had the motorcycles been equipped with ABS. No

results were presented on how increased stability or stopping power provided by ABS might have

decreased the severities of the crashes that were deemed inevitable.

A study by the Highway Loss Data Institute (HLDI), conducted in conjunction with the present

study, found that motorcycles equipped with optional ABS had 19 percent fewer insurance claims for

collision damage per insured vehicle year than the same motorcycle models without ABS (Moore and

Yan, in process). The goal of the present study was to evaluate the effectiveness of ABS in reducing the

rate of fatal motorcycle crashes on public roads in the United States. Specifically, rates of fatal crash

involvement per registered vehicle were compared for motorcycle models with and without ABS installed

as optional equipment.

2

METHODS

Data on fatal motorcycle crashes were extracted from the Fatality Analysis Reporting System

(FARS), a national census of fatal crashes occurring on public roads that is maintained by the National

Highway Traffic Safety Administration. Exposure data consisted of national motorcycle registration

records obtained from R.L. Polk and Company. Each vehicle record in both databases was indexed by its

vehicle identification number (VIN), which encodes vehicle information, and the first 10 digits of the

VINs were used to determine make, model name, and model year according to records in a motorcycle

features database created and maintained by HLDI. Vehicles with missing or invalid VINs were

excluded.

To be included in the study, a motorcycle model was required to have ABS as an option and the

presence of that option must have been discernable directly from the model name (e.g., Honda Gold Wing

vs. Honda Gold Wing ABS). This eliminated bias due to the comparison of different makes or styles of

motorcycles. Although ABS has been an option on BMW models for much longer than the study period,

the BMW systems cannot be identified from the model name alone. All BMW models were excluded.

The final study population (Table 1) included eight make/model motorcycles, each with both ABS and

non-ABS versions. Some vehicles were excluded due to zero registrations of the ABS model during the

study years in the Polk records. Because none of the study vehicles with ABS were available in model

year 2000 or earlier, the analysis was restricted to 2001 or later model year vehicles. Among the

motorcycles included, all of the Hondas (both ABS and non-ABS) were equipped with standard CBS;

CBS was not available on any of the others.

Table 1

Study motorcycles Non-ABS motorcycles ABS motorcycles Model years Make/model Model years Make/model 2001-06 Honda Gold Wing 2001-06 Honda Gold Wing ABS 2001-06 Honda Interceptor 800 2002-06 Honda Interceptor 800 ABS 2001-06 Honda Reflex 2001-06 Honda Reflex ABS 2003-06 Honda ST1300 2003-06 Honda ST1300 ABS 2002-06 Honda Silver Wing 2003-06 Honda Silver Wing ABS 2003-06 Suzuki Burgman 650 2006 Suzuki Burgman 650 ABS 2001-06 Triumph Sprint ST 2006 Triumph Sprint ST ABS 2003-05 Yamaha FJR1300 2004-06 Yamaha FJR1300 ABS

At the time this study was conducted, registration data were available only for 2000 and 2005-07,

and FARS data were not yet available for 2007. There were no registrations of the ABS versions of these

motorcycles in 2000. Therefore, data were analyzed for years 2005-06. Fatal crash rates per 10,000

registered vehicle years for each motorcycle model, both ABS and non-ABS versions, were calculated by

dividing 10,000 times the number of motorcyclist fatal crash involvements in 2005-06 by the number of

3

motorcycles registered during these years. Because registration counts spanned 2 years, the denominator

was interpreted as registered vehicle years instead of registrations.

Fatal crash rates per vehicle registrations for ABS and non-ABS motorcycle models were

compared by calculating a rate ratio (RR) equal to the crash rate for ABS models divided by the crash rate

for non-ABS models. If ABS has no effect, then the rate ratio should be 1.0. A rate ratio of less than 1.0

would indicate the fatal crash rate for ABS models is lower than the rate for non-ABS models. Similarly,

a rate ratio greater than 1.0 would indicate the fatal crash rate for ABS models is higher than the rate for

non-ABS models. One way to calculate the rate ratio is to calculate the rate for non-ABS motorcycles as

total crash involvements divided by total registered vehicle years, and analogously for ABS motorcycles.

However, to reduce any bias that may have occurred from relative differences in registrations among

motorcycle models, an alternative rate for non-ABS motorcycles was calculated as the weighted average

of fatal crash rates for each vehicle model, where the weights were taken as the number of registered

vehicle years of ABS-equipped motorcycles. Thus, for any given motorcycle model the ABS and non-

ABS fatal crash rates received the same weight in calculating the overall fatal crash rates for motorcycles

with and without ABS.

Ninety and 95 percent confidence intervals (CIs) for the rate ratios were calculated based on

standard error estimates derived by assuming that the number of fatal crash involvements follows a

Poisson distribution as follows:

Var(ln(RR)) = VarABS + Var on-ABS n

where VarABS = ∑ X ,ABS81

and Varnon-ABS = ∑ r X ,non‐ABS81

∑ r X81

with Xi,ABS as the number of fatal crash involvements of the ith model motorcycle

equipped with ABS, analogously for Xi,non-ABS, and ri as the ratio of registered vehicle

years of the ABS model to registered vehicle years of the non-ABS model for model i.

lower 95% confidence limit = exp[ln(RR) – 1.96(Var(ln(RR)))1/2]

upper 95% confidence limit = exp[ln(RR) + 1.96(Var(ln(RR)))1/2]

RESULTS

Table 2 presents fatal crash involvements, registered vehicle years, and the rate of fatal crash

involvements per 10,000 registered vehicle years for the study motorcycles during 2005-06. Motorcycles

manufactured by Honda, particularly the Gold Wing model, dominated the sample, but the pattern for all

but two of the motorcycles was a lower fatal crash rate for ABS-equipped motorcycles. Across all ABS-

equipped motorcycles, the rate of fatal crash involvements per 10,000 registered vehicle years was 4.1,

compared with 6.7 for the same motorcycles not equipped with ABS.

4

Table 2Motorcycle fatal crash involvements and registered vehicle years, 2005-06

Non-ABS models ABS models

Fatal crash

involvements Registered

vehicle years Rate

(x 104) Fatal crash

involvements Registered

vehicle years Rate

(x 104) Honda Gold Wing 63 93,608 6.7 6 19,547 3.1 Honda Interceptor 800 8 10,437 7.7 3 2,307 13.0 Honda Reflex 6 14,858 4.0 1 2,644 3.8 Honda ST1300 2 7,003 2.9 2 3,580 5.6 Honda Silver Wing 11 12,273 9.0 1 1,278 7.8 Suzuki Burgman 650 8 9,618 8.3 0 309 0.0 Triumph Sprint ST 1 4,476 2.2 0 135 0.0 Yamaha FJR1300 8 8,734 9.2 2 6,486 3.1

Total 107 161,007 6.7* 15 36,286 4.1 *Overall non-ABS rate is weighted by registered vehicle years of ABS-equipped motorcycles.

The effect of ABS on fatal crash involvement is given by the rate ratio estimate for ABS-

equipped motorcycles against non-ABS motorcycles. This estimate and associated 90 and 95 percent

confidence intervals are provided in Table 3. The rate ratio estimate corresponds to an approximate 38

percent reduction (computed as (RR-1)×100%) in the rate of fatal crash involvements per 10,000

registered vehicle years for the ABS models over the (weighted) non-ABS models.

Table 3

Estimated rate ratios and confidence intervals for those estimates for comparing ABS and non-ABS fatal crash rates Rate ratio 0.615 95% confidence interval (0.352, 1.074) 90% confidence interval (0.385, 0.982)

DISCUSSION

Results of this analysis provide evidence that ABS is effective in reducing fatal motorcycle

crashes. Study motorcycles with ABS had a fatal crash involvement rate 38 percent lower than that for

their non-ABS versions during the study years.

Although the estimated effect of 38 percent is large, it is not statistically significant at the

customary 0.05 level. ABS is a relatively recent option on motorcycles, and the option was purchased on

only 18 percent of registered study vehicles during 2005-06. More data are required to obtain a more

precise estimate of ABS effectiveness in reducing fatal motorcycle crashes. However, as the estimate

becomes more precise, it is quite likely that it will continue to indicate a benefit of ABS. If there were no

effect of ABS on fatal crash involvement, an estimate as large as the 38 percent reduction in this study

would be expected to occur by chance less than 10 percent of the time. Thus, there is considerable

confidence that ABS is preventing fatal crashes among motorcyclists. This confidence is bolstered by the

fact that a separate analysis of insurance collision coverage losses among crashes of all severities also

shows a reduction in crashes of about 19 percent for motorcycles equipped with ABS (Moore and Yan,

5

2008). These results provide confirmatory evidence of the expected benefit of ABS from engineering

principles, test-track trials, and a crash reconstruction analysis.

The substantial effectiveness estimate observed in this study is not, however, without limitations.

ABS was studied as optional equipment, so the cohort of motorcyclists who choose to purchase ABS may

differ from those who decline to purchase it. In particular, motorcyclists who choose ABS may be more

concerned about safety than those who decline, thus leading to lower fatal crash rates due to safer riding

practices. Because of the small sample of ABS-equipped motorcycles, it was not possible to carefully

examine how rider factors such as helmet use and speeding differ between the two groups. For example,

78 percent of the non-ABS riders were helmeted, compared with 60 percent (9 of 15) of the ABS riders.

Also, the prevalence of these factors is not known for riders of study motorcycles that were not involved

in fatal crashes. Therefore, it was not possible to study how such factors influenced the observed

reduction in fatal crash rate for ABS-equipped motorcycles. Aside from differences in rider factors, it is

also possible that riders who choose ABS accumulate more miles than those who decline, which would

result in an upward bias in the fatal crash rate for the ABS cohort relative to the non-ABS cohort. As may

have occurred in passenger vehicles (Grant and Smiley, 1993), motorcyclists may tend to drive ABS

motorcycles more aggressively than non-ABS motorcycles, also resulting in a higher than expected crash

rate for the ABS group. Without more extensive data, it is not possible to know the magnitude or nature

of any bias of the estimated rate ratio comparing crash rates for ABS and non-ABS motorcycles.

With or without ABS, CBS also may reduce the likelihood of certain types of crashes. However,

due to the small sample of non-CBS motorcycles in this study, the effect of CBS could not be evaluated.

Still, CBS is not expected to bias the results because the braking systems of the ABS and non-ABS

motorcycles differed only by whether or not they were equipped with ABS. In other words, each ABS/non-

ABS pair either did or did not have CBS. ABS showed a benefit in both the CBS and non-CBS groups,

suggesting the presence of CBS on some of the motorcycles did not confound the observed effect of ABS.

ABS cannot be expected to prevent or mitigate all types of crashes, as demonstrated in the

Allianz Center for Technology (2005) study. For example, ABS would not affect the outcome or

likelihood of a crash involving a motorcycle struck from behind. The small sample of ABS motorcycles

in FARS and the lack of detailed information on precrash events in FARS precluded examination of the

effects of ABS on crashes that would likely be influenced by its presence.

ACKNOWLEDGMENTS

Motorcycle VINs were decoded by Marvin Campbell of the Highway Loss Data Institute, which

also maintains the database used to decode these VINs. Adrian Lund, Anne McCartt, and Charles Farmer

of the Insurance Institute for Highway Safety contributed helpful comments and suggestions. This work

was supported by the Insurance Institute for Highway Safety.

6

7

REFERENCES

Allianz Center for Technology. (2005) AZT Accident Research on Effectiveness of ABS in Motorbikes. Presented at 2005 RCAR Conference,

Association of European Motorcycle Manufacturers. 2004. MAIDS in-depth investigations of accidents involving powered two wheelers. Final Report 1.2. Brussels, Belgium

Billheimer, J.W. 1998. Evaluation of the California motorcyclist safety program. Transportation Research Record 1640:100-09.

Douglas, J. and Schafer, T. 1971. The Chrysler ‘Sure-Brake:’ The first production four-wheel anti-skid system. SAE Technical Paper Series 710248. Warrendale, PA: Society of Automotive Engineers.

Farmer, C.M. 2001. New evidence concerning fatal crashes of passenger vehicles before and after adding antilock braking systems. Accident Analysis and Prevention 33:361-69.

Farmer, C.M.; Lund, A.K.; Trempel, R.E.; and Braver, E.R. 1997. Fatal crashes of passenger vehicles before and after adding antilock braking systems. Accident Analysis and Prevention 29:745-57.

Grant, B. and Smiley, A. 1993. Driver response to antilock brakes: a demonstration of behavioural adaptation. Proceedings of the Canadian Multidisciplinary Road Safety Conference VIII. Ottawa, Ontario: Transport Canada.

Green, D. 2006. A comparison of stopping distance performance for motorcycles equipped with ABS, CBS and conventional hydraulic brake systems. Presented at the 2006 International Motorcycle Safety Conference, Long Beach, CA.

Hurt, H.H.; Ouellet, J.V.; and Thom, D.R. 1981. Motorcycle accident cause factors and identification of countermeasures; Volume 1: technical report. Report no DOT HS-805-862. Washington, DC: National Highway Traffic Safety Administration.

Insurance Institute for Highway Safety. 2008. Fatality facts, 2006: motorcycles. Arlington, VA. Available: http://www.iihs.org/research/fatality_facts_2006/motorcycles.html. Accessed: July 21, 2008.

Maslen, J. (Spring 2008) ABS: 30 years of life saving. Roadsafe Magazine. Available: http://www.roadsafe.com/magazine/2008spring/abs.htm. Accessed: July 23, 2008.

Mayhew, D.R. and Simpson, H.M. 1996. Effectiveness and role of driver education and training in a graduated licensing system. Ottawa, Ontario: Traffic Injury Research Foundation.

Moore, M. and Yan, Y. In process. Antilock braking systems for motorcycles and insurance collision losses. Arlington, VA: Highway Loss Data Institute.

Tuttle, M. 2001. One track mind. Rider Magazine July;28(7), p. 8.

Vavryn, K. and Winkelbauer, M. 2004. Braking performance of experienced and novice motorcycle riders – results of a field study. Presented at the 2004 International Conference on Transport and Traffic Psychology, Nottingham, United Kingdom.