The Case Against Dog Breed Discrimination 2 by Homeowners’ Insurance … · 2019-12-16 ·...

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies

Larry Cunningham

2CHAPTER

In spring 2003 I moved from Vir-ginia to Texas to begin work as atenure-track faculty member at

Texas Tech University School of Law.I brought my two dogs with me: Saffy(a four-year-old mixed breed whoseparents were a fluffy red Chow Chowand a big black Labrador retriever)and Semona (a two-year-old rot-tweiler). Neither Semona nor Saffyhas ever bitten anyone. Neither hasshown any aggressive tendencies.Both are extremely playful andfriendly animals.

After I placed a bid on a house inLubbock, Texas, I began the searchfor homeowners’ insurance—aprocess that I thought would bestraightforward and easy. Much tomy surprise, dozens of insurancecompanies denied my applicationoutright. The reason? Semona is arottweiler and Saffy is half-Chow.Rottweilers and Chow Chows are onthe “blacklist” of dog breeds. Someinsurance companies believe they,along with pit bulls, huskies, Dober-man pinschers, and other specifiedbreeds, are more likely to bitehumans and, in turn, cause liabilityclaims to be brought against theirowners. Even mixed breeds, like myhalf-Chow, Saffy, are blacklisted.This practice is known by many dogowners as “breed discrimination.”

Thankfully, the story ended hap-pily for my dogs and me. After weeksof calling nearly every insuranceagent in Lubbock, I was able toobtain insurance through the TexasFarm Bureau, an organization thatadvocates for farmers and farmingissues.1 Had it not been for the FarmBureau, I would have found myselfon the horns of a horrible dilemma:whether to buy a home or give upmy dogs. Anyone who knows me canconfirm that this dilemma wouldhave been easy to resolve; I wouldhave chosen my furry family mem-bers over home ownership.2 Sadly,however, many Americans are find-ing themselves in similar positionsand are opting to give up their dogsto animal shelters.3

Breed discrimination by insur-ance companies is on the rise in theUnited States. Insurers are refusingto write homeowners’ policies forpeople who own breeds that theinsurance industry considers to bedangerous. Their decisions arebased solely on the breed of the ani-mal, not the individual characteris-tics of the particular dog. Dog bitesare certainly a public health con-cern. However, the insurance indus-try’s approach to the problem isbased on faulty assumptions andimproper use of dog-bite statistics.

The insurance industry has pre-judged entire breeds of dogs asbeing “too risky,” instead of takinga more reasonable dog-by-dogapproach to risk assessment.

Major veterinary and breed reg-istry organizations have stronglyopposed breed discrimination ininsurance. Authors of scientificstudies on dog bites have evenargued against the use of theirdata to support breed-based deci-sion-making by insurers and legis-latures. Dog owners across thecountry have spoken out about thehorrible choice they have beenforced to make between obtaininginsurance and keeping their dogs.

There has existed a historic ten-sion between risk classification andsocial policy. Classification andinsurability decisions are usually“actuarially justified”—that is, theinsurance company has identified astatistical correlation between acharacteristic and increased risk.Actuarial justification is frequentlycited by insurers as a reason toavoid social regulation. Insurersexist to make a profit for theirshareholders. They do so by mini-mizing risk, which, in turn, mini-mizes claims paid out.

Actuarial justification is only thefirst step in determining the social

This essay was originally published in the Connecticut Insurance Law Journal (Vol. ll, No. l, 2004–2005). Theviews expressed in this essay are the author’s own.

25

26 The State of the Animals IV: 2007

propriety of a proposed underwrit-ing mechanism. Social utility of therisky conduct must also be consid-ered. Statutes across the UnitedStates are replete with examples oflegislatures overruling actuariallyjustified practices in favor of com-peting social policies. “Red-lining”is a classic example. Actuaries iden-tified statistical correlations be-tween living in certain neighbor-hoods and increased risk for claimsagainst homeowners’ policies. As aresult, insurance companies beganto refuse to write policies in thesehigh-risk neighborhoods. Theneighborhoods in question wereoften economically depressed andoccupied by members of racial orethnic minorities. Legislatures andcourts stepped in to prohibit red-lining, despite the actuarial justifi-cation for the practice.4

Breed discrimination is a differ-ent animal altogether. Even with-out considering the high socialutility of pet ownership, insurershave been unable to demonstratean actuarial justification for dis-criminating based on breed. Asthe multidisciplinary Task Forceon Canine Aggression and Human-Canine Interactions concluded,“[D]og bite statistics are notreally statistics, and they do notgive an accurate picture of dogsthat bite.”5 The popular notionthat pit bulls and rottweilers areinherently more likely to bite issimply not supported by the avail-able statistics.

When the social utility of pets isadded to the equation, breed dis-crimination becomes even moreunreasonable. Dogs and other do-mesticated animals provide im-measurable joy and happiness tothe families that own them. Evensome components of the legal sys-tem itself have evolved to recog-nize pets as being more than merechattel.6 In addition, the failure toobtain homeowners’ insurance is adeath knell for homeownership—no insurance, no mortgage; nomortgage, no house.

My argument is quite simple:decisions regarding the provision,rating, termination, or renewal of ahomeowners’ insurance policyshould not be based on ownershipor possession of a particular breedof dog unless there is evidence ofdog-specific risk. Insurers wouldconcededly be actuarially justifiedin charging higher premiums ordeclining coverage for people whoown dogs that have unjustly bittenin the past. After all, the best pre-dictor of future behavior is past be-havior. Breed discrimination, as itcurrently stands, is not actuariallyjustified because scientists havenot been able to accurately deter-mine whether certain breeds areinherently more dangerous, or,instead, whether a breed’s highpopulation is making it appearthat the breed is more dangerous.

The consequences of breed dis-crimination could not be greater.Homeowners’ insurance is thegatekeeper to homeownership.Without homeowners’ insurance, abuyer cannot get a mortgage. Formost Americans, if a person can-not obtain a mortgage, he cannotbuy a home.

In Part I of this article, I give anoverview of the problem: dog breeddiscrimination by insurers, as wellas a related problem of breed-spe-cific legislation by some states. InPart II, I analyze the major scien-tific studies on dog bites, showingthat no one has adequately proventhat some breeds are more inher-ently dangerous than others. InPart III, I show that breed discrim-ination and breed-specific legisla-tion are opposed by most veteri-nar y and animal protectiongroups. Part IV demonstrates thatinsurers have been ignoring theunique and special role that petsplay in millions of Americanhomes. I draw upon not only theprofoundly personal argumentsadvanced by myself and others, butalso the way in which the law itselfis evolving by recognizing pets asmore than mere property. Part V

shows how the insurance industryis a highly regulated industry thatsubjects itself to legislative controlwhere, as here, the public is beingharmed by underwriting decisionsnot driven by actuarial justifica-tion. I also offer a number of alter-natives to breed discrimination.

I. Dog BreedDiscriminationBreed discrimination in insuranceis a recent phenomenon that waspreceded by the enactment of“breed-specific legislation” (BSL)by some state legislatures andmunicipalities. Both breed dis-crimination and BSL are a per-ceived response to highly publi-cized attacks by certain breeds,particularly pit bulls.

Highly PublicizedAttacks by Pit BullsIn the 1980s there were a numberof high-profile attacks on humansby pit bulls. These attacks led to anear-hysterical reaction by mem-bers of the communities thatwere affected by the attacks andby the legislators who repre-sented them.

I n M a rc h 1 9 8 4 , p i t b u l l sattacked Angie Hands, a nine-year-old girl in Tijeras, New Mexico.7The dogs bit her right leg to thebone, ripped flesh from her arms,and tore her ear in half.8 The childsurvived but had to undergo yearsof reconstructive surgery.9 She hadbeen attacked by her uncle’s fourpit bulls in between her bus stopand her home.10 The small commu-nity of Tijeras, located outside ofAlbuquerque, responded with anoutright ban on pit bull owner-ship.11 Dog owners challenged thelaw in court, but the law wasupheld as a constitutional exerciseof the town’s police power.12

27

The attack on Angie Hands fol-lowed a number of other pit bullattacks around the country. A four-year-old girl was killed in OregonCity, Oregon, when she fell into ayard where a pit bull was chained.13

Two pit bulls mutilated their ownerin Edgemere, Maryland.14 A re-cent, widely publicized attack inSan Francisco has also brought theissue of aggressive dogs to the fore-front of public attention. In Janu-ar y 2001, Diana Whipple wasmauled to death by two PresaCanario dogs. The dogs wereowned by a pair of lawyers. Evi-dence at the owners’ murder trialsshowed that the dogs had tried toattack other people and animalsin the past. Both defendants wereconvicted and served prisontime.15 A subsequent civil lawsuitbrought by Ms. Whipple’s motherwas settled out-of-court.16

“Breed-SpecificLegislation”Highly publicized pit bull attacksin the 1980s led to knee-jerk reac-tions by many communities.17

Attacks led to editorials, which ledto public outrage, which led toswift and spontaneous legislativeaction that was based on neithergood science nor good law. BSLbegan to emerge in the 1980s andearly 1990s. These laws targetedspecific breeds for regulation or, insome cases, outright bans. BSL ison the rise in the United States.States and municipalities acrossthe country have considered—and,in some cases, enacted—breed-specific legislation designed toprotect the public against dogbites.18 Commonly, these statutesand ordinances have banned, orplaced restrictions on, pit bulls,rottweilers, Doberman pinschers,Chow Chows, German shepherds,and shar-peis.19

Ohio has aggressively targetedpit bulls for regulation. Ohio lawdeclares any dog that “[b]elongs to

a breed that is commonly known asa pit bull dog”20 is automatically a“vicious dog.”21 “Vicious dogs”must be penned or tied up when ontheir owners’ premises.22 If off-premises, they must be tethered,caged, or muzzled.23 Owners mustobtain liability insurance to providecoverage in the event of a bite.24

BSL has also occurred at the localmunicipal level. Denver passed anoutright ban on the ownership, pos-session, keeping, control, mainte-nance, harboring, transportation, orsale of pit bulls.25 A “pit bull” isdefined as an American pit bull ter-rier, American Staffordshire terrier,Staffordshire bull terrier, or any dogdisplaying the majority of physicaltraits of one of those breeds.26 Thisordinance is in addition to Denver’s“dangerous dog” ordinance that reg-ulates “[a]ny dog with a knownpropensity or disposition to attack un-provoked, to cause injury or to other-wise endanger the safety of humansor other domestic animals.”27 “Dan-gerous dogs” must be confined whileat home and must be leashed andmuzzled while traveling.28

Not all states have followed theBSL trend. Some legislatures haveprohibited BSL enacted by munici-palities. Florida enacted a statutethat permits localities to regulatedogs “provided that no such regu-lation is specific to breed.”29 Somelegislators have attempted, with-out success, to repeal this anti-BSLstatute in response to severalhighly publicized attacks.30 Min-nesota also has the following pro-hibition against BSL:

A statutory or home rule char-ter city, or a county, may notadopt an ordinance regulatingdangerous or potentially dan-gerous dogs based solely onthe specific breed of the dog.31

Court challenges to BSL havebeen largely unsuccessful.32 Oppo-nents of BSL have brought lawsuitsclaiming the legislation is unconsti-tutional because it violates due pro-cess (substantive and procedural),the Takings Clause,33 equal protec-

tion, and the vagueness doctrine.34

Plaintiffs have challenged BSL ondue process grounds by arguingthat there was no “rational relation-ship” to a legitimate legislative goalor purpose.35 Courts have ruled thatBSL is a rational response to a per-ceived problem of dog bites by cer-tain breeds.36 They have alsorejected plaintiffs’ arguments thatthe statutes and ordinances do notprovide dog owners with sufficientnotice and an opportunity to beheard, which are the requirementsfor procedural due process.37 TheTijeras ordinance, for example, pro-vides that a pit bull may be de-stroyed by the village only after ahearing to determine whether thedog is, in fact, a pit bull.38 Plaintiffshave also contended that BSLamounts to a taking without justcompensation. Courts have rejectedthis argument, noting that personalproperty is subject to regulationunder the police power of a state.39

Challenges based on vagueness haveargued that identifying a dog’s breedis difficult.40 Most courts have foundBSL to be sufficiently specific toenable a reasonable dog owner todetermine if his or her dog is coveredby the particular statute.41 Plaintiffshave also alleged that BSL violatesequal protection by singling out pitbulls but not other breeds.42 Courtshave noted that pit bull ownership isnot a “suspect classification,” and,therefore, BSL need only have somereasonable basis to be constitutional.Courts have concluded that sufficientevidence exists to support a findingthat pit bulls can be regulated by leg-islatures and municipalities.43

One significant decision foundBSL to be unconstitutional. InAmerican Dog Owners Association,Inc. v. City of Lynn,44 the Massachu-setts Supreme Judicial Court up-held a trial court’s finding that theCity of Lynn’s attempt to regulatepit bulls was unconstitutional.45

The Court noted that it is particu-larly problematic to determine adog’s breed. The Court held,

[T]here is no scientific means,



by blood, enzyme, or other-wise, to determine whether adog belongs to a particularbreed, regardless of whether“breed” is used in a formalsense or not.46

The Court upheld the trial court’sfinding that animal-control officershad no real standards to identify pitbulls, in part because they had notraining in breed identification.47

The ordinance included a ban onmixed-breed dogs that contained“any mixture” of pit bull.48 This pro-vision was likewise found to be un-constitutional since it is scientifi-cally “impossible to ascertain”whether a dog is part pit bull.49 Theordinance was also unconstitutionalbecause it tried to define “pit bull”as including any breed where “com-mon understanding and usage” dic-tated that the dog was, in fact, a pitbull.50 The combination of thesefacts led the court to conclude thatthe statute was too vague to passconstitutional muster.51

The Reaction of InsurersWhile some communities andstates have responded to dog biteswith breed-specific legislation de-signed to regulate or outlaw cer-tain breeds, insurance companieshave also reacted to the problem ofdog bites in a breed-specific man-ner. Dubbed “breed discrimina-tion” by dog owners, insurancecompanies have started makingcoverage and renewal decisionsbased on one’s ownership of cer-tain breeds of dog.

A Rise in BreedDiscriminationDuring 2003 and 2004, the mediabrought breed discrimination tolight. The CBS Evening News withDan Rather aired a story in June2003 that featured a family thathad difficulty obtaining insurancebecause they owned a dalmatian.52

The report stated, “[A]nimal

lovers have a term for what the in-surance company did. They call it‘breed discrimination’—arbitrarilypunishing all dogs of certainbreeds because some are vicious.”53

In the months that followed, sev-eral newspaper stories discussedthe prevalence of breed discrimina-tion and documented the effectsthis practice has had on families.54

These news reports replicate theexperience I had in trying to gethomeowners’ insurance. Multipleinsurers denied coverage becauseof the dogs I owned. I literallycould not find a carrier in the Lub-bock market willing to write a pol-icy for me until I stumbled uponthe Farm Bureau on the advice ofone insurance broker who sympa-thized with my plight.

The practice of breed discrimina-tion produces absurd results. Con-sider the case of Chris and NormCraanen of San Antonio, Texas.55

They own a twelve-year-old dognamed Bukarus. He is a rottweiler,a breed often targeted for discrim-ination by insurance companies.Yet, Bukarus does not pose much ofa threat: he is deaf, partially blind,and has arthritis.56 Despite his bite-free history, his owners lost theirhomeowners’ insurance.57

Some of the most well-known in-surers are engaging in breed dis-crimination.58 Some insurers haveoutright bans on specific breeds,59

while others take a more realisticand logical dog-by-dog approach.These decisions are predicated oninsurers’ assessment of relativerisk.60 The “usual suspects” forbreed discrimination are pit bulls,rottweilers, German shepherds,Doberman pinschers, Chow Chows,wolf hybrids, and Presa Canarios.61

The Humane Society of theUnited States (HSUS) has docu-mented an increase in the numberof people being denied insurancebecause they own certain breeds ofdog.62 As a result, The HSUS hasstarted collecting data through theInternet, in the hopes of eventuallyconvincing the insurance industry

that there are alternatives to thecurrent practice and that it muststop.63 To achieve their goal, TheHSUS and the American Societyfor the Prevention of Cruelty to An-imals (ASPCA) have created a jointgrass-roots campaign designed toeducate the insurance industry.64

The InsuranceIndustry’s Defense ofBreed DiscriminationHomeowners’ insurance protects apolicyholder in the event of finan-cial loss. Most policies include twoprovisions, property damage andliability. Property damage provi-sions protect the policyholder inthe event of fire, lightning, wind,water, or hail damage, theft, andvandalism. Liability provisions pro-tect the policyholder in the eventthat a claim is made against ahomeowner for negligence. Liabil-ity coverage typically pays for bod-ily injury, medical payments, andproperty damage that are sus-tained because of the negligence ofthe property owner.65 Absent breeddiscrimination, most homeowners’insurance policies would cover in-juries due to dog bites on the pre-mises between the amounts of$100,000 and $300,000.66 In 1995the average policyholder paid $418in homeowners’ insurance premi-ums.67 By 2004 the average pre-mium climbed to $608.68

“Insurance is a business.”69 Insur-ers must make profits in order tocontinue in existence.70 Companiessurvive by minimizing risk, whichreduces the likelihood of claims.Some companies have decided thatcertain breeds of dog are simply“too much of a risk” to insure.71 Anindustry representative claims thatthe issue of dog bites “is a majorconcern for insurers.”72

The industry defends its posi-tion, in part, on a series of studiesfrom the Centers for Disease Con-trol and Prevention (CDC), whichthe industry claims as support forthe proposition that certain breedshave a propensity to bite.73 As I

29

demonstrate in Part II, however,the industry’s reliance on the CDCstudies is misplaced. Even theauthors of the CDC studies havestated that breed discriminationis wrong and is not supported byscientific evidence.74

The industry has also pointed tothe large amount of money thathas been paid out in recent yearsfor dog-bite claims.75 The Insur-ance Information Institute (III), atrade group of the insurance indus-try, stated that in 2002 $345.5 mil-lion was paid out in dog-bite liabil-ity claims, up from $250 million in1995.76 The group argues that dog-bite lawsuits are on the rise andjuries are awarding larger claims.77

It claims, therefore, the need tocurtail its risk.

The industry’s cost statistics aremisleading, however. The III states,“[D]og bites now account foralmost one quarter of all home-owner’s insurance liability claimscosting $345.5 million.”78 Someperspective is in order. Forevery $100 in premiums, insurersspend $77 paying claims. Of that$77, the overwhelming majority($72, or 93.5 percent) is spenton paying property damageclaims. Liability claims onlyamount to $5, or 6.5 percent, oftotal claims.79 Even then, dogbites only constitute a percentageof that figure. Put into perspec-tive, the money paid out in dog-bite claims is negligible whencompared to the overall amount ofmoney paid out for other types ofclaims. Damage due to lightning,fire, and mold all individuallyaccount for more claims payoutsthan all liability claims combined.80

The insurance industry has notbeen consistent in the reasons forits defense of breed discrimination.One report from the III’s websiteseems to defend breed-specificresponses based on the aggregateclaims paid81 and stories of severalhigh-profile and tragic bites.82

However, in a statement to anewsletter of veterinary medicine,

the III defended breed discrimina-tion on the basis that certainbreeds cause more damage whenand if they do bite.83 Ultimately, aspokesperson for the III conceded,“[t]he industry isn’t positioned todetermine which dogs should bedeemed vicious....[W]e’re certainlynot dog experts or veterinari-ans.”84 This, however, has notstopped many insurers from engag-ing in breed discrimination.

Some Exceptions to the Rule?It appears that not all insurershave followed the breed discrimi-nation trend. DVM reported thatNationwide Insurance changed itsbreed discrimination policy inOctober 2003. While Nationwidenow insures all dog owners, itspecifically excludes dog bites fromits liability coverage.85

State Farm’s national represen-tatives have repeatedly stated thatthe company does not practicebreed discrimination.86 However,when I searched for homeowners’insurance in 2003, a State Farmagent in Lubbock refused to eventake my application because of thebreeds I owned.

Other Instances ofBreed DiscriminationThere are other examples where aperson’s ownership of a particularbreed of dog can have negativeconsequences. Families seeking toadopt children can face roadblocksif they own dogs that belong to cer-tain breeds. In Massachusetts theAdoption and Foster Care Unit ofthe Department of Social Serviceswill not place children in homeswith certain breeds of dog.87 Thestate relied upon data provided bythe insurance industry when itmade its decision to discriminatebased on breed.88 Some airlinesalso practice breed discriminationby prohibiting some dogs from fly-ing, even though they are stored incargo and in a closed carrier.89

II. The Lack ofScientific EvidenceNumerous scientific studies haveattempted to identify the numberof annual dog bites, the dogs mostlikely to bite, the people mostlikely to be bitten, and the circum-stances under which bites are mostlikely to occur. Such studies havenot reached a uniform consensusand have left us with more ques-tions than answers. Even the stud-ies that have attempted to reporton breeds’ proclivity to bite havecautioned that their research isincomplete and should not be usedto justify breed discrimination bylegislatures or insurers.90

CDC StatisticsThe CDC commissioned a numberof studies during the 1980s and1990s to determine the scope andnature of the problem of dog bitesin the United States.

Fatality StudiesFour separate studies attempted tochronicle the number of fatal dogbites dur ing the per iods o f1979–1988,91 1989–1994,92 1995–1996,93 and 1997–1998.94 Thestudies were specifically limited tofatal dog attacks because fatalitystatistics are easier to track.95 Non-fatal bites were excluded from thestudies, although other scientistshave attempted to use emergencydepartment reports and othersources to determine the numberof nonfatal bites per year.96

The authors combed three setsof sources in an attempt to deter-mine the number of fatal dog bitesper year. First, they searchedNEXIS for news reports of dog bite-related fatalities.97 Second, theyused the National Center forHealth Statistics’ (NCHS) single-cause mortality tapes (SCMTs) toidentify deaths where the underly-ing cause was listed as a dog bite.98

Finally, the authors supplemented



these reports with information col-lected by The HSUS to help iden-tify the breed of dog involved ineach incident.99 From these threesources, the authors tried to piecetogether the number of peoplewho died each year in the UnitedStates from dog bites.

The authors concluded that dogbites caused approximately sevendeaths per year per hundred mil-lion people.100 They discerned noidentifiable trend that would indi-cate an increase in the incidenceof fatal bites over the years of thestudies.101 During the first report-ing period (1979–1988), approxi-mately 70 percent of victims wereunder the age of ten.102 Males,under the age of twenty-nine, weremore likely than females to be vic-tims.103 These findings as to ageand gender were consistentthroughout the study periods.

Many of the fatal bites of childreninvolved horrific attacks on the veryyoung. A three-week-old girl waskilled in her crib by the family’sChow Chow.104 A two-year-old boy inSouth Dakota wandered into aneighbor’s yard, where he wasattacked and killed by two Germanshepherd-wolf hybrids.105 The elderlywere also victims of several fatalattacks. In March 1996 two rottweil-ers killed an eighty-six-year-old Ten-nessee woman. One month beforethe assault, the dogs had attackedand injured the same woman.106

In the twenty-year period of theCDC studies, the breed responsiblefor the most number of bites haschanged.107 From 1979 to 1980,Great Danes caused the most num-ber of fatalities, with three deathsfor the period. However, fourbreeds were tied with two deathseach: pit bulls, rottweilers, huskies,and malamutes.108 In 1981 pit bullstook over as the breed with themost number of fatal bites.109 Pitbulls remained in that positionuntil 1993, when rottweilers begancausing approximately ten fatalbites per two-year reportingperiod.110 The last available report-

ing period, 1997—1998, showsthat rottweilers caused ten fatalbites per two-year period, while pitbulls caused six, and Saint Ber-nards caused three.111 During thetwenty-year study, ninety deathswere excluded because the breedwas “unavailable.”112

The authors of the CDC studiesacknowledged that the methodsthey used in their studies had anumber of limitations. NEXIS, theypointed out, was not designed forscientific research. News reportswould only be flagged if their textcontained certain keywords.113

Further, rel iance on NEXISassumes that newspapers accu-rately reported the breed of doginvolved in a particular attack.114

SCMTs have a one- to two-year lagtime, which means that some fatal-ities may have been missed.115 Theauthors believed that, on average,their methods only uncoveredapproximately 74 percent of dog-bite-related fatalities.116

Even if one accepts the CDC sta-tistics as definitive on the subject,they have a number of other limita-tions in answering the questionof whether certain breeds are moredangerous than others. First, thestudies were limited to fatal dogattacks.117 Second, the breed ofthe dog could not be accurately de-termined in every case.118 Finally,the number of fatal attacks peryear is so low that it is problematicto statistically extrapolate conclu-sions from the data. For example,in the first two years of the study(1979–1980), Great Danes ac-counted for the most number offatal bites (three).119 Four breeds,however, followed closely behindwith two fatal bites each (pit bull,German shepherd, husky, and mal-amute).120 It would be statisticallyquestionable to conclude thatGreat Danes were inherently moredangerous than the other breeds,based on a net difference of onlyone fatality.

NonfatalityStudiesThe CDC fatality studies acknowl-edged that, while death rates fordog bites do not appear to haveincreased over time,121 nonfatalbites were becoming more of apublic health problem.122 The CDCconducted a study of nonfatal dogbites in 2001.123 The study useddata from the National ElectronicInjury Surveillance System-All In-jury Program (NEISS-AIP) to iden-tify the number of nonfatal dogbites during the 2001 calendaryear. NEISS-AIP collects data frominitial visits to emergency depart-ments (EDs) across the country.124

NEISS-AIP data are drawn from anationally representative sample ofNEISS hospitals.125 The CDC ana-lyzed every case where “dog bite”was listed as the external cause ofinjury.126

In total, NEISS-AIP data revealedthat hospital EDs treated 6,106patients for dog-bite-related in-juries during 2001.127 Since theNEISS-AIP data did not includeevery hospital in the nation, theauthors used these data to extrap-olate to the general population.128

They estimated that 368,245 peo-ple were treated for dog-bite-related injuries in 2001.129 Thelargest cohort of victims was chil-dren between the ages of five andnine.130 Boys, under the age offourteen, were more likely thangirls to be seen in EDs for dog-bite-related injuries.131

The NEISS-AIP data includednarratives for many of the attacks.One case involved a four-year-oldwho was bitten by a dog guardingher puppies.132 Another involved athree-year-old girl who was bittenwhen she tried to take away adog’s food.133 A thirty-four-year-oldman was bitten while trying tobreak up a dogfight. Some victimswere bitten by their own dogs. Atwenty-seven-year-old woman wasbitten by her dog after he hadbeen hit by a car and became dis-

31

oriented.134 A seventy-five-year-oldwoman was attacked while tryingto prevent her dog from biting anemergency medical technician(EMT) who was attempting to putthe woman in an ambulance.135

The Morbidity and Mortalityreport describing the study doesnot document the number ofattacks per breed. This is likely dueto the fact that the ED reports didnot specify the breed of dog. Anattempt to determine the numberof bites per breed would depend onvictims accurately self-reportingthe breed of the attacking dog.136

The study had a number of limi-tations. First, the authors excludedfatal dog bites. Second, the studyonly examined cases where the vic-tim sought treatment in an ED.Victims may have gone to otherhealth care providers, such as pri-vate physicians or urgent-care cen-ters. Third, 26 percent of reportswere missing an injury diagnosis.Many cases had limited data on thecircumstances of the attack or theidentity of the dog involved.137

Thus, the CDC’s estimates may beboth overinclusive (“just cause”bites may have been included)138

and underinclusive (insofar as vic-tims may have sought treatment atother facilities).

Another CDC study attempted toidentify the incidence of dog bitesin a particular locality: Denver, Colorado.139 The authors examinedreports from the Denver MunicipalAnimal Shelter in 1991.140 Therewere a total of 991 bites during thestudy period.141 However, only 178were eligible for the study,142 asthe authors excluded several cate-gories of bites: bites involvinghousehold members, attacks invol-ving multiple dogs, attacks before1991, dogs who had been ownedfor less than six months, cases inwhich the owner did not live inDenver County, attacks where theowner’s phone number was notlisted on the report, and cases inwhich the victim did not receivemedical treatment.143

The study created a controlgroup of dogs to try to determinewhether certain characteristics(such as breed) made a dog morelikely to bite.144 Using a multivari-ate statistical analysis, the studyconcluded that biting dogs weremore likely than control dogs tobe German shepherds or ChowChows, male, intact (not neu-tered), and reside in a house withone or more children.145 Denverhad (and still has) a ban on pitbulls, so it is not surprising that nocases involved that breed.146

The authors acknowledged thattheir results had several problems.First, they were only able to speakto owners of approximately half ofthe biting dogs. They excludedcases in which the victim did notseek medical attention. In thisrespect, the authors believed thatseeking medical attention was a“surrogate” for “real bites.”147 Theauthors did not verify the breeds ofthe dogs involved, but, instead,“identified predominant breed aswhatever breed the owner consid-ered the dog.”148 Because of thesmall number of bites per breed,the authors could not assess thestatistical significance of breedsother than German shepherd andChow Chow.149

Another CDC study attempted todetermine the frequency of dogbites by conducting a random tele-phone survey of households.150 Theauthors used the Injury Control andRisk Survey (ICARIS), a random-digit-dialing telephone survey.151

They asked each adult respondentwhether he (or his children) hadbeen bitten by a dog in the previoustwelve months and whether the vic-tim had sought medical atten-tion.152 Out of 5,328 completed in-terviews, ninety-four adults andninety-two children reported beingbitten in the previous twelvemonths.153 Of these, twelve adultsand twenty-six children sought med-ical care.154 From these data, theauthors extrapolated that 1.8 per-cent of the American population

(4,494,083 people) had been bittenin the previous twelve months, and0.3 percent had sought medicalattention.155 This shows that nonfa-tal bites are a public health problemthat “is five orders of magnitudegreater” than fatal dog bites.156 Thestudy concluded that several factorshad no statistical significance onthe likelihood of being bitten: cen-sus region, urbanicity, race/ethnicgroup, and household income.157

The study did not attempt to corre-late between the number of bitesand the breed of dog. The authorsacknowledged that the study reliedon the self-reporting of data, whichwere not validated, and that theyreceived a poor response rate (only56 percent of people responded tothe survey).158

Other StudiesOther studies have attempted to doc-ument the total number of dog bitesand the number of bites per breed.

A study of ED visits for dog-biteinjuries159 confirmed many of theconclusions of the previously dis-cussed CDC study of ED visits.160

The study noted that a lack of a na-tional reporting system for dog-bite injuries makes gathering andanalyzing data on the subject diffi-cult.161 The authors, in reviewingthe literature on the subject, foundthat previous studies concludedthat between 0.3 percent and 1.1percent of all ED visits are due todog-bite-related injuries.162 Todetermine the true percentage,they collected data from the Na-tional Hospital Ambulatory Med-ical Care Survey (NHAMCS), a ran-dom surveying instrument that isused to calculate the number ofED visits per year.163 They esti-mated that between 1992 and1994, 333,687 annual visits weremade to EDs seeking medicaltreatment for dog-bite-relatedinjuries.164 This amounted to 0.4percent of all ED visits nation-wide.165 Looking at the monetarycost of dog bites, they found that



the average cost for a dog-bite-related ED visit was $274, result-ing in an annual cost of $102.4million.166 The study, however, didnot address the quest ion ofwhether certain breeds are particu-larly more dangerous than others.This is partly due to the unavail-ability of data through NHAMCS.Moreover, the study most likelyundercounted the number of non-fatal dog bites because victims mayhave sought treatment from placesother than EDs.167

Other studies have attempted toexamine the problem at a morelocalized level. A July 1991study168 found that dog bites wereresponsible for 0.3 percent of allED visits at The Children’s Hospi-tal of Philadelphia.169 Of those vis-its, 77 percent involved caseswhere the victim knew the bitingdog.170 The study found one statis-tically significant conclusion: morepit bull injuries were the result ofunprovoked attacks as comparedto such attacks by other breeds.171

“Unfortunately, the absence of reli-able dog breed-specific populationfigures prevent[ed] the calculationof breed-specific injury rates.”172

An October 1997 study tried todetermine the number of dog bitesin Alleghany County, Pennsylvania(Pittsburgh), by using the “cap-ture-recapture” method of statisti-cal analysis.173 The authors foundthat 790 dog bites were reportedto the Alleghany County HealthDepartment in 1993.174 Using thecapture-recapture method, alongwith log-linear modeling, the studyconcluded that the number of un-reported dog bites was 1,388 (witha 95 percent confidence interval ofbetween 1,010 and 1,925).175 Theauthors cautioned, however, thatthe self-reporting sources are prob-lematic in that “whether or not acase is reported depends largely onthe severity of the event and theattitude, knowledge, or educationlevel of the victim.”176 Accordingly,the authors suggested that theactual Pittsburgh dog-bite inci-

dence rate must be higher thanthat found in the study.177

Another survey178 in Pennsylva-nia polled children in order todetermine an overall bite rate fromthe perspective of bite victims.179

The survey, conducted in 1981,found that 46.1 percent of childrenreported that they had been bittenby a dog during their lifetime.180

The study concluded that [B]eing bitten by a dog is arather common occurrence forchildren, especially those be-tween the ages of seven andtwelve years, and the event isgreatly underestimated by offi-cial bite statistics.181

Nevertheless, the authors did notattempt to catalog bites per breed.182

Unfortunately, not all scientistshave used statistically sound meth-ods to draw conclusions about therelative dangerousness of breeds.Two physicians, Lee E. Pinckneyand Leslie A. Kennedy, from theDepartment of Radiology at theUniversity of Texas SouthwesternMedical School and Children’sMedical Center, sent letters to theeditors of 245 major newspapersrequesting copies of all storiesabout dog-bite-related fatalities.183

The number of fatalities reportedby the responding newspapers be-tween March 1966 and June 1980totaled seventy-four.184 Of the sev-enty-four fatalities, sixteen werecaused by German shepherds, nineby husk i e s , e i gh t b y Sa in tBernards, six each by bull terriersand Great Danes, and five by mala-mutes.185 The remaining dog-bitefatalities were caused by a varietyof breeds, including ten attacks bymixed breeds and five attacks bydogs of unknown breeds.186 Inaddition to acquiring bite fatalitystatistics from newspapers, theauthors used American KennelClub (AKC) registration data tocompare the relative number offatalities per breed.187

The CDC authors criticized thePinckney/Kennedy study as being“primarily anecdotal” rather than

“systematic” in its approach.188

Indeed, Pinckney and Kennedy con-ceded that their database was“incomplete” and “may not beentirely reliable.”189 Their datadepended on newspaper reports,which may themselves be incom-plete or inaccurate. Thus, theauthors said their data required“cautious interpretation.”190 Anexample of such “cautious interpre-tation” is represented by theauthors’ observation that eventhough German shepherds wereinvolved in more fatalities than anyother breed in the study, such largefrequency could be reflective of thefact that German shepherds hadthe highest AKC registration of anylarge breed.191 Hence, the use ofAKC data to draw comparisonsbetween breeds is problematic,192

as demonstrated by the high num-ber of registrations for breeds suchas German shepherds, and lownumber of registrations for a popu-lar breed, such as the pit bull.193

William Winkler’s study194 in1977 has also been criticized forits lack of scientific method.195 His“study” involved compiling newsreports from eleven dog-bite-re-lated fatalities from January 1974through December 1975.196 Fromthese data, he made various con-clusions about the breeds responsi-ble, finding that, “not unexpect-edly,” German shepherds were thebreed most often responsible forfatal dog attacks.197 Because SaintBernards were responsible for twodeaths during this twenty-four-month period, he concluded,“[t]his relatively uncommon breedmay be a greater hazard thanothers.”198

A common thread runningthrough several studies is theattempt to extrapolate conclusionsabout breeds based on limiteddata. For example, an April 2000epidemiological study in Philadel-phia used reports from the Depart-ment of Health to conclude thatbetween 1995 and 1997 there wereapproximately 5,390 bites.199 The

33

authors concluded that pit bulls,German shepherds, and rottweilerscombined were responsible for 59percent of bites each year.200 Theauthors felt comfortable drawingthis conclusion despite the factthat they could not determine thebreed in 74 percent of cases.201

The UnknownOrigin ofAggressivenessDespite all of the research and stud-ies on the subject, scientists andveterinarians cannot state with cer-tainty or confidence why certaindogs are more aggressive than oth-ers.202 It seems that a particulardog may be aggressive because of avariety of factors.203 According tothe American Veterinary MedicalAssociation’s multidisciplinary TaskForce on Canine Aggression andHuman-Canine Interactions, “Adog’s tendency to bite depends onat least five interacting factors:heredity, early experience, later so-cialization and training, health(medical and behavioral), and vic-tim behavior.”204

While breed (as an inheritedcharacteristic) is one componentof predicting a dog’s dangerous-ness, it is not the only factor.205

There is no way to scientificallydetermine whether a dog is likelyto bite in the future, any morethan psychologists can predictwhether certain people will com-mit crimes of violence. The excep-tion to this rule is the axiom thatthe best predictor of future behav-ior is past behavior. For this rea-son, many veterinary and scientificgroups support “dangerous doglaws” that target individual dogswho have demonstrated a propen-sity to bite or attack innocent victims.206 The problem with BSLand breed discrimination is thatlegislatures and insurers haveattempted to prophylactically deter-mine which breeds are most likely

to bite without any evidence of in-dividual dangerousness.

Numerators andDenominators in Dog-BiteStatisticsTo date, no scientific study hasbeen able to resolve what I term tobe the problem of “numeratorsand denominators.” A person wish-ing to determine whether certainbreeds are more likely to bite thanothers must first determine thenumber of bites per breed (thenumerator) and then compare thatnumber to the total number ofdogs of that breed in the generalpopulation (the denominator).This can be expressed as a ratio:

This ratio (RDR) allows for acomparison between breeds. Thehigher the RDR, the greater pro-clivity a particular breed has tobite. It allows for a comparison of“oranges to oranges” and “applesto apples.” Otherwise, it is likelythat highly popular breeds willappear to be more dangerous,when in fact the number of bites isreflective of the overall populationof the particular breed.

A study that tried to extrapolatebreed data from the previously dis-cussed CDC studies agreed thatthe proper method for determin-ing a breed’s dangerousness wasthe use of a comparative ratio:

Ideally, breed-specific bite rateswould be calculated to comparebreed and quantify the relativedangerousness of each breed.For example, 10 fatal attacks byBreed X relative to a populationof 10,000 X’s (1/1,000) imply agreater risk than 100 attacks byBreed Y relative to a populationof 1,000,000 Y’s (0.1/1,000).Without consideration of the

population sizes, Breed Y wouldbe perceived to be the moredangerous breed on the basis ofthe number of fatalities.207

Using the RDR normalizes theeffect of a breed’s popularity, orlack thereof. Dogs of popularbreeds are going to bite moreoften simply because there aremore of them.208 A January 1997article warned that, as dalmatiansbecome more popular, peopleshould expect to see more bitesfrom that breed.209 This is not tosay that dalmatians are inherentlymore dangerous than other breeds.Rather, an increase in their popula-tion should also result in a propor-tional increase in bites from thatbreed.210 Similarly, the Pinckney/Kennedy study211 cautioned that,despite the fact that German shep-herds accounted for the most num-ber of deaths, their finding mustbe read in conjunction with thepopularity of the breed, as evi-denced by AKC registrations of thesame time period.212

The problem of numerators anddenominators is that it is diffi-cult—if not impossible—to accu-rately determine the number ofbites per breed and the number ofdogs in a particular breed. Withoutan accurate count for either thenumerator or denominator, oneruns the risk of stigmatizing anentire breed as “overly dangerous”based on the breed’s absolute num-ber of bites, instead of examiningthe breed’s number of bites rela-tive to its overall population.

The Numerator ProblemThe principal problem in determin-ing the total number of bites by a particular breed is that there is nonational reporting system for dogbites.213 The CDC studies214 demon-strate that, while fatal dog bites areeasier to track than nonfatal bites,even the methodology used touncover fatalities misses approxi-mately 26 percent of cases.215

Further, news accounts—on whichthe CDC relied, in part, to deter-


Number of Bites by Breed

Total Population of Breed

Relative Dangerousness =

Ratio


mine the number of fatal dog bitesand the breeds involved—may bebiased toward reporting attacks bycertain breeds.216

The numerator may also bebiased against dogs who causemore damage, while ignoringbreeds that bite more often but donot cause victims to seek emer-gency treatment.217 If a dog bitedoes not cause serious injury, it isnot likely that the victim wouldseek medical treatment.218 Thisthen skews the results of studiesthat use emergency departmentvisits to track the incidence of dogbites.219 “The problem with self-reporting sources is that whetheror not a case is reported dependslargely on the severity of the eventand the attitude, knowledge, or ed-ucation level of the victim.”220

Studies that have used randomsampling221 are equally problem-atic because they, too, depend onaccurate self-reporting of theirsample groups. The low responserates of these studies also lead toquestions about the accuracy ofthe results that are extrapolated tothe general population.222

The DenominatorProblemNo one knows how many dogs arepresent in the United States at anyone time. This should not be sur-prising, as even the constitution-ally mandated223 decennial censusof human beings is known to un-dercount people.224

Determining the true number, oreven an accurate estimate, of dogscan be problematic. While manydogs are kept as household pets,225

others are used as service animalsor guard dogs; kept in animal shel-ters or animal stores; or simplyallowed to wander the streets asstrays. The dog population is con-stantly changing and moving,which makes obtaining an accuratecount difficult and expensive.

Even if it was possible to deter-mine how many dogs exist in thecountry at any one time, the prob-

lem then becomes how to deter-mine how many of those dogs be-long to each breed. Determiningthe breed of one dog is difficultenough.226 To take a census of alldogs and identify their breedswould be an impossible task.

Some scientists have suggestedusing AKC or municipal registra-tion data to determine the numberof dogs in a particular breed in aparticular community.227 However,one study concluded that city reg-istrations account for only 29.1percent of all dogs.228 Further,owners of breeds considered “dan-gerous” may be reluctant to regis-ter their animals.229 This may beparticularly true of dogs used forillicit purposes, such as thoseowned by drug dealers, dogfight-ers, and gang members.230

AKC registration data is alsoproblematic because the AKC onlyregisters purebred dogs231 anddepends on owners taking the ini-tiative to register their dogs.232

Mixed breeds, for which there arenumerous combinations, are noteligible for registration.233 Pit bullsare often registered with organiza-tions other than the AKC. If ownersdo register them, they registerwith the United Kennel Club or theAmerican Dog Breeders Associa-tion.234 If a breed is undercountedin the denominator of the ratio, itwill make a breed appear moredangerous than it actually is.235

The Problem of BreedsBreed is a human construct that isused to conveniently group dogsbased on similar physical character-istics.236 There is no scientific testto determine a dog’s breed.237 Theonly way to determine a dog’s breedis to examine its heredity. This taskis made possible but is expensiveand time-consuming,238 if a dog isregistered with the AKC.239

As examples of the problem ofdefining and identifying breed,consider the case of huskies and

pit bulls. “Husky” refers to a classof dogs, not any one particularbreed. Siberian huskies, Alaskanmalamutes, and Samoyeds are allconsidered to belong to the“husky” family, yet they are all dif-ferent breeds.240 Similarly, there isno AKC-standard breed called “pitbull.” “Pit bull” is a collective clas-sification of the American Stafford-shire terrier, Staffordshire pit bullterrier, and bull terrier.241

Scientists have not been able todetermine if victims of dog bitescan accurately report the breeds ofdogs that attacked them. Many sci-entists, particularly the CDCauthors, have stated that misiden-tification is a likely problem, espe-cially under the stress of a dogattack.242 Part of the problem maybe that as a particular breed getsa reputation for dangerousness,some victims jump to the conclu-sion that they were bitten by a dogof that breed.243

Even under ordinary, low-stressconditions, many people have diffi-culty identifying a dog’s breed.

For the average person any-thing with prick ears and blueeyes automatically becomes a“husky”....Any smooth coatedbrown dog, medium sized, andmuscular becomes a “pitbull”....Any tall dog becomes aGreat Dane, fuzzy or hairy, andit’s a Chow Chow. If it’s blackand tan and heavy, it’s a rot-tweiler, etc.244

One survey of bite reports foundthat medium-size black and tan ani-mals were likely to be recorded asGerman shepherds. Stocky, short-haired dogs were listed as pit bulls.Media reports of pit bull attacks areoften accompanied by pictures ofboxers or pugs instead of AmericanStaffordshire terriers.245 One enter-taining website, called “Find the pit bull,” displays twenty-one pic-tures of purebred dogs and chal-lenges the user to identify the pitbull among them.246

Even veterinarians and otherexperts have difficulty determining

35

whether a particular dog belongs toa particular breed.247 This was a cen-tral concern of the MassachusettsSupreme Judicial Court in Ameri-can Dog Owners Association v. Cityof Lynn.248 The Court declared thecity of Lynn’s pit bull ordinance tobe unconstitutional in part becausethe animal-control officers desig-nated to enforce the ordinance usedconflicting and subjective standardsto determine and identify breed.249

The problem of mixed breedcomplicates the issue even further.In determining a relative danger-ousness ratio, it is unclear how tocount mixed breeds.250 Should theybe counted once per breed? Not atall? Create a new category for eachpossible combination of breeds?Aside from how to use the raw dataon attacks by mixed breed, there isthe additional problem of misiden-tification by laypeople.251 Victimssometimes inadvertently reportmixed-breed dogs as purebreds252

due to the heat of the moment andtheir lack of training in identifyingsubtle breed characteristics.

There is good reason to believethat the raw data being used to cal-culate relative dangerousnessratios are incomplete and inaccu-rate. If the data being input intothe calculation are flawed, the re-sults (claiming to show somebreeds are more dangerous thanothers) are equally flawed.253

The Problem of“Just Cause” BitesEven if an accurate count could beobtained of the number of bites perbreed, there is the additional prob-lem of how to handle “just cause”bites in the resulting statistics. Ifthe purpose is to determine whichbreeds are inherently more danger-ous, just by virtue of the breedsthemselves, then the statisticsshould exclude bites by the dog thatwere justified. If a rottweiler bitesan intruder who is attacking thehomeowner, we would expect the

rottweiler to be praised for defend-ing its owner. This is not the type ofbite that we should be trying to pre-vent. It is also not the type of bitethat is likely to lead to an insuranceclaim. Similarly, if a dog is beingphysically tormented by a neighbor-hood child who is poking it in theeye, we would not deny that the doghas an inherent right to defenditself by growling, snarling, barking,or biting back.254 These are “justcause” bites, bites in which the doghas a legitimate reason to defenditself or its owners.

It is possible that the statisticsare being skewed because propertyowners who wish to purchase“guard dogs” may be self-selectingcertain breeds based on the popularnotions of relative dangerousness.Guard dogs are trained to protectproperty by scaring away would-beintruders and, if necessary, to bitean actual trespasser. Owners whodesire to have guard dogs mayrationalize the purchase of onebreed over another based on thedegree to which they subjectivelybelieve that the dog will be “mean”or “scary.” This creates a self-fulfill-ing prophecy. The “scarier” a breedis considered by a community, themore likely a dog of that breed willbe purchased for protection, usedfor protection, and actually bite anintruder. This will skew the statis-tics in a way that purports to showthat the particular breed is, in fact,inherently more dangerous.

Despite these concerns, itappears that the studies to datehave not excluded this category ofbites from their datasets.255 This isa fatal flaw in the statistics, for itconfuses the issue between inher-ent dangerousness (due to breed)and legitimate animal behavior.

Breed Switchingby Bad OwnersAssume for the moment that anaccurate relative dangerousnessratio could be determined for each

breed, and that it could be scientif-ically determined that certainbreeds are inherently more danger-ous than others. What about theowners? Does this not excuse themfrom the responsibility to properlytrain and care for their pets?

The reality is that there is a widespectrum of responsible pet owner-ship. For some people, occasionallyproviding food and water for a dogis considered sufficient. On theopposite end of the spectrum,some people spend thousands ofdollars on luxuries such as petspas, advanced dog agility classes,and elaborate beds. Somewhere inthe middle of the spectrum arepeople who actively ensure thattheir pets have food, water, andshelter; get exercise; are welltrained; and receive adequate vet-erinary care.256

Unfortunately, a small percent-age of pet owners breed and usetheir pets for illicit purposes. Theyintentionally seek out vicious dogswho will attack and maim humansand other animals.257 Dogfightingenthusiasts, gang members, anddrug dealers will purposely select,breed, and train dogs to be vicious.The purpose may be to intimidaterivals (in the case of gangs anddrug dealers), to defend illegaldrugs (in the case of drug dealers),or to make money (in the case ofpromoters of dogfights).258 Forsome, having a vicious dog is sim-ply a status symbol.259 In order tomake dogs into vicious weapons,they use “revolting and painfultechniques to bring the animals tothe verge of bloodlust.”260 Drugdealers in Philadelphia during the1980s had pit bulls named “Mur-der, Hitler, and Scarface.”261 Theywore collars that concealed crack,cocaine, and money.262 In Chicago,gang members “brandish[ed] theirfierce pit bulls just as they would aswitchblade or a gun.”263

Current statistics do not take in-to consideration the degree towhich the source of a dog’s aggres-siveness is the torturous upbringing



described above, as opposed to thedog’s breed.264 In those situations,the problem is clearly with the dogowner—not the dog itself or itsbreed. These problem owners aredangerous with any breed of dog.265

One solution would be for insur-ers to write policies that excludeinjuries related to dogfighting. Thiswould limit the claims paid out forthese high-risk animals, yet it wouldleave potential plaintiffs without anadequate source of compensation.This result might be a sociallyacceptable solution because of theunclean hands of the “victims.” Ifdogfighting exclusions are incorpo-rated into standard homeowners’insurance contracts, the languageshould be narrowly written toexclude only those bad faith actorswho, as a matter of social policy,should not be rewarded or compen-sated for injuries attendant to anillegal activity. The key would be towrite language that would still pro-tect innocent passersby.

One of the arguments againstBSL is that once a breed becomesbanned, problem owners will sim-ply switch to another breed.266 Inthe 1930s, pit bulls were far fromconsidered a “vicious breed.” Infact, a pit bull named “Pete”starred in the Our Gang films ofthe time.267 Fifty years ago theDoberman was considered themost vicious dog.268 During the1980s the focus turned to pitbulls.269 In short, today’s publictarget may be tomorrow’s favoritepet, and vice versa.

Do the InsuranceCompanies HaveBetter Data?It is quite possible that one ormore insurance companies havetheir own proprietary data purport-ing to show that one breed oranother is disproportionately re-sponsible for bites. I am skepticalthat their data would be any betterthan the CDC’s. The problems

associated with the CDC and non-CDC studies are inherent to theproblem of trying to determine thenumber of bites per breed and thenumber of dogs per breed.

III. The WidespreadOpposition to BreedDiscriminationBreed discrimination by insurancecompanies and breed-specific leg-islation by state and local gov-ernments have attracted natio-nal attention and outrage by vet-erinarians, animal groups, anddog owners.

The American Veterinary Med-ical Association’s Task Force onCanine Aggression concluded thatBSL and other breed-specificactions are “inappropriate andineffective.”270 The Task Force con-sisted of a diverse coalition of vet-erinarians, academics, physicians,insurers, representatives from ani-mal rights advocates, CDC scien-tists, and lawyers.271 The TaskForce agreed that to properly de-termine the relative dangerousnessof breeds, one must first determinethe number of bites per breed andthe total population of each breed.As noted above,272 the accuratecalculation of both numbers is animmense challenge.273

The Task Force rejected the no-tion that a dog’s breed is the soledeterminant of dangerousness.“[A] dog’s tendency to bite de-pends on at least five interactingfactors: heredity, early experience,later socialization and training,health (medical and behavioral),and victim behavior.”274 They also pointed to the problems ofmixed breeds, misidentification ofbreeds, and shifting popularity ofbreeds.275 The Task Force also ex-pressed concern about making de-cisions based solely on breed, sincethere is a lack of scientific meansto identify breed.276 The Task Forcerecommended, instead, that local

governments focus on individualdogs and dog owners.277

The very scientists who haveauthored studies trying to deter-mine a link between breed andaggressiveness oppose breed dis-crimination and BSL. In many of the CDC studies, the scientistscautioned against using their in-complete data on attacks to makeknee-jerk legislative or policy deci-sions based solely on breed.278

They pointed to the lack of reliabledata on bites per breed (the“numerator problem”) and theabsence of a reliable count of dogsper breed (the “denominator prob-lem”).279

Animal groups have also opposedBSL and breed discrimination. TheAKC has taken a strong stanceagainst breed discrimination by in-surance companies:

The American Kennel Clubbelieves that insurance compa-nies should determine cover-age of a dog-owning householdbased on the dog’s deeds, notthe dog’s breeds. If a dog is awell-behaved member of thehousehold and the community,there is no reason to deny orcancel coverage. In fact, insur-ance companies should con-sider a dog an asset, a naturalalarm system whose bark maydeter intruders and preventpotential theft.280

The AKC also issued this state-ment concerning BSL:

The American Kennel Club(AKC) strongly supports dan-gerous-dog control. Dog-con-trol legislation must be reason-able, non-discriminatory, andenforceable as detailed in theAKC Position Statement.

To provide communitieswith the most effective danger-ous-dog control possible, lawsmust not be breed specific.Instead of holding all dog own-ers accountable for their be-havior, breed specific lawsplace restrictions only on theowners of certain breeds of

37

dogs. If specific breeds arebanned , owners o f thesebreeds intent on using theirdogs for malicious purposes,such as dog fighting or crimi-nal activities, will simplychange to another breed ofdog and continue to jeopard-ize public safety.281

In response to a perceived rise inbreed discrimination, The HSUSand the ASPCA developed a grass-roots campaign to educate the in-surance industry.282 Both groupsoppose breed discrimination.283

Other groups that have spoken outagainst breed discrimination in-clude the American VeterinaryMedical Association, the AmericanDog Owners Association, the West-minster Kennel Club, and theAmerican Humane Association.284

IV. The Unique and Special Role of Pets in SocietyFor at least twelve thousand years,the history of the domestic dog,Canis familiaris, has been inter-t w i n e d w i t h t h a t o f h u m a nbeings.285 The law has generallytreated dogs as mere prop-erty286—or worse, as nonprop-erty.287 As the popularity of dogsas pets has grown, the law hasresponded in kind by recognizingthe importance of dogs, cats, andother pets. The insurance indus-try, by practicing breed discrimi-nation, has failed to appreciatethe unique and special role ofdogs to their owners and to soci-ety. This section is offered to pro-vide some context for the implica-tions of breed discrimination.This is a problem that has thepotential for affecting a large seg-ment of the population and forhaving damaging effects on themental, physical, and emotionalhealth of people.

The GrowingPopularity of PetsPopulationA study estimated that in 1998there would be 53.6 million dogsin the United States, a 2.1 percentincrease since 1991.288 Approxi-mately 34.3 percent of homes haveone or more dogs.289 Dog ownersare thus a significant portion ofthe United States population.They are also a significant pool ofcustomers (actual and potential)for insurers.

SpendingTo understand the scope andpower of the pet-owning popula-tion, consider the amount ofmoney that is spent on pets eachyear. In 1998, Americans spent$11.1 billion on veterinary carealone, a 61 percent increase from1991.290 There are more than 35“pet vacation resorts” where dogsand cats can go to be pampered.291

There are also more than 650 pet cemeteries in the UnitedStates, indicating the extent towhich owners will go to memorial-ize their pets.292

Dogs: Members of the American FamilyBreed discrimination ignores thereality that most pet owners con-sider their pets to be members oftheir immediate family.293 Indeed,this “coexistence has contributedsubstantially to humans’ quality oflife.”294 Dogs were initially domes-ticated to be work animals, assist-ing humans with farming, herdinglivestock, and providing security atnight.295 In time, dogs became“four-legged members of the fam-ily.”296 Some dogs provide assis-tance to humans with disabili-ties.297 Service dogs serve as atangible resource for people, notjust a source of companionship.298

Dogs can have positive effects onthe health of their owners,299 suchas alleviating loneliness and de-

pression, reducing high blood pres-sure, and addressing obesity.300 Onthe other hand, these effects mustbe balanced against the negativehealth effects of dogs, such as bitesand the transmission of zoonoticdiseases.301 When the positives areweighed against the negatives, atleast one physician has concludedthat dogs probably are beneficial tohuman health.302 Some owners willforgo their own health in order tocare for their pets—a demonstra-tion of how much pets mean tosome owners. “Most physicians arefamiliar with at least one exampleof a person refusing hospitaliza-tion...because there was no oneelse in the home to care for theirpet.”303

The loss of a pet can have pro-found effects on an owner. A num-ber of organizations providebereavement support for peoplewhose pets have died,304 and atleast three greeting card compa-nies make sympathy cards specifi-cally for the loss of a pet.305

Breed discrimination forces petowners to choose between theirhomes and their dogs. Forcingowners to make this choice repre-sents a significant misunderstand-ing of the role of pets in our soci-ety. For some pet owners, giving upa pet is like losing a child, sibling,or spouse.

The Consequencesof BreedDiscriminationWhen a dog bites, it can have lastingconsequences for both the dog andits owner’s family. When an insur-ance company refuses to insure orrenew a household based on a par-ticular breed of dog, it, too, can havefar-reaching consequences.

Most people do not respondappropriately if their dog bitessomeone. Most punishment is toosevere and too late to be of anyvalue to the dog in preventingfuture occurrences.306 The dog is



usually isolated from the familyand visitors. By limiting inter-action with humans, the dogdoes not learn how to deal withpeople appropriately.307 Isolationmay also lead to inadequate med-ical care, which may in turn leadto serious health problems for the dog.308

Some owners abandon theirdogs or euthanize them either outof frustration at not being able tocorrect aggressive behavior or be-cause an insurance company tellsthem to do so in order to gethomeowners’ insurance.309 WhenBSL goes into effect or insurancecompanies discriminate, it causessome owners to purposely assumea sheltered and low profile in thecommunity to avoid being caughtwith an unauthorized pet.310 Shel-ter drop-offs are common afterBSL goes into effect or insurersbegin to discriminate based onbreed.311 The humane society inAtchison, Kansas, reported a 40percent increase in drop-offs of rot-tweilers because of breed discrimi-nation.312 This is unfortunate be-cause many shelters can only keepdogs a certain number of daysbefore euthanizing them. Breeddiscrimination can have a chillingeffect on ownership of certainbreeds,313 which means certainbreeds are not likely to be adoptedand will have to be euthanized.

Breed discrimination will likelyhave an effect on homeownershipin states that permit this practice.Homeowners’ insurance is the“gatekeeper” to homeownership.Without homeowners’ insurance,a person cannot get a mortgage.Without a mortgage, most peoplecannot buy a house.314 An insuredwho chooses to lie about a dog’sbreed or the existence of a dogaltogether is committing policyfraud, running the risk of criminalprosecution315 and the completecancellation of his or her policy.316

Pets: More thanMere PropertyThe problem of breed discrimina-tion should be viewed in light ofmodern developments in animallaw, which is beginning to recognizethat animals are more than mereproperty. Until recently, the legalstatus of animals was governed byan 1897 Supreme Court case,Sentell v. New Orleans & C.R. Co.317

The case involved a Newfoundlandnamed Countess Lona who waskilled by a railroad car.318 Her ownerbrought suit against the railroad fornegligence. The railroad defendedby relying on a statute that prohib-ited an owner from recovering formore than the declared value on theanimal’s registration form.319 Anowner whose dog was not registeredcould not recover anything for theloss of or damage to the animal.320

Countess Lona’s owner broughtsuit, challenging the constitutional-ity of the law.321

The Supreme Court held that thestatute was constitutional as a validexercise of the state’s policepower.322 The Court declared thatdogs are a form of quasi-propertythat is “imperfect or qualified” innature.323 The Court relied on thecommon law rule that dogs couldnot constitute stolen property forpurposes of larceny statutes.324 Thecommon law held that wild animalshad no property value until killedor subdued.325 Domesticated ani-mals, such as horses, cattle, sheep,and other “work” animals, wereconsidered “perfect and complete”property.326 Dogs fell in a third cat-egory, that of “cats, monkeys, par-rots, singing birds, and similar ani-mals, kept for pleasure, curiosity,or caprice.”327 The Court saw nouseful, social value for dogs, exceptfor companionship, which theCourt dismissed as unsatisfactoryfor the establishment of a propertyinterest. Thus, the Court held thatproperty interests in animals are ona continuum: wild animals (animalsferae naturae) on one end, domes-

ticated animals (such as horses,cattle, and sheep) on the otherend, and dogs somewhere in be-tween.328 To the Sentell Court,dogs hold “their lives at the will ofthe legislature, and properly fall-ing within the police powers of theseveral states.”329 The Court con-cluded, “It is purely within the discretion of the legislature to sayhow far dogs shall be recognizedas property, and under what re-strictions they shall be permittedto roam the streets.”330

The question of the legal statusof dogs and other pets has recentlybeen addressed by courts in thecontext of family disputes. Bennettv. Bennett331 and Arrington v.Arrington332 typify the majorityrule with respect to the “custody”of pets upon their owners’ divorce.In both cases, divorcing couplessought both custody and visitationof their dogs. In Bennett the trialcourt awarded legal custody of thedog, Roddy, to the husband, withthe wife receiving every-other-weekend and holiday visitationrights.333 Subsequent squabblingbetween the parties led the Courtto modify its order to have the par-ties swap custody of the dog everymonth.334 The appellate courtreversed the trial court’s order andaffirmed the Sentell doctrine:“While a dog may be considered bymany to be a member of the family,under Florida law, animals are con-sidered to be personal property.”335

The court found that the trialcourt lacked authority to order vis-itation rights in mere property.336

The court in Arrington reacheda similar conclusion. Arrington in-volved a custody dispute over Bon-nie Lou, “a very fortunate littledog with two humans to showerupon her attentions and genuinelove frequently not received byhuman ch i l d r en f rom the i rdivorced parents.”337 The trialcourt had awarded custody of Bon-nie Lou to Mrs. Arrington. Mr.Arrington appealed, claiming heshould have been appointed “man-

39

aging conser vator” (primar yguardian) of Bonnie Lou.338 TheCourt held that managing conser-vatorships were designed for hu-mans, not animals.339 The Courtheld, “A dog, for all its admirableand unique qualities, is not ahuman being and is not treated inthe law as such....A dog is personalproperty, ownership of which isrecognized under the law.”340

There is an indication that thelegal status of dogs and other petsmay be beginning to change. InRaymond v. Lachmann, the courthad to determine the custody of acat named Merlin.341 The defen-dant originally owned Merlin, butleft him for one and a half yearswith a former roommate, the plain-tiff.342 During that time, the plain-tiff renamed him “Lovey” and grewto be quite attached to him.343 Thetrial and appellate courts bothheld that Lovey should remain inthe custody of the plaintiff, whohad taken care of him for a lengthyperiod of time. What is remarkableabout this case is that the courtused a “best interests of the cat”standard to decide the issue. Thecourt discarded strict applicationof property law and in its placeadopted a version of the “bestinterests of the child” standardfrom (human) family law. Thecourt held:

Cognizant of the cherishedstatus accorded to pets in oursociety, the strong emotionsengendered by disputes of thisnature, and the limited abilityof the courts to resolve themsatisfactorily, on the recordpresented, we think it best forall concerned that, given hislimited life expectancy, Lovey,who is now almost ten yearsold, remain where he has lived,prospered, loved, and beenloved for the past four years.344

Some courts have also recog-nized that pets are more thanmere property in the context oftort awards. In Corso v. CrawfordDog & Cat Hospital, Inc.,345 a New

York City civil court judge awarded$700 in damages to the owner of adeceased poodle. The dog hadbeen euthanized by the defendant,on instructions from the plain-tiff.346 “The plaintiff had arrangedfor an elaborate funeral...includ-ing a headstone, an epitaph, andattendance by plaintiff ’s two sis-ters and a friend.”347 When theplaintiff opened the casket, how-ever, she saw the body of a deadcat.348 She brought suit, allegingthat she had suffered emotionaldistress as a result of the inci-dent.349 The Court held that theplaintiff was entitled to sue notjust for the market value of thedog (for conversion of her prop-erty) but also for her mentalanguish and suffering in seeingthe cat instead of her dog. TheCourt stated:

This court now overrules priorprecedent and holds that a petis not just a thing but occupiesa special place somewhere inbetween a person and a pieceof personal property....A pet isnot an inanimate thing thatjust receives affection it alsoreturns it....To say that [thepoodle] is a piece of personalproperty and no more is a re-pudiation of our humanness.This I cannot accept.350

Dicta in other cases demon-strate that courts are beginning torethink the concept that pets aremere property. In Bueckner v.Hamel,351 the Texas Court ofAppeals had to decide the amountof damages to be awarded theowner of then-deceased dogs, adalmatian and an Australian shep-herd.352 The defendant shot thedogs while they were chasing adeer.353 The plaintiffs brought suitto recover damages for the loss oftheir property, which the trialcourt found “had special value tothe Plaintiffs and were loved aspets by the Plaintiffs.”354 Themajority concluded that “Texaslaw recognizes a dog as personalproperty”355—a holding consis-

tent with Sentell. The majoritywent on to hold that the plaintiffscould recover only for the loss ofvalue of prospective puppies butonly in the context of how muchthe animal itself would be worthas breeding stock.356

A concurring judge took abroader view of damages in thecase. He said the award for dam-ages should be based on “theintrinsic or special value of domes-tic animals as companions and be-loved pets.”357 The market valuewas inadequate to compensate theplaintiffs for the full extent of theirloss.358 “It is common knowledgeamong pet owners that the deathof a beloved dog or cat...can be agreat loss.”359 He called for theacknowledgment of pets as a spe-cial form of property360 based onthe relationship between humansand their pets:

Many people who love andadmire dogs as family mem-bers do so because of the traitsthat dogs often embody. Theserepresent some of the best ofhuman traits, including loy-alty, trust, courage, playful-ness, and love. This cannot besaid of inanimate property. Atthe same time, dogs typicallylack the worst human traits,including avarice, apathy, pet-tiness, and hatred....Losing abeloved pet is not the same aslosing an inanimate object,however cherished it may be.Even an heirloom of great sen-timental value, if lost, does notconstitute a loss comparableto that of a living being. Thisd i s t inct ion app l ies eventhough the deceased living be-ing is a nonhuman.361

Juries have been following thistrend. In cases where harm hadbeen done to pets, juries have beenawarding damages as high as$35,000. In contrast, the averageaward in the early 1990s was only afew hundred dollars.362



V. Breed DiscriminationShould Be EndedThrough Legislationor AdministrativeRegulationA central principle of insurancelaw is that insurance companiesoperate at the pleasure of thestates.363 “Indeed, the organiza-tion of an insurance company andthe conduct of the business of writ-ing insurance is not a right but aprivilege granted by the State sub-ject to the conditions imposed byit to promote the public wel-fare.”364 The power to regulate in-surance is so strong that a statemay take over the entire businessof insurance if it decides it is in thepublic interest to do so.365

States have the power to regulateinsurers as an exercise of their po-lice power.366 Although insurancelaw is governed in part by contractlaw,367 it is also quasi-public in na-ture.368 States have the power notjust to regulate insurance con-tracts, but also to declare theterms and conditions of those con-tracts and to impose additionalduties and obligations.369 On theother hand, when a state does notregulate a particular practice of theinsurance industry, companies arefree to contract as they see fit.370

States regulate and legislateinsurance on behalf of the publicinterest. Regulations counterbal-ance free market forces to protectthe public at large.371 Some statesprohibit unfair and deceptive tradepractices.372 Some administra-tively set rates.373 In determiningwhether a rate is reasonable, stateswill look to see if the rate is basedon “legitimate cost factors.”374

Some states require insurers towrite policies for particular risks,even though the marketplace mayhave determined such insureds are

poor risks or that they are simplyuninsurable.375

In 1997 D.S. Hellman evaluatedthe widespread practice of thetime of insurers in denying health,life, and disability coverage to vic-tims of domestic abuse.376 She pre-sented a compelling and detailedanalysis of the philosophical andlegal implications of this practice,ultimately concluding that statelegislatures should intervene andprevent underwriting decisionsbased on a customer’s history ofdomestic abuse.377

Hellman’s analysis started withthe premise that insurers had beenable to draw an actuarially justifiedconclusion that domestic abusevictims were, from a statisticalstandpoint, more likely than oth-ers to be victimized in the futureand, thus, to result in claimsagainst their insurers.378 Domesticabuse victims were a higher risk—so high, the insurers concluded,that the insurance pool could notbear to have them as a risk, nomatter how high the premium.379

Breed discrimination is an entirelydifferent problem altogether.There is a lack of statistically andscientifically sound data to showthat certain breeds are more dan-gerous than others. Even if suchdata existed, a plausible case couldbe made that the breed of a fam-ily’s dog should not be used as afactor in underwriting.

Insurers’ DutyIn making underwriting decisions,insurers decide which of manyrisks to insure in order to protecttheir fiscal solvency and profitabil-ity.380 When an underwriter de-cides not to insure a particularrisk, the would-be insured is left tofind insurance elsewhere. If noinsurer will underwrite or acceptthe risk, the result may be a cost-shifting to society381 or the lossof an economic opportunity to aconsumer.382

The question then becomeswhich factors an insurer may con-sider in making its underwritingdecisions. Insurance is a highlyregulated industry. It does notoperate in a regulatory vacuum,free to let the give-and-take of themarketplace decide who gets in-surance, how much coverage theyget, and how much it will costthem. There is social utility inmaking insurance available to thehighest number of people possi-ble.383 Insurance allows people tobuy homes, afford health care, anddrive automobiles.384 The highstakes and high social utility ofinsurance have historically justi-fied strict government regulationof the industry.385

All states require underwritingdecisions to be based on actuari-ally sound data.386 In Maryland, for example,

An insurer or insurance pro-ducer may not cancel or refuseto underwrite or renew a partic-ular insurance risk or class ofrisk for a reason based wholly orpartly on race, color, creed, sex,or blindness of an applicant orpolicyholder or for any arbi-trary, capricious, or unfairlydiscriminatory reason.387

Maryland law also provides thatunderwriting must be accom-plished “by the application of stan-dards that are reasonably relatedto the insurer’s economic and busi-ness purposes.”388

Actuarially justified underwrit-ing is not only the law, it is goodbusiness. By accurately separatingout risks into “not insurable” and“insurable” (and, then, in turn,separating out insurable risks intovarious risk classifications), actu-arially justified underwriting pro-motes efficiency and profit. Con-sumers are not allowed into theinsurance pool when the likelihoodof loss is so high that inclusion oftheir risks threatens the viability of the pool itself.389 For those in-sureds allowed in the pool, actuar-ially justified underwriting pro-

41

motes efficiency by assigning lowpremiums to low-risk insureds andhigh premiums to insureds morelikely to have a claim.390 This cre-ates a market incentive for low-riskinsureds to participate in the poolas opposed to engaging in adverseselection.391 Accurate risk classifi-cation also maximizes profits forthe insurer. By eliminating thehighest-risk insureds from thepool, an insurer keeps premiumslow for the low-risk insureds whoremain. An insurer that does notmaintain its “classification edge”faces the potential of having itslow-risk insureds leave to joinother companies that are able tocharge lower premiums due to bet-ter risk classification decisions.392

The insurer is stuck with its high-risk insureds as well as the high-risk insureds who migrate overfrom the insurer’s competition.393

This means that the insurer is notmaximizing its profitability.

How much statistical correlationis required for a rating factor to be“actuarially fair”? How legitimatedo “legitimate cost factors” haveto be?394 Certainly, perfect 1:1 cor-relation is not required.395 Thus, Ido not suggest that insurers mustbe able to demonstrate that everyChow Chow will have an unjust bitein its lifetime. Risk classificationnecessarily will involve some “falsepositives.”396 Otherwise, insurerswould be very limited in the classi-fications they could use, therewould be insufficient stratificationof the rate pool, and the dangers ofmoral hazard397 and adverse selec-tion398 would increase dramati-cally. On the other end of the spec-trum is the insurers’ position, thatany correlation is sufficient.399

This is not an economically viableposition for an insurer, since low-risk insureds may be incorrectlyclassified as high-risk customers,and high-risk insureds might bepriced out altogether.400 For exam-ple, my ownership of a rottweilerand a half-Chow put me in an irra-tionally high-risk classification—so

high that every insurer except theFarm Bureau declined to providecoverage. The dozen or so insurersthat I contacted in Lubbock whodeclined to provide coverage lostout on what would otherwise be alow-risk insured, simply becausethey adhered to a hypothesis (rott-weilers and Chow Chows are moredangerous than other dogs) thathas not been scientifically proven.In my case, the insurer who used amore actuarially sound rate classi-f ication structure (the FarmBureau) benefited by offering alow-risk consumer a low-risk pre-mium, thus gaining a market ad-vantage over its competition.401

I do not believe there exist suffi-cient data for an insurer to evenjustify a weak correlation betweenbreed and bite risk. Insurers shouldwork to minimize the risk of falsepositives so as to “fine tune” theirrisk classifications to the greatestextent possible.402 Risk classifica-tions should be sufficiently refinedso as not to be overbroad. Exclud-ing all dogs would clearly be over-broad and would come with highsocial costs. Excluding some breedsis also unsound, based on my reviewof the scientific literature.403 WhatI propose—and what the Task Forceon Canine Aggression and Human-Canine Interactions proposed404—is the refinement of breed-specificactions by legislatures and insurersto control and regulate “dangerousdogs.” Dangerous dogs are thosewho have demonstrated (on anindividual, dog-by-dog basis) a pro-pensity for violence. This would beactuarially fair because adequateevidence exists that a dog with ahistory of unjustified bites is likelyto be dangerous in the future.

As demonstrated in Part II, thereis insufficient evidence to supportthe insurance industry’s argumentthat certain breeds bite more often.In other words, the current riskclassification (by breed) is too gen-eral and is generating too manyfalse positives while at the sametime having unnecessary social

costs. A spokesperson for the IIIrecently conceded, “[T]he industryisn’t positioned to determine whichdogs should be deemed vicious....[W]e’re certainly not dog experts orveterinarians.”405 Unless and untilthe industry can demonstrate thatdifferent breeds have different rela-tive dangerousness ratios with somedegree of accuracy, breed discrimi-nation should be opposed by thegeneral public, insurers themselves,and regulators.

Arguments to SupportRegulation in thePublic InterestThe law is full of examples where“actuarially fair” factors have never-theless been prohibited in under-writing because of overriding publicinterests. Statistical correlationbetween behavior and risk, there-fore, is only the first step in a muchbigger, public policy analysis. Drive-through deliveries,406 preexistingmedical conditions,407 civil rights,408

and witness intimidation409 are allexamples of where otherwise actuar-ially justified practices were prohib-ited by state legislatures and courtsdue to overriding interests in equal-ity, health, and fairness.

Part IV demonstrated the impor-tance of dogs and other pets insociety. Pets provide physical andemotional benefits to humans andare not mere property. Even ifbreed discrimination were actuari-ally justified, I think a plausibleargument would exist that thepractice should be regulated be-cause of the public interest in pro-tecting animal-human bonds.

There is an additional, andarguably more important, socialvalue that is compromised bybreed discrimination: homeowner-ship. Most home buyers requirehomeowners’ insurance in order topurchase a home. This require-ment comes from mortgagors, who



require some protection in theevent their security (the homeitself) is destroyed, damaged, orotherwise made unavailable forcollection.410 As the Seventh Cir-cuit stated in NAACP v. AmericanFamily Mutual Insurance Co.,411

“No insurance, no loan; no loan, nohouse; lack of insurance thusmakes housing unavailable.”412

The issue in American Family wasa practice known as “red-lining”where homeowners’ insurancecompanies were charging higherrates, or declining to write insur-ance altogether, based on geo-graphic location of insureds.413

The boundaries (“redlines”) thatdefined the no-insurance zones fre-quently fell along racial and socioe-conomic lines, and the NAACPbrought suit alleging that thispractice was discriminatory andillegal. The Seventh Circuit heldthat red-lining violated the FairHousing Act, a statute passed byCongress to prohibit discrimina-tion in the housing market.414

It is quite possible that red-lin-ing was actuarially justified; that is,it may have in fact cost insurancecompanies more to write policiesin certain areas than others. This,however, did not end the inquiryfor Congress or the Court ofAppeals. The Seventh Circuit heldthat homeowners’ insurance is aservice that has the power to makehomeownership available.415 If aplaintiff can demonstrate that anapplication for homeowners’ insur-ance was rejected or unfairly ratedon the basis of race or another pro-hibited factor, the practice consti-tutes discrimination in housing.416

Homeownership is a worthwhilepublic interest. People who owntheir homes develop roots in agiven community. A homeowner isless likely to leave than is someonewho is in a year-to-year or month-to-month lease. The homeowner,therefore, has a personal invest-ment in the well-being of the com-munity. Homeownership providesan incentive for civic involvement

and community-wide improvement.For many families, homeownershipis the way to accumulate wealth forthe future.417 Home equity can beborrowed against for emergencies,higher education, or retirement.418

The family home is often the mostsignificant component in an estateafter a parent dies.419

Breed discrimination should,thus, be viewed in a larger socialcontext. There is a high social costwhen someone is denied homeown-ers’ insurance: he is unable to buya home.420 The social harm in pre-venting the dream of homeowner-ship must be weighed against thesmall risk of a dog-bite claim.There are over fifty million dogs inthe United States, yet only a fewdogs have been responsible for bit-ing people.

This is not a simple matter of de-ciding to throw away the familytrampoline or forgo the purchaseof an in-ground pool. Pets are notmere property. To make peoplechoose between the family pet andhomeownership is unfair, unneces-sary, and goes against an impor-tant social value: homeownership.

How Else Can InsurersControl Risk?Let me assume for the momentthat insurers could demonstratewith some degree of actuarial con-fidence that some breeds are morelikely to bite than others. Couldthere be other ways of controllingthis risk, short of outright denialof coverage?

ExclusionsWhen I was shopping for homeown-ers’ insurance, one of the firstquestions I asked insurers waswhether they would write a policywith an exclusion for dog bites. Idid this because I was desperate—I needed insurance and I was will-ing to assume the risk that mydogs were not dangerous and were

not likely to bite someone. Insur-ers still turned me away. They re-fused to write a policy with a dog-bite exclusion in it.

There are several good reasonswhy exclusions may not be goodpublic policy or wise businesssense. Exclusions operate to thedetriment of third parties, thosewould-be plaintiffs who are injuredand need compensation for theirloss. Exclusions would createpockets of plaintiffs who would, ineffect, have no way to satisfy ajudgment if they could prove liabil-ity. This is not an insignificantpublic policy, for the same reasonthat states require certain profes-sionals to have liability insur-ance421 and drivers to carry mini-mum limits on their automobilepolicies,422 to provide a source ofrecovery for third parties in theevent of a legitimate claim. If weexclude dog bites or even thosedog bites from breeds we canprove are the most dangerous, wewould run the risk of creating aspecial class of plaintiffs whowould have no source of recovery.Plaintiffs would have to turn toother sources in order to havetheir basic medical needs met.423

Exclusions are also bad for busi-ness because they make insuranceless attractive to consumers. A per-son with cancer is a much higherrisk than a healthy individual. If ahealth insurer began excludingcoverage for cancer treatment, fewemployers or individual consumerswould purchase that company’sinsurance. My decision to try tobargain my way into the insurancerisk pool by excluding dog bitesfrom coverage was, in reality,pretty stupid. In the rare eventthat I was found liable for one ofmy dogs biting someone, I wouldbe solely responsible for the judg-ment against me. I would losewhatever equity I had in my house,my car, my savings, and I couldhave my wages garnished. In retro-spect, an exclusion would not havebeen a good choice for me.

43

Insure but ReclassifyAnother option would be for insur-ers to write policies for familieswith “dangerous” breeds butcharge them higher premiums.Risk classification is an acceptedpractice in the insurance indus-try.424 By separating and groupingpeople of similar risks, insurerskeep rates low for the desirable,low-risk insureds, and insure ade-quate resources in the event thathigh-risk insureds cause a claim.425

I would have the same objection tohigh-risk classification for ownersof certain breeds as I would for out-right refusals to insure, that is, thelack of actuarial justification forthe practice of breed discrimina-tion. Classifying certain dog own-ers in a higher category is unfairbecause it places those insureds inan artificially higher rate bracket.This is economically inefficient,although perhaps more profitablefor the insurer.

Where I think risk classificationcould work is if insurers coulddemonstrate—to the veterinaryand CDC communities with a suffi-cient degree of scientific cer-tainty—that certain breeds, whenthey do bite, cause more damage.It is hypothesized, for example,that the jaw structure of pit bullscauses them to inflict more injurythan other breeds.426 This wouldstill be breed discrimination427

but, in my view, an acceptable formof risk classification...providedthere is a scientific/veterinarybasis for the conclusion. To date,the studies in this area have fo-cused on determining the numberof bites per breed, not the amountof damage per bite.

I believe insurers would also beactuarially justified in classifyinghomeowners based on whether ornot they own a dog, period. Onedoes not need to be an actuary tostate that a dog owner is morelikely than a non-owner to have abite claim against him. Insurerscould simply classify all dog ownersat a higher rate level because they

are more likely to have claimsagainst them. Let’s be clear: thisis not what is going on right now.The current practice of breed dis-crimination is to differentiateamong breeds, even though thereis no statistical evidence to provethat certain breeds are more dan-gerous than others. This creates anartificial risk classification thatcharges owners of certain breedsmore than others.

If all dog owners were classifiedat a higher rate than non-dog own-ers, I think there would be a greatpublic outcry. Then the social valueof dogs as pets—and as securityalarms on four paws—would cometo the forefront of the debate.

Allow the Marketplace to Correct ItselfIf, as I conclude, there are no reli-able data to support breed discrim-ination, then there is a market ofconsumers (owners of rottweilers,pit bulls, etc.) being overchargedor not served altogether. This cre-ates an economic inefficiency. Aninsurer with good business judg-ment would seek to corner thisunderserved market by writingpolicies with low-risk premiums.

There are a number of reasonswhy the market is not correctingitself. The number and identity ofpeople being affected by breed dis-crimination is unknown. Withoutthese data it would be difficult foran insurer to market itself to thoseconsumers. Also at work is the factthat insurers try to market them-selves to the lowest-risk con-sumers. Although these consumerspay lower premiums, they are re-sponsible for fewer claims. Everyinsurer tries to maximize its num-ber of low-risk insureds while max-imizing the number of high-riskinsureds who are serviced by itscompetitors.428 The insurance in-dustry as a whole appears to becaught up in this breed discrimina-tion hysteria. Individual companiesmay fear that the assumptionsbehind breed discrimination are in

fact true and therefore see littleincentive to market themselves topeople they view as high-risk. Forthese reasons, it is unlikely thatthe marketplace will correct itselfto end breed discrimination.

Other SolutionsPreventing law-abiding homeown-ers from obtaining insurance is notthe answer to the problem of dogbites. Better and more effectivealternatives exist.429

Collect Better DataAn initial first step would be toimprove surveillance and reportingof dog bites. Until accurate num-bers for the numerator and denom-inator in the relative dangerous-ness ratio can be ascertained,insurers and governments will bewithout realistic data on which tobase meaningful decisions. Theneed for more accurate data col-lection has been championed bythe very scientists who have triedto calculate the scope of the dog-biting problem.430 In addition,studies should be commissioned todetermine if certain breeds, whenthey do bite, cause more physicalinjury or damage.431

Enforce Existing LawsAgainst Dogfighting and Dogs at LargeThere are existing laws that, ifenforced more vigorously, couldreduce the number of dog bites.Dogfighting explains why somedogs are vicious. This undergroundindustry brings some dogs “to theverge of bloodlust.”432 By shuttingdown criminal organizations of ille-gitimate breeders, promoters, andowners, local governments couldtake a first step toward reducingbites by dogs that have been pur-posely bred to be dangerous.433 TheAKC and other groups support theuse of existing laws to break updogfighting rings.434

Many attacks appear to be causedby strays or dogs who have been per-



mitted to run off-leash.435 Theenforcement of existing laws against“dogs at-large” could reduce thenumber of bites.436 While these lawsexist in many places, they are notadequately enforced.

Owners are sometimes to blamefor socializing a dog to be danger-ous or for permitting it to get intosituations where it can causeinjury. Dogfighting, leash, and at-large laws address the root of theproblem, which is irresponsible dogownership. A dog is just as good ashis owner trains him to be. Oneproblem dog can be seized anddestroyed. One problem owner,however, can continually breed,adopt, or purchase dog after dog.Replacing one dog with another, orone breed with another, will nothelp to reduce the overall problemof owner irresponsibility.437 Exist-ing laws can and should be used toaddress the behavior.

Regulate Problem Dogs with Existing“Dangerous Dog” LawsSome dogs, as a result of socializa-tion (or lack thereof), bad tem-perament, or genetics, demon-strate that they are dangerous.They have a history of bites orattacks against people or otheranimals.438 By regulating theseindividual dogs, municipalities canfocus their efforts on the specificdogs likely to cause injuries in thefuture.439 Instead of targeting anentire breed, governments canaddress the handful of dogs whoare really the problem.

There are existing laws that per-mit local governments to regulate,or in some cases seize and destroy,dogs who have demonstrated apropensity to bite without justcause. Michigan enacted a statuteto permit local governments toseize “dangerous animals” andhave them tattooed, insured,fenced, sterilized, destroyed, “orany other action appropriate toprotect the public.”440 The statute

provides due process protectionsto the owner—requiring a hearingby a judge and a finding of danger-ousness before a disposition isordered.441 A dangerous animal isone who, without just cause,442

bites or attacks a person, or a dogwho bites or attacks and causesserious injury or death to anotherdog while the other dog is on theproperty or under the control ofits owner.443 Oklahoma has a simi-lar statute that allows for height-ened regulat ion o f an imalsdec lared dangerous by the irconduct,444 but prohibits localgovernments from enacting breed-specific legislation.445

“Most of the approximately 55million dogs in the United Statesnever bite or kill humans.”446 Dan-gerous-dog laws are narrowly tai-lored to address the real problem,which is the small percentage ofthe overall dog population that isresponsible for bites, injuries, anddeaths.447 Dangerous-dog lawsexist in many states. Insurers couldwork with local governments tofund additional animal-control offi-cers or work with owners of dan-gerous dogs to help take steps toprevent future dangerous acts.448

Educate the Public,Particularly Children,about Animal BehaviorInsurers and local governmentscould partner together to educatethe public about proper ways ofsocializing and approaching dogs.Proper training is essential for afamily with a new dog.449 Publiceducation about the importanceof neutering can reduce the inci-dence of dog bites450 because adisproportionate number of bitesare caused by intact dogs.451 Newowners should also be educatedabout the steps in picking theright dog for a household.452

“[T]here is no all-around bestbreed.”453 Certain breeds will bemore compatible with certaintypes of families.454

Children must also be educatedabout dealing with dogs safely.455 Atleast one study has demonstratedthe effectiveness of public educationas a way to improve children’s behav-ior around and toward dogs.456 Thestudy, conducted in Australia, exam-ined the reactions of children, agesseven to eight, to a dog that was tiedup in their playground.457 Half of thestudy group received a thirty-minuteclassroom lesson seven to ten daysbefore on how to safely approachand treat dogs.458 Researchers ob-served the reactions of the childrento the dog.459 The group that re-ceived the classroom instruction dis-played greater precautionary behav-ior than did the control group.While 79 percent of the controlgroup hastily patted the dog andtried to excite it, only 9 percent ofthe group that received instructiondid so.460

ConclusionWhile dog bites are serious eventsfor those who are bitten, the dog-bite problem is not the publichealth crisis that the insuranceindustry has made it out to be.Some perspective is in order. Thenumber of fatalities due to dogbites is very low when compared tothe number of people who die fromheart disease, cancer, accidents,suicide, and diabetes. Likewise,nonfatal bites are responsible for asmall number of injuries when com-pared to other accidental, uninten-tional injuries. Falls (11.5 million),motor vehicle accidents (4.3 mil-lion), drugs (3.3 million), sports(2.0 million), insect bites (1.7 mil-lion), bicycle accidents (1.4 mil-lion), poisoning (.7 million), andknives (.6 million) all individuallyoutrank dog bites (.5 million) aspublic health problems.461 Simi-larly, claims paid out by homeown-ers’ insurance companies for dogbites are miniscule when comparedto payouts for property damage.Damage due to fire, water, wind,and theft represents much larger

45

problems for homeowners’ insur-ance companies.

One way to eliminate the entireproblem of dog bites would be tooutlaw all dogs.462 Without dogs,there would be no dog bites and no dog-bite-related insuranceclaims.463 While this would result inan elimination of the perceivedfinancial burden to insurers, itwould not be “practical, realistic, ordesirable” to the average layman,scientist, or dog owner.464 Unless weas a society are willing to disregardthe social and health benefits ofdogs as pets, then we must be will-ing to accept a certain number ofbites. While “[t]he dog bite problemas a whole is not preventable, it iscontrollable.”465 Better alternativesto breed discrimination exist, suchas education and enforcement ofexisting dangerous dog laws.

With over 34 percent of house-holds owning at least one dog as apet, dogs have become valued four-legged members of our society. Tothe families that love them, petsare not mere chattel. Refusing towrite homeowners’ insurance poli-cies, therefore, should be a nar-rowly curtailed remedy, limited tothose families that own dogs whohave proven to be dangerous to lifeor property. The insurance industryhas chosen to paint with a verybroad brush. Breed discriminationis an overreaction, an attempt tosolve a small problem by prejudg-ing all dogs of certain breeds aslikely to be dangerous in the future.

When insurers develop under-writing standards and decide whichrisks to insure, they have a respon-sibility to the public interest. In-surers do not contract with con-sumers in a vacuum. A long historyof state regulation of the industryserves as a backdrop for this issue.Underwriting decisions should bethe product of reason, not specula-tion. In other words, if insurers aregoing to engage in breed discrimi-nation, they better have hard sci-ence to back up their practice.

The science behind dog bites isinconclusive at best. Most of thescientists authoring studies on dogbites have acknowledged that theirdata are incomplete and should notbe used to enact breed-specific leg-islation or to deny insurance tofamilies with certain dogs. No studyhas accurately or completely deter-mined the number of bites perbreed, or the number of dogs perbreed. Without these numbers, it isimpossible to compare breeds onthe basis of dangerousness. Insur-ers who are making judgmentsabout certain breeds are doing sowithout adequate scientific evi-dence. This is the Achilles’ heel ofbreed discrimination; by actingwithout adequate evidence, the in-surance industry has left itself opento regulation by the states.

State regulation is necessary tocorrect this injustice in the mar-ketplace. Insureds are being shutout of entire markets because ofthe near-hysteria that has grippedthe insurance industry. This is nota new phenomenon for the indus-try. In the past insurers have cutbenefits and denied applicationsfor insurance based on fiscal cost-benefit analyses that have had col-lateral social and health conse-quences. It was more costly tokeep new mothers in the hospitalfor forty-eight hours. Our societycame to the recognition, however,that discharging new mothers andtheir newborns within six hours ofdelivery was against public policy.Legislatures stepped in to correctthe injustice in the marketplace,knowing full well that it would costthe industry more money. Thesame should be done here.

To the insurance industry, breeddiscrimination reflects a belief thatdenying coverage to families withcertain breeds of dogs will savethem money. Insurers have notproduced scientific proof that dogsof certain breeds bite more oftenor cause more damage. The evi-dence simply does not exist be-cause of the problems of data col-

lection that I have highlightedhere. The irony is that insurers whoare practicing breed discrimina-tion are turning away good cus-tomers who pay premiums. Leg-islative action to correct thispractice will benefit both familieswith dogs and the shareholders ofinsurance companies.

Legislative action in this area isboth appropriate and necessary.What happened to me is happeningacross the country to thousands ofother families. To some insurers,dogs are mere property—like anold can of paint that can be leftbehind when a family moves. Thetruth is that dogs are members ofthe American family and deserve tobe treated as such. When familiesare forced to make the choicebetween owning a home and havinga dog, some have no choice at all;they must give up their beloved petto an animal shelter. There are doc-umented increases in “shelterdrop-offs” due to breed discrimina-tion. These animals cannot behoused indefinitely, so many haveto be euthanized.

The social cost to families is toomuch to ride on incomplete statis-tics and hunches by insuranceexecutives. Legislative action isnecessary. Luckily, many state leg-islators have become aware of thisproblem and have taken steps toend breed discrimination. Pennsyl-vania enacted a statute prohibitingbreed discrimination, which statesthe following:

No liability policy or suretybond issued pursuant to thisact or any other act may pro-hibit coverage from any spe-cific breed of dog.466

New York is considering legisla-tion that would outlaw breed dis-crimination as well. Bill 6761would prohibit insurers from refus-ing to issue or renew, canceling, orcharging or imposing an increasedpremium or rate for owning a dogof a specific breed.467 A New Hamp-shire bill would prohibit nonre-newal or cancellation of a policy



“based solely on the insured own-ing a certain breed of dog.”468

Other states should follow suit andenact legislation or administrativeregulations to prohibit the prac-tice of breed discrimination.

Notes1The Farm Bureau provides a number of

services to its members, including insuranceand banking. See Texas Farm Bureau,http://www.txfb.org (accessed July 7, 2004).

2On the other hand, renting would notnecessarily have been an easy task either.Many landlords prohibit certain breeds fromliving on their property or forbid dogs alto-gether. Obtaining renters’ insurance wouldalso have been difficult because of breed dis-crimination by insurance companies.

3See infra Part IV. for a discussion of theeffects of breed discrimination.

4See 42 U.S.C. §§ 3601-31 (2003); Mich.Comp. Laws § 37.2302 (2001); N.J. Stat. Ann.§§ 10:5-4, 10:5-9.1 (2002); see also NAACP v.American Family Mutual Insurance Company,978 F.2d 287, 297-98 (7th Cir. 1992) (FairHousing Act is applicable to allegations ofracially discriminatory red-lining).

5American Veterinary Medical Association,Task Force on Canine Aggression and Human-Canine Interactions, “A Community Approachto Dog Bite Prevention,” Journal of the Ameri-can Veterinary Medicine Association 218(2001), 1732, 1733 (hereinafter Task Force).

6See Richard Willing, (2000). “Under Law,Pets Are Becoming Almost Human,” USAToday (Sept. 13, 2000), 1A.

7Fred Bayles, “Pit Bullterriers: Too Fierceto Live? Call for Ban Follow Maimings,Death,” The Record (Dec. 30, 1985), B16.

8Ibid.9Ibid.

10Ibid.11Ibid.12Garcia v. Vill. of Tijeras, 767 P.2d 355,

362 (N.M. Ct. App. 1988).13Bayles, supra note 7, B16.14Ibid.15Bob Egelko, “Appeal for Murder Rap in

Dog-Maul Case/Attorney General Says JudgeHad No Right to Let Knoller Off So Easy,” SanFrancisco Chronicle (April 12, 2003), A15.

16Kenneth Phillips, “Diane Whipple Case,”http://www.dogbitelaw.com/PAGES/Whipple.html (accessed Jan. 8, 2004).

17Karyn Grey, Note, “Breed-Specific Legis-lation Revisited: Canine Racism or the Answerto Florida’s Dog Control Problems?,” NovaLaw Review 27 (2003), 415, 417 (describinga series of dog attacks in Florida); see Julie A.Thorne, Note, “If Spot Bites the Neighbor,Should Dick and Jane Go to Jail?” SyracuseLaw Review 39 (1988), 1445, 1445 (assert-ing that media attention fueled public out-rage against pit bulls); see generally Jeffrey J.Sacks et al., “Fatal Dog Attacks, 1989–1994,”Pediatrics 97 (June 1996), 891 (hereinafterCDC 1989–1994).

18For a list of communities that haveadopted or are considering BSL, see JanCooper, “Breed Specific Legislation,”http://www.rott-n-chatter.com/rottweilers/laws/

breedspecific.html (accessed June, 2004).19Ibid.20Ohio Rev. Code Ann. § 955.11(a)(1)(B)

(4)(a)(iii) (West 2004).21Ibid. § 955.11. For a general discussion

of Ohio’s regulation of pit bulls, See PaulaLynn Wilson, Note, “A.M. Sub. H.B. 352: AnOverview—Dogs Under Control,” Universityof Dayton Law Review 13 (1988), 297.

22Ohio Revised Code Annotated §955.22(C) (West 2004).

23Ibid. § 955.22(D)(2).24Ibid. § 955.22(E).25Denver, Colorado, Revised Municipal

Code § 8-55(a) (2003). The ordinance pro-vides several exceptions, including a grandfa-ther clause, possession by animal shelters orhumane societies, public exhibition, or trans-portation through the city. Ibid. § 8-55(c).

26Ibid. § 8-55(b)(2). For an analysis of theproblem of defining “pit bull,” see infra notes240–241 and accompanying text.

27Ibid. § 8-52(a)(1).28Ibid. § 8-52(c)-(d).29Fla. Stat. Ann. § 767.14 (West 2003).30See Grey, supra note 17, 418.31Minn. Stat. Ann. § 347.51(8) (West 2004).32See generally Russell G. Donaldson,

Annotation, “Validity and Construction ofStatute, Ordinance or Regulation Applying toSpecific Dog Breeds, Such as ‘pit bulls’ or‘Bull Terriers,’” 80 American Law Reports 4th70 (2004). For discussion and background onchallenges to BSL, see Thorne, supra note 17;Lynn Marmer, Comment, “The New Breed ofMunicipal Dog Control Laws: Are They Consti-tutional?” University of Cincinnati LawReview 53 (1984), 1067; Heather K. Pratt,Comment, “Canine Profiling: Does Breed-Spe-cific Legislation Take a Bite Out of CanineCrime?” Pennsylvania State Law Review 108(2004), 855.

Cases in which courts have upheld the con-stitutionality of BSL include: Vanater v. Vil-lage of South Point, 717 F. Supp. 1236 (S.D.Ohio 1989); Colorado Dog Fanciers, Inc. v.City and County of Denver, 820 P.2d 644(Colo. 1991) (en banc); State v. Peters, 534So. 2d 760 (Fla. Dist. Ct. App. 1988); Hearnv. City of Overland Park, 772 P.2d 758, 766(Kan. 1989); Garcia v. Village of Tijeras, 767P.2d 355 (N.M. Ct. App. 1988); State v. Ander-son, 566 N.E.2d 1224 (Ohio 1991); State v.Robinson, 541 N.E.2d 1092 (Ohio Ct. App.1989); American Dog Owners Association v.City of Yakima, 777 P.2d 1046 (Wash. 1989)(en banc).

The Massachusetts Supreme JudicialCourt, however, found a pit bull ban in Lynn,Massachusetts, to be unconstitutional. Ameri-can Dog Owners Association, Inc. v. City ofLynn, 533 N.E.2d 642, 646 (Mass. 1989).

33U.S. Constitution, amendment V.34Sallyanne K. Sullivan, “Banning the Pit

Bull: Why Breed-Specific Legislation Is Con-stitutional,” University of Dayton Law Review13 (1988), 279, 280–93.

35See, for example, Garcia, 767 P.2d at358. Since dog ownership is not a “fundamen-tal right,” BSL need only meet the “rationalrelationship” test to be constitutional. Sulli-van, supra note 34, 281.

36See, for example, Garcia, 767 P.2d at358–62.

37See, for example, ibid., 361.

38Ibid. 39See, for example, ibid., 361–63.40See, for example, Hearn v. City of Over-

land Park, 772 P.2d 758, 762–65 (Kan. 1989).41See, for example, ibid.42See, for example, ibid., 766.43See, for example, ibid.44533 N.E.2d 642 (Mass. 1989).45The ordinance prohibited the sale of pit

bulls and required the registration of any pitbulls within city limits. Ibid., 644. Transporta-tion of pit bulls was permitted only if they weremuzzled and leashed. Even then, they couldonly be transported to a veterinarian. Ibid.

46Ibid., 644. For further discussion of theproblem of identifying dogs by breed, see infraPart II.

47American Dog Owners Association, Inc.,533 N.E.2d, 644.

48Ibid., 646.49Ibid.50Ibid.51Ibid. The purpose of this article is not to

address the constitutionality of BSL, but ratherto focus on the insurance aspect of breed dis-crimination. For further discussion of BSL, seeLinda S. Weiss, “Breed-Specific Legislation in theUnited States,” Animal Legal and Historical WebCenter (2001), http://www.animallaw.info/articles/aruslweiss2001.htm (accessed June8, 2004); Randall Lockwood, “Humane Con-cerns About Dangerous Dog Laws,” Universityof Dayton Law Review 13 (1988), 267; Grey,supra note 17; Ohio Valley Dog Owners, Inc.,“OVDO Is Opposed to Breed-Specific Bans”(2003), http://www.canismajor.com/orgs/ovdo/bslho.html (hereinafter OVDO); CindyAndrist, Note, “Is There (and Should ThereBe) Any ‘Bite’ Left in Georgia’s ‘First Bite’Rule?” Georgia Law Review 34 (2000), 1343;Marmer, supra note 32, at 1067; Pratt, supranote 32, at 855; Sullivan, supra note 34, 279;Diane Blackman, “Practicality of Breed Specific Legislation in Reducing or Eliminat-ing Dog Attacks on Humans and Dogs,”http://www.dog-play.com/pitbull.html (accessedJuly 14, 2003).

52“Insurance Industry DiscriminatesAgainst Dogs,” CBS News: Evening News withDan Rather (CBS television broadcast, June 3,2003), available at 2003 WL 5212276.

53Ibid.54See, for example, Jeff Bertolucci, “Man’s

Best Friend but Insurers’ Foe; Their AssemblyBill Has Failed, But Dog Lovers Continue toRail Against Breed Discrimination,” Los Ange-les Times (June 6, 2004), K1 (discussing fail-ure of Bill AB 2399, prohibiting breed discrim-ination but allowing a higher premium andrequiring a discount for canines that pass anobedience test); Vincent J. Schodolski, “‘Bad’Dogs Put Costly Bite on Insurers, Homeown-ers,” Chicago Tribune (May 17, 2004), 1; JeffBertolucci, “Is Nothing Private? Home Insur-ers Ask About Everything from Rover toRolexes. And the Answers Matter,” Los AngelesTimes (May 9, 2004), K1; Allan Woods, “Rot-tweilers, Pit Bulls New Insurance Liability,”National Post (March 26, 2004), A3; MicheleDerus, “Dog Bites Giving Insurers Pause,” Mil-waukee Journal and Sentinel (Feb. 29, 2004),1F; Gloria Campisi, “Beware of Dog WhenSeeking Insurance; Some Firms Have ‘BadBreed’ Lists,” Philadelphia Daily News (Oct. 7,2003), 14; Ryan Slight, “Liability Factor Can

47

Hurt Homeowners,” Springfield News-Leader(Sept. 28, 2003), 6A; William Sweet, “Insurersin Doghouse with Some Pet Owners,” Spring-field Union-News (Aug. 19, 2003), A01; PurvaPatel, “A Bite to the Pocket, Home InsurersOften Charge Higher Premiums Because ofDogs,” Sun-Sentinel (Ft. Lauderdale, Florida)(Aug. 12, 2003), 1D; Charles Toutant, “Insur-ers Attempt to Leash Dog-Bite Claims: Small-Scale Nuisance Litigation Turning Into BigBusiness,” Connecticut Law Tribune 29 (Aug.11, 2003), 8; Jim Spencer, “HomeownersInsurance Rules Not for the Dogs,” Daily Press(Virginia) (Jan. 10, 2003), C1.

Breed discrimination is not confined toAmerican insurers, either. Reports have alsocome from Canada describing the practicethere by Canadian insurance companies. “Cal-gary Man Denied Home Insurance RenewalDue to the Type of Dogs He Has,” CanadianPress (March 25, 2004), available at 2004 WL74118306; “Some Insurance CompaniesWon’t Cover You if You Own Certain Dogs,”CTV News-PM (CTV television broadcast,March 24, 2004), available at 2004 WL60848476.

55See Patrick McMahon, “Dog Owners’New Policy: Bite Back,” USA Today (May 20,2003), 03a.

56Ibid.57Ibid. I wonder if breed discrimination is

encouraging homeowners, who are in a situa-tion like mine or the Craanens’, to engage inpolicy fraud by lying about their pet’s breed orby hiding the dog altogether. In Mutual Bene-fit Insurance Co. v. Lorence, No. 02-1734,2003 WL 1354845 (4th Cir. Mar. 20, 2003),homeowners lied on their insurance applica-tion about owning dogs. One of their dogs (apit bull) subsequently attacked a mother andher son, resulting in a claim against thehomeowners’ insurance. The insurer broughtsuit against the owners, seeking a declaratoryjudgment that it was not required to defendthe claim because of the misrepresentation.Ibid., *596. The courts, however, declined toexercise jurisdiction and instead deferred toongoing proceedings before the MarylandInsurance Administration. Ibid., 597. Theowners had filed a complaint with the MIAalleging that the insurance company’s prac-tice of breed discrimination is against Mary-land law. Ibid., 596–597.

58Stephanie Davis, “ASPCA/HSUS Cam-paign Tackles Insurance Industry,” DVM: TheNewspaper Magazine of Veterinarian Medicine34 (November 2003), 1. (“Some householdname insurance companies either have can-celed or refused to write homeowners’ poli-cies for individuals with certain dog breeds.”)

59See Brian Sodergren, “Insurance Compa-nies Unfairly Target Specific Dog Breeds,”http://www.hsus.org/ace/18624 (accessedJune 11, 2004) (perceived increase in thenumber of claims, or risk, exists for peoplewho own certain breeds of dogs).

60See infra Part I. for the insurance indus-try’s arguments in favor of breed discrimina-tion.

61Davis, supra note 58.62Sodergren, supra note 59.63Dog owners wishing to report incidents

of breed discrimination to The HSUS can doso at the following website: http://files.hsus.org/web-files/PDF/insurance_incident_form.pdf

(accessed July 5, 2004). See Sodergren, supranote 59.

64Davis, supra note 58.65Insurance Information Institute, “Home-

owners Insurance,” http://www.iii.org/media/facts/statsbyissue/homeowners/ (accessedJune 11, 2004) (hereinafter III, HomeownersInsurance).

66Insurance Information Institute, “DogBite Liability,” http://www.iii.org/media/hottopics/insurance/dogbite/ (accessed June11, 2004) (hereinafter III, Dog Bite Liability).

67III, Homeowners Insurance, supra note65.

68III, Dog Bite Liability, supra note 66. Theinsurance industry points to rising construc-tion costs and natural disasters, such as windand hail, as principal causes for the rise in pre-miums. Ibid.

69Davis, supra note 58, 36 (quoting Ale-jandra Soto, spokesperson for the InsuranceInformation Institute, a trade group of theinsurance industry).

70Dan Hattaway, “Dogs and Insurance,”Journal of the American Veterinary MedicalAssociation 210 (April 15, 1997), 1143.

71Sodergren, supra note 59. 72III, Dog Bite Liability, supra note 66.73Nationwide Insurance stated to USA

Today that it relies on CDC data to support itsbreed-specific policies. McMahon, supra note55, 03a.

74Infra Part II.75III, Dog Bite Liability, supra note 66.76Ibid. One insurance company, State

Farm Fire and Casualty Co., reported that in1995 it paid out $70 million on eleven thou-sand claims. That year, it insured a total of fivemillion dog-owning homes. See Hattaway,supra note 70, 1144.

77III, Dog Bite Liability, supra note 66.78Ibid. (emphasis added).79III, Homeowners Insurance, supra

note 65.80Ibid. A plausible argument could be

made that dog bites are a preventable risk, ina way that damage due to fire, mold, and theftare not. Property damage claims are the resultof true accidents—unforeseeable acts forwhich insurers have little to no control over.Dog bites are one small way in which insurerscan try to minimize risk.

In reality, big-ticket risks can be con-trolled. Insurers could decide, for example,not to insure a home unless it has smokedetectors. This would provide an incentive forpeople to install smoke detectors, whichwould, in turn, reduce the number of fire-related claims. Insurers could also justifyrefusing to write policies for homes wherethere is a high risk of wind or hail damage.Any risk is controllable, but only to the extentthat the insurance market can bear the loss ofthis business.

81As shown, supra, while the industry’saggregate figures may sound “scary,” theymisstate the scope of the dog-bite problem inthe larger context of total claims paid.

82III, Dog Bite Liability, supra note 66.83Davis, supra note 58, 36 (“We just know

that certain breeds, when they do attack, tendto cause a lot more damage when they do bite,not because they bite most often.”); RandallLockwood and Kate Rindy, “Are ‘Pit Bulls’ Dif-ferent? An Analysis of the Pit Bull Terrier Con-

troversy,” Anthrozoös 1 (1987), 2, 4 (damagecaused by pit bulls is generally more severedue to the attack behavior of the breed).

84Davis, supra note 58, 36 (quoting Ale-jandra Soto, spokesperson for the III).

85Ibid., 38. This policy of exclusion is not without problems, as it leaves plaintiffswithout a remedy if a dog bite does occur and subjects a homeowner to loss of the homeand bankruptcy. Kenneth Phillips, “Breed Specific Laws, Regulations and Bans,” avail-able at http://www.dogbitelaw.com/PAGES/breedlaws.html (accessed July 21, 2003)(Phillips is a lawyer and expert on dog-bitelaw). Litigation over exclusions can be costlyand lengthy and can lead to uncertainty in the marketplace.

86See Davis, supra note 58, 36 (“We don’tdiscriminate or deny coverage based on breedof dog.”); McMahon, supra note 55; Hattaway,supra note 70 (author works for State Farm).

87Franco Ordonez, “Pet Peeve: A New StateRegulation Prohibiting Owners of Some DogBreeds From Adopting Draws Fire,” BostonGlobe (Jan. 15, 2003), B1.

88Ibid.89Phillips, supra note 85.90Infra Part III.91Jeffrey J. Sacks et al., “Dog Bite-Related

Fatalities From 1979 Through 1988,” Journalof the American Medical Association 262(1989), 1489 (hereinafter CDC 1979–1988).

92CDC 1989–1994, supra note 17.93Centers for Disease Control, “Dog Bite

Related Fatalities—United States, 1995–1996,”Morbidity and Mortality Weekly Report 46(1997), 463 (1997) (hereinafter CDC1995–1996).

94Jeffrey J. Sacks et al., “Breeds of DogsInvolved in Fatal Human Attacks in the UnitedStates Between 1979 and 1998,” Journal ofthe American Medical Association 217 (2000),836 (hereinafter CDC 1979–1998).

95Ibid., 838.96See infra Part II.97CDC 1979–1988, supra note 91, 1489.98Ibid. The NCHS uses reports from hospi-

tals and other sources to classify deaths.There is a category of deaths just for dogbites: “E906.0.” National Center for HealthStatistics et al., International Classification ofDiseases, 9th Revision, Clinical Modification(2003), http://ww.eicd.com.

99CDC 1979–1988, supra note 91, 1489.The HSUS reports are particularly detailedand are believed to be the most accurate indetermining the true breed of a biting dog.CDC 1989–1994, supra note 17, 891. HSUSstaff reviewed police reports, animal controlreports, statements by dog owners and vic-tims, and photographs. Ibid.

100CDC 1979–1988, supra note 91, 1490;CDC 1989–1994, supra note 17, 892.

101CDC 1979–1988, supra note 91, 1490.102Ibid. Between 1989 and 1994, that fig-

ure had dropped to 56.9 percent. CDC1989–1994, supra note 17, 892. From 1995through 1996, that number climbed to 80percent. CDC 1995–1996, supra note 93,463. From 1997 through 1998, the numbersettled at 70 percent. CDC 1979–1998, supranote 94, 837.

Victims were often elderly, suggesting that,“The main victims of fatal dog bites were thevery young and very old, those least able to



protect themselves.” CDC 1979–1988, supranote 91, 1492.

103CDC 1979–1988, supra note 91, 1490. 104CDC 1995–1996, supra note 93, 463.105Ibid. 106Ibid.107CDC 1979–1998, supra note 94, 839.

The Sacks/Sinclair study attempted to piecetogether the data on attacks by breed fromthe previous three studies as well as new datafrom 1997 and 1998. Ibid., 837.

108CDC 1979–1988, supra note 91.109Ibid.110CDC 1979–1998, supra note 94, 837.111Ibid.112Ibid. This accounted for approximately

28 percent of the deaths during the studyperiod.

113CDC 1979–1988, supra note 91, 1492.114Ibid., 1489. For further discussion of

the problem of breed misidentification, seeinfra Part II.

115CDC 1979–1988, supra note 91, 1492.116CDC 1979–1998, supra note 94, 838.117Ibid. (Fatal attacks are easier to track

than nonfatal bites because fatal attacksresult in SCMT coding and are more likely tobe reported by the news media).

118CDC 1979–1988, supra note 91, 1492.119Ibid., 1491.120Ibid.121See supra note 101 and accompanying

text.122CDC 1989–1994, supra note 17, 894.

They estimated that, in 1994, 1.8 percent ofthe population was bitten by a dog and 0.3percent of the U.S. population sought somemedical care for a dog bite.

123Centers for Disease Control, “NonfatalDog Bites-Related Injuries Treated in HospitalEmergency Departments—United States,2001,” Morbidity and Mortality Weekly Report52 (July 4, 2003) 605 (hereinafter CDC Nonfatal).

124Ibid. The NEISS-AIP is operated by theU.S. Consumer Product Safety Commission.

125Ibid.126Ibid. Deaths were excluded from the

study. 127Ibid.128Ibid. 129Ibid.130Ibid.131Ibid. 132Ibid., 607.133Ibid.134Ibid.135Ibid.136See infra Part II.137CDC Nonfatal, supra note 123.138See infra Part II. for a discussion of what

I call “just cause” bites.139Kenneth A. Gershman, Jeffrey J. Sacks,

and John C. Wright, “Which Dogs Bite? ACase-Control Study of Risk Factors,” Pedi-atrics 93 (1994), 913. Two of the threeauthors, Gershman and Sacks, were with theCDC at the time.

140Ibid., 913.141Ibid., 914.142Ibid.143Ibid., 913.144Ibid. To create a control group, the

authors used the first five digits of a biting-dog owner’s phone number and randomized

the last two digits. They called numbers untilan eligible control dog was found.

145Ibid., 915.146Ibid., 914; see also supra notes 25–28

and accompanying text.147Gershman, supra note 139, 916.148Ibid., 914.149Ibid., 916.150Jeffrey J. Sacks, Marcie-jo Kresnow, and

Barbara Houston, “Dog Bites: How Big a Prob-lem?” Injury Prevention 2 (1996), 52 (two ofthe three scientists who conducted the surveywere with the CDC at the time).

151Ibid., 52.152Ibid.153Ibid.154Ibid.155Ibid., 53.156Ibid.157Ibid.158Ibid.159Harold B. Weiss, Deborah L. Friedman,

and Jeffrey H. Coben, “Incidence of Dog BiteInjuries Treated in Emergency Departments,”Journal of the American Medical Association279 (1998), 51.

160CDC Nonfatal, supra note 123.161Weiss, supra note 159, 51.162Ibid.163Ibid., 52.164Ibid.165Ibid.166Ibid., 53. The conclusions about the

monetary impact of dog bites were confirmedby a CDC study in 1999. See Kyran P. Quinlanand Jeffrey J. Sacks, “Hospitalization for DogBite Injuries,” Journal of the American Med-ical Association 281 (1999), 232. A sample of904 hospitals in seventeen states found that5,991 hospital discharges in 1994 were theresult of dog bite injuries. The average lengthof stay was 3.6 days, costing a total of $164.9million in direct care per year.

167In this respect, it suffers from many ofthe same flaws as the CDC nonfatal bite studydiscussed supra in Part II.A.2.

168Jeffrey R. Avner and M. Douglas Baker,“Dog Bites in Urban Children,” Pediatrics 88(1991), 55.

169Ibid., 56.170Ibid. Consequently, the authors con-

cluded that in such cases the patient’s accu-racy of breed identification should be high.Ibid., 57. I disagree with this conclusion. Justbecause Little Johnnie knows Spot from theneighborhood does not necessarily mean thatJohnnie knows with any certainty what breedSpot is.

171Ibid, 56.172Ibid., 57.173Yue-Fang Chang et al., “Dog Bite Inci-

dence in the City of Pittsburgh: A Capture-Recapture Approach,” American Journal ofPublic Health 87 (1997), 1703.

174Ibid., 1703.175Ibid., 1703–1704.176Ibid., 1704.177Ibid.178Alan M. Beck and Barbara A. Jones,

“Unreported Dog Bites in Children,” PublicHealth Report 100 (1985), 315.

179The authors described previous studiesas relying principally on official reports ofbites. Ibid., 316. This resulted in significantunderreporting of bites.

180Ibid., 317. The survey included childrenin preschool to twelfth grade.

181Ibid., 319. The study went on to findthat children’s attitudes toward dogs were notaffected by being bitten: “Children appear toaccept being bitten by dogs much as they doother accidents such as falling off a bike.Being bitten had little influence on their lik-ing for dogs.”

182See Ibid., 317–320.183Lee E. Pinckney and Leslie A. Kennedy,

“Traumatic Deaths From Dog Attacks in theUnited States,” Pediatrics 69 (1982), 193.Only 48 percent of the newspapers responded.

184Ibid.185Ibid., 194.186Ibid.187Ibid.188CDC 1979–1988, supra note 91, 1489.189Pinckney and Kennedy, supra note 183,

195.190Ibid.191Ibid.192See infra Part II.193See Pinckney and Kennedy, supra note

183, 195.194William G. Winkler, “Human Deaths

Induced by Dog Bites, United States, 1974–75,”Public Health Report 92 (1977), 425.

195The CDC authors criticized the Winklerstudy as being “anecdotal” rather than “sys-temic.” CDC 1979–1988, supra note 91, 1489.

196Winkler, supra note 194.197Ibid., 428.198Ibid. 199Jason W. Stull and Robert R. Hodge, “An

Analysis of Reported Dog Bites: ReportingIssues and the Impact of Unowned Animals,”Journal of Environmental Health 62 (2000), 17.

200Ibid., 19.201Ibid. The authors were truthful when

they stated, “[f]or cases in which breed wasgiven, pit bulls, German shepherds, and rot-tweilers combined were responsible for over59 percent of bites each year.” Nevertheless,as a lawyer, I am concerned by scientists draw-ing this conclusion. Such statements can betaken out of context and used by the insur-ance industry and legislatures to justify breeddiscrimination and BSL when such actions arenot supported by statistics.

202Lockwood and Rindy, supra note 83, 7(“The genetics of canine aggression are stillpoorly understood...”).

203Ibid.204Task Force, supra note 5, 1736, (citing

J.C. Wright, “Canine Aggression Toward Peo-ple: Bite Scenarios and Prevention,” Veteri-nary Clinics of North America: Small AnimalPractice 21 (1991), 299.

205Lockwood and Rindy, supra note 83, 7(discussing the factors that can lead a dog tobe aggressive).

206See infra Part V. for a discussion of dan-gerous dogs laws as an alternative to BSL andbreed discrimination.

207CDC 1979–1998, supra note 94, 838.At the time of the study, Sacks and Gilchristwere with the CDC’s National Center forInjury and Control, Division of UnintentionalInjury Prevention; Sinclair and Lockwoodwere with The HSUS; and Golab was with theAmerican Veterinary Medical Association.

208Task Force, supra note 5, 1733.209“Dalmatian Popularity May Spur

49

Increase in Bites, HSUS Says,” DVM: TheNewsmagazine of Veterinary Medicine (Jan.1997), 40.

210The CDC authors noted, “[c]onsideringAmerican Kennel Club registration data forrottweilers in parallel with fatality data forthat breed indicates that as the breed hassoared in popularity, so have rottweiler-relateddeaths.” CDC 1979–1998, supra note 94,838–839.

211Pinckney, supra note 183.212Id. at 195.213Weiss, supra note 159, 51.214Supra Part II.A.215CDC 1979–1998, supra note 94, 838.216Ibid.217Task Force, supra note 5, 1733.218Ibid.219See supra notes 137–138 and 167 and

accompanying text.220Chang, supra note 173, 1704.221See supra notes 150–158 and accompa-

nying text.222Ibid.223U.S. Constitution, art. I, § 2, cl. 3.224Department of Commerce v. United

States House of Representatives, 525 U.S. 316,322 (1999).

225One study estimated that as of 1998approximately 34.3 million homes had dogsand a total of 53.6 million dogs in the UnitedStates being kept as pets. See J. Karl Wise andJih-Jing Yang, “Dog and Cat Ownership,1991–1998,” Journal of the American Veteri-nary Medical Association 204(8) (1994),1166, 1166–67.

226Supra Part II.227See, for example, Pinckney, supra note

183; Winkler, supra note 194. The Winkler andPinckney/Kennedy studies were also criticizedby Sacks et al. for taking a more anecdotal,rather than systematic, approach to determin-ing the relative dangerousness of breeds. SeeCDC 1979–1988, supra note 91, 1489.

228John C. Wright, “Canine Aggressiontoward People: Bite Scenarios and Preven-tion,” Veterinary Clinics of North America:Small Animal Practice 21(2) (1991), 299, 301.

229CDC 1979—1998, supra note 94, 839.230David Brand, “Time Bombs on Legs,”

Time (July 27, 1987), 60. Call me cynical, butI cannot envision drug dealers or gang mem-bers registering their dogs with the localmunicipality or the AKC.

231American Kennel Club, “Register a Dog,”http://www.akc.org/registration/registeradog.cfm(accessed July 9, 2004).

232While there were approximately fifty-seven million dogs in the United States in1990, only twelve million were registered withthe AKC. Mark Derr, “The Politics of Dogs,”Atlantic Monthly (March 1990), 49, 50.

233See ibid.234Lockwood & Rindy, supra note 83, 3.235Ibid.236Derr, supra note 232, 51.237Ibid., 52 (discussing interview with Joe

W. Templeton, professor of veterinary pathol-ogy and genetics at Texas A&M University, inwhich Professor Templeton stated that scien-tists cannot distinguish between breeds byusing genetic fingerprinting or examiningchromosomes); American Dog Owners Asso-ciation, Inc. v. City of Lynn, 533 N.E.2d 642,644 (Mass. 1989) (“After trial, the [trial]

judge found that there is no scientific means,by blood, enzyme, or otherwise, to determinewhether a dog belongs to a particular breed,regardless of whether ‘breed’ is used in a for-mal sense or not.”) Professor Templetondescribed one study in which two AmericanStaffordshire terriers were compared with awhippet. The profile of one of the terriersmore closely matched the whippet than theother terrier.

238CDC 1979–1998, supra note 94, 839. 239However, only some dogs have their

“AKC papers.” See supra notes 231–233 andaccompanying text.

240Katharine Dokken, “Dog Bite Statistics: Bad Logic,” http://www.thedogplace.com/library/articles156.htm (accessed June 1, 2004).

241OVDO, supra note 51; American DogOwners Association, Inc., 533 N.E.2d at 644.

242Wright, supra note 228, 301; TaskForce, supra note 5, 1733; Gershman, supranote 139, 916.

243CDC 1979—1988, supra note 91,1492.

244Dokken, supra note 240.245Lockwood and Rindy, supra note 83, 2.246“Find the Pitbull,” http://www.pitbulls

ontheweb.com/petbull/findpit.html (accessedJune 1, 2004) (“Breed misidentification is ascary thing in a time breed specific legislationis growing....Pit [B]ull dogs are often blamedfor dog attacks that may very well have beencaused by another breed”).

247CDC 1979–1998, supra note 94, 838(“even experts may disagree on the breed of aparticular dog”).

248533 N.E.2d 642 (Mass. 1989).249Ibid., 644.250CDC 1979–1998, supra note 94, 838

(stating that it is unclear how to countattacks by mixed-breed dogs).

251Ibid., 839.252Task Force, supra note 5, 1736.253Dokken, supra note 240 (“To go by sta-

tistics alone assumes that the majority of dogbites are reported and that the majority ofbreeds identified are correct”).

254There is reason to believe that incidentsof “just cause” bites are not uncommon. TheCDC described a series of incidents wherenonfatal dog bites that involved “just cause.”Stated conversely, these are bites where thehuman “victim” is to blame. CDC Nonfatal,supra note 123.

255Dokken, supra note 240.256Task Force, supra note 5, 1742.257Brand, supra note 230.258Ibid.259See, for example, ibid.260Ibid.261Ibid.262Ibid.263Brand, supra note 230.264Lockwood and Rindy, supra note 83, 3.265Ibid.266OVDO, supra note 51.267Brand, supra note 230, 60.268Ibid.269Ibid.270Task Force, supra note 5, 1736.271Ibid., 1732.272See supra Part II.273Task Force, supra note 5, 1736.274Ibid.275Ibid.

276Ibid.277Ibid., 1733.278CDC 1989–1994, supra note 17, 894

(“Although several breeds appear over-repre-sented in the population of animals involvedin fatal attacks, this representation fluctuatesover time. Thus, it may be unproductive toview this as a problem that is unique to anyone breed”); CDC 1995–1996, supra note 93(“Although some breeds were disproportion-ately represented in the fatal attacksdescribed in this report, the representation ofbreeds changes over time. As a result, target-ing a specific breed may be unproductive; amore effective approach may be to targetchronically irresponsible dog owners”); CDCNonfatal, supra note 94 (arguing for regula-tion of individual dogs over BSL).

279CDC Nonfatal, supra note 123; Lock-wood and Rindy, supra note 83, 2 (describingBSL as “controversial” and attributing theproblem to a lack of “good data” for thenumerator and denominator).

280American Kennel Club, “Canine Legisla-tion Position Statements,” http://www.akc.org/love/dip/legislate/canleg.cfm (accessed July2003).

281American Kennel Club, “American Ken-nel Club Statement on Dangerous Dogs,” http://www.akc.org/love/dip/legislate/dangerous.cfm (accessed June 8, 2004).

282Davis, supra note 58, 1.283Ibid.284Weiss, supra note 51; Lockwood, supra

note 51, 276.285Weiss et al., supra note 159, 51; Task

Force, supra note 5, 1733. In contrast, thedomestic cat has been with us for only aboutfour thousand years. Gina Spadafori and PaulD. Pion, Cats for Dummies, 2d ed. Hoboken,N.J.: For Dummies, 2000, 14.

286See infra Part IV.287Ibid.288Wise and Yang, supra note 225.289Ibid.290Willing, supra note 6. 291Ibid.292Ibid.293Task Force, supra note 5, 1739.294Ibid., 1733.295Ibid., 1739.296Ibid.297Susan D. Semmel, Comment, “When

Pigs Fly, They Go First Class: Service Animalsin the Twenty-First Century,” Barry LawReview 3 (2002), 39.

298Ibid., 44.299CDC 1995–1996, supra note 93, 466

(“Dogs provide many health and social benefits”).300David T. Allen, “Effects of Dogs on

Human Health,” Journal of the American Vet-erinary Medical Association 210(8) (1997),1136, 1138.

301Ibid.302Ibid.303Task Force, supra note 5, 1740.304See, for example, “In Memory of Pets,”

http://www.in-memory-of-pets.com/ (accessedJuly 21, 2004); Association for Pet Loss andBereavement, http://www.aplb.org/ (accessedJuly 21, 2004).

305Willing, supra note 6.306Wayne Hunthausen, “Effects of Aggres-

sive Behavior on Canine Welfare,” Journal ofthe American Veterinary Medical Association



(1997), 1134.307Ibid.308Ibid., 1135.309Ibid.310Ibid.311Sodergren, supra note 59; McMahon,

supra note 17, 03a.312Sodergren, supra note 59.313Ibid.314See infra Part V.315Robert R. Googins, “Fraud and the

Incontestable Clause: A Modest Proposal forChange,” Connecticut Insurance Law Journal2 51 (1996), 69 n. 86.

3162 Couch on Insurance § 31:10 (3d ed.2004).

317Sentell v. New Orleans & C.R. Co., 166U.S. 693 (1897).

318Ibid.319Ibid.320Ibid.321Ibid., 694.322Ibid., 696.323Ibid., 694.324Ibid. The Court noted, however, the

dogs can be considered property for purposesof the tort of conversion.

325Ibid.326Ibid. 327Ibid. 328Ibid., 694–695. 329Ibid. The Court’s distinction among

full, incomplete, and no status is interesting,but ultimately irrelevant, as the Court heldthat even if dogs were considered completeproperty, they (like all forms of property) aresubject to the police power of the state.

330Ibid., 696331Bennett v. Bennett, 655 So. 2d 109 (Fla.

Dist. Ct. App. 1995).332Arrington v. Arrington, 613 S.W.2d 565

(Tex. App. 1981).333Bennett, 655 So. 2d at 110.334Ibid.335Ibid. 336Ibid. 337Arrington, 613 S.W.2d at 569.338Ibid. In Texas family law a “managing

conservator has the right to establish thechild’s residence and has primary custody ofthe child. A possessory conservator typicallyhas visitation rights under terms and condi-tions set by the court.” In re V.L.K., 24 S.W.3d338, 340 n.1 (Tex. 2000) (citations omitted);see generally Texas Family Code Annual(2004) §§ 153.132, 153.371, 153.192.

339Arrington, 613 S.W.2d at 569.340Ibid.341Raymond v. Lachmann, 264 A.D.2d

340 (N.Y. App. Div. 1999).342Ibid. 343Ibid.344Ibid., 341.345Corso v. Crawford Dog & Cat Hospital,

Inc., 97 Misc.2d 530 (NY. Civ. Ct., QueensCounty 1979).

346Ibid.347Ibid.348Ibid., 530–31. 349Ibid., 531.350Ibid. 351Bueckner v. Hamel, 886 S.W.2d 368,

370 (Tex. App. 1994).352Ibid., 370.353Ibid.

354Ibid.355Ibid.356Ibid., 371.357Ibid., 373 (J. Andell concurring).358Ibid. 359Ibid., 376.360Ibid., 377.361Ibid., 377–378.362Willing, supra note 290.363Blue Cross & Blue Shield of Central New

York, Inc. v. McCall, 674 N.E.2d 1124, 1126(N.Y. 1996) (hereinafter Blue Cross & BlueShield). For further discussion of insurance asa “privilege” instead of a “right,” see 43,American Jurisprudence 2d ed., Title Insur-ance (2004) § 23. “The organization of aninsurance company and the conduct of thebusiness of writing insurance is not a rightbut a privilege granted by the state, subject tothe conditions imposed by it.” Ibid. § 24.

Regulation of the insurance industry is pri-marily a power of states, not the federal gov-ernment. 43 American Jurisprudence 2d ed.,(2004) §§ 29, 30. The McCarran-FergusonAct, 15 U.S.C. §§ 1011–1015, declared thatregulation of insurers by the states was in thepublic interest. 43 American Jurisprudence2d ed., (2004) § 30. McCarran-Ferguson cre-ates a reverse preemption system where thepower to regulate is vested in the states, notthe federal government. Ibid. However, planscreated under the Employee RetirementIncome Security Act (ERISA) are still subjectto joint federal-state regulation. John Han-cock Mutual Life Insurance Co. v. Harris Trust& Savings Bank, 510 U.S. 86 (1993) (“We aresatisfied that Congress did not order theunqualified deferral to state law that Hancockboth advocates and attributes to the federallawmakers. Instead, we hold, ERISA leavesroom for complementary or dual federal andstate regulation, and calls for federalsupremacy when the two regimes cannot beharmonized or accommodated”).

364Blue Cross & Blue Shield, 674 N.E.2d,1126.

36543 American Jurisprudence 2d ed.,(2003) § 25.

366Bekken v. Equitable Life Assurance Soci-ety, 293 N.W. 200, 211 (N.D. 1940); 43 Amer-ican Jurisprudence 2d ed., (2003) §§ 24, 25.

367Bekken, 293 N.W., 209.368Ibid., 210.369Ibid., 211. (“It will be noted that the

State, under its police power, in the interestof the general welfare, may regulate not onlythe business of insurance in a general way, butit may, also, regulate contracts of insurance,and in a large measure prescribe the termsand conditions of such contracts, and it mayimpose duties and obligations incident to therelation created by the contract or the nego-tiations for a contract different from thosearising or existing under other contracts, andit may prohibit the parties from altering suchduties or obligations.”). In Bekken, the NorthDakota Supreme Court had to decide whethera beneficiary of a life insurance policy couldenforce the policy even though the insurancecompany had not acted promptly in writingthe policy before the deceased died. TheCourt held that the insurance company hadbreached its duty to act promptly on thedeceased’s application and ordered the com-pany to pay the beneficiary. Ibid., 218.

One court has stated, “Such regulation,even down to ‘the minutest particular in theinterest of the public,’ has never been ques-tioned.” Blue Cross & Blue Shield, 674 N.E.2d,1126 (quoting People v. Formosa, 30 N.E. 492[N.Y. 1892]).

370Greer v. Aetna Life Insurance, 142 So.393, 395 (Ala. 1932) (“The generally recog-nized rule is that ‘in the absence of statutoryprohibitions discriminations or rebating as topremiums is not illegal’”). In Greer, the statecomplained that Aetna’s policy of not charg-ing higher premiums for mortgage insurancefor older people (who pose a greater risk) inessence discriminated against the young.Ibid., 394–395. Since state law did not pro-hibit discrimination based on age or lifeexpectancy, the Alabama Supreme Court heldthat the insurance company did not do any-thing illegal.

371Blue Cross & Blue Shield, 674 N.E.2d,1126 (stating that the conduct of insurancecompanies is subject to conditions imposedby the state to promote public welfare).

37244 C.J.S. Insurance § 45 (1993).373Ibid. § 66 (“The rates charged by insur-

ance companies must not deviate from thoseestablished by state authority. Rates may notbe unreasonable, excessive, inadequate, ordiscriminatory”). In states where rates areregulated, insurers bear the burden of provingthe reasonableness of their rates. Ibid. §69(b). Legislatures can prohibit discrimina-tion against insureds of the same “class.”Ibid. § 64. Some states provide a private causeof action to enable insureds to recover forunfair trade practices or violations of rate reg-ulations. Ibid. § 45.

37443 American Jurisprudence 2d ed., § 43(2003).

375Ibid. § 25. Most states require insurersto contribute to an “assigned risk plan” thatwrites “high-risk” policies for drivers who can-not get insurance in the free market, oftenbecause of DWI convictions. See, for example,Texas Insurance Code Annotated art. 21.81(2004–2005).

376Deborah S. Hellman, “Is Actuarially FairInsurance Pricing Actually Fair?: A Case Studyin Insuring Battered Women,” 32 HarvardCivil Rights and Civil Liberties Law Review(1997), 355.

377Ibid., 410.378Ibid., 356, 377–378.379See ibid., 358.380Deborah A. Stone, “The Rhetoric of

Insurance Law: The Debate over AIDS Testing,”Law and Social Inquiry 15 (1990), 385, 388.

381If a person cannot obtain health insur-ance, his or her health care costs shift to thepublic (through Medicare or Medicaid) or tohospitals, which may be required to treatevery patient regardless of the ability to pay.In the latter case, the costs are likewiseshifted to the public because hospitals mustincrease costs for paying customers (and theirinsurers) in order to make up for their lossesfrom their nonpaying customers. Ibid., 390.

382In the case of breed discrimination, thelatter results. If I had not been able to secureinsurance with the Farm Bureau, I would havelost out on a significant economic opportu-nity—the ability to purchase a house. Seeinfra Part V.B. for a discussion of insurance asgatekeeper to homeownership. In the alterna-

51

tive, I could have given up my dogs, whichlikewise would have been a significant loss ofhappiness and opportunity.

383See Stone, supra note 380, 386 (statingthat insurance is the primary mode peopleprovide for needs not affordable through nor-mal work income).

384Brian J. Glenn, “The Shifting Rhetoricof Insurance Denial,” Law and Society Review34 (2000), 779, 782.

385Lee R. Russ and Thomas F. Segalla,Couch on Insurance, 3d ed. (1995) § 2, 1.

386Sonia M. Suter, “Disentangling PrivacyFrom Property: Toward a Deeper Understand-ing of Genetic Privacy,” George WashingtonLaw Review 72 (2004), 737, 795 (“All statesrequire underwriting decisions to be actuari-ally, or rationally, based; they cannot be arbi-trary. Insurers must engage in good-faithpractices in deciding whether to underwrite,at what rate, and for what conditions.”); 43American Jurisprudence 2d ed., § 43 (2003)(underwriting and rate setting may take intoaccount only legitimate cost factors). To someextent, this requirement is based on a prohi-bition of unfair trade practices, as enumer-ated in the model Unfair Trade Practices Act,developed by the National Association ofInsurance Commissioners. See Unfair TradePractices Act (National Association of Insur-ance Commissioners, 2001) (hereinafterUTPA). The NAIC, a nongovernmental organi-zation of insurance regulators, drafted thismodel act to regulate unfair trade practices inthe insurance industry. Most states haveadopted UTPA in some form or another. SeeStone, supra note 380, 392; Richard J. Wirth,“My Customer’s Keeper: The Search for a Uni-versal Suitability Standard in the Sale of LifeInsurance,” New England Law Review 24(2002), 47, 79.

387Maryland Code Annotated - Insurance(2000)§ 27-501(a)(1) (emphasis added).

388Ibid. § 27-501(a)(2).389Tom Baker, “Containing the Promise of

Insurance: Adverse Selection and Risk Classi-fication,” Connecticut Insurance Law Journal9 (2002/2003), 371, 377.

390Otherwise insurers would run a risk oflow-risk consumers being priced out of themarket and only high-risk insureds remaining,a danger known as the “death spiral.” SeePeter Siegelman, “Adverse Selection in Insur-ance Markets: An Exaggerated Threat,” YaleLaw Journal 113 (2004), 1223, 1224.

391Baker, supra note 389, 373.392Kenneth S. Abraham, Distributing Risk:

Insurance, Legal Theory, and Public Policy(New Haven, Conn.: Yale University Press,1986), 67–68.

393Baker, supra note 389, at 377.394See, for example, 43 American Jurispru-

dence 2d ed., (2003) § 43.395Hellman, supra note 376, 380. 396Ibid., 378.397Moral hazard is the change in incen-

tives that can result because an individual isprotected by an insurance contract. Baker,supra note 389, 373. Insureds without a suffi-cient personal investment in a particularrisk—through co-insurance, for example—may be more likely to engage in the risky con-duct in a negligent manner.

398Adverse selection is the tendency forlow-risk insureds to drop out of insurance

pools altogether, unless administrators pro-vide financial incentives for them to remain inthe pools. Ibid., 375–376.

399See Hellman, supra note 376, 380.400See Ibid., 380–381.401Ibid.402Ibid., 380 (noting that “statistical dis-

crimination is justified when the social utility,including the cost to the insurer of employinga particular classification, could not beincreased by further refinement of the classifi-cation,” a view expressed by Steven Meitzen in“The Ethics of Statistical Discrimination,”Social Theory and Practice 17 [1991], 23, 26).

403Frederick Schauer argues that BSL andbreed discrimination are justified practices.Frederick F. Schauer, Profiles, Probabilities,and Stereotypes (Cambridge, Mass.: HarvardUniversity Press, 2003), 55–78. Schauer startswith the premise that there is a statistical cor-relation between pit bulls and unprovokedbites. Ibid., 55 (“Looking at what evidence wehave, it turns out that the generalizationsunderlying pit bull restrictions do indeed havethe kind of empirical support that distin-guishes them from purely spurious generaliza-tions”). He then goes on to argue that BSLand breed discrimination, while both underin-clusive and overbroad, are justifiable from theperspective of statistical correlation and theneed to protect people from dangerousevents. Ibid., 59. As I have demonstrated, how-ever, the correlation between breed and dan-gerousness is weak at best. See supra Part II.Schauer’s citation for his proposition is sev-eral of the CDC studies that have been dis-cussed in this article. Schauer, supra, at 59n.7; see supra Part II. Yet, these very studiescautioned that their data were incompleteand that it would not be statistically sound tomake generalizations about breeds. See supraParts II & III.

404Task Force, supra note 5, 1737.405Davis, supra note 58, 36 (quoting Ale-

jandra Soto, spokesperson for the III).406The practice of “drive-through deliver-

ies” required doctors to discharge new moth-ers within hours after giving birth. VickiLawrence MacDougall, “Medical Gender Biasand Managed Care,” Oklahoma City Univer-sity Law Review 27 (2002), 781, 882,892–894. States responded by passing lawsrequiring insurance companies to allow newmothers to stay in the hospital for a minimumperiod of time. Elizabeth C. Price, “The Evolu-tion of Health Care Decision-Making: ThePolitical Paradigm and Beyond,” TennesseeLaw Review 65 (1998), 619, 626. Connecti-cut, for example, passed a law that requiredinsurance companies to pay for forty-eighthours of inpatient care following vaginal deliv-eries and ninety-six hours following caesariandeliveries. Connecticut General Statutes Anno-tated §§ 38a-503c, 38a-530c (West 2004).Congress, using its power to regulate ERISA-based insurance policies, followed suit bypassing the Newborns’ and Mothers’ HealthProtection Act of 1996, which included the48-/96-hour rule that had been adopted byseveral states. Newborns’ and Mothers’ HealthProtection Act of 1996, Pub. L. No. 104-204,110 Stat. 2874, 2935 (codified at 29 U.S.C. §§1185[a][1][A][i]-[ii][2004]).

407Customers with preexisting medicalconditions represent a high-risk class of

insureds. They are people likely to need addi-tional care in the future. The marketresponded by charging higher rates for peoplewith preexisting medical conditions or flat-out rejecting them for coverage. The resultwas that many people found themselves with-out insurance when they switched insurers.See Grey, supra note 17, 421 (insurers aregiven wide latitude to refuse to cover certainhigh-risk customers, absent state legislationor regulation to the contrary). Some statesresponded by passing legislation prohibitinginsurers from refusing to provide coveragebased on preexisting medical conditions. Con-necticut requires health insurers to providecoverage for preexisting conditions if the per-son was previously covered by his or her pre-ceding plan. Connecticut General StatutesAnnotated §§ 38a-476 (West 2004).

408Legislatures have acted to protect con-sumers from discrimination on the basis ofrace, color, national origin, gender, and otherprotected classes. Maryland, for example, hasa relatively far-reaching statute prohibitingdiscrimination in underwriting based on race,color, creed, sex, blindness, or for any arbi-trary or capricious reason. Maryland CodeAnnotated, Insurance § 27-501 (2004).

States have enjoined insurers from charg-ing higher automobile insurance rates to meneven though actuarial statistics show thatwomen are a lower risk. For example, in Hart-ford Accident and Indemnity Company v.Insurance Commissioner, 482 A.2d 542 (Pa.1984), the Pennsylvania Supreme Court heldthat the practice of charging different ratesbased on gender violated the state’s statute,which prohibited “unfairly discriminatory”rates. Ibid., 549. The Court looked to the pur-poses of the act and insurance regulation ingeneral: to promote the public welfare. Ibid.,547. The Court held, “[P]ublic policy consid-erations require more adequate justificationfor rating factors than simple statistical corre-lation with loss.” Ibid., 548 (quoting NationalAssociation of Insurance Commissioners,Report of the Rates and Rating ProceduresTask Force of the Automobile Insurance (D3)Subcommittee, November 1978, 5–6 [foot-notes omitted]).

A Michigan court declared unlawful thepractice of refusing to write automobile insur-ance for adults under twenty-one unless theyresided with parents. Detroit Automobile Inter-Insurance Exchange v. Commissioner ofInsurance, 326 N.W.2d 444, 445 (Mich. Ct.App. 1982). The Court found that this prac-tice violated the state’s unfair trade practiceslaw, which permitted refusing to insure agroup “only if the cost is unreasonable.” Ibid.,447. The Court affirmed a central notion ofinsurance law that the free market is not thesole determinant of insurance rates. Thecourt stated, “The mere fact that one group ismore expensive to insure than another doesnot preclude fixing a reasonable rate.”

Plaintiffs have also relied on federal law forrelief against discrimination in the provisionof homeowners’ insurance. In National FairHousing Alliance, Inc. v. Prudential InsuranceCompany of America, 208 F. Supp. 2d 46(D.D.C. 2002), a U.S. district judge held thatdiscrimination by a homeowners’ insurancecompany on the basis of race, color, religion,sex, handicap, familial status, or national ori-



gin was made illegal by the Fair Housing Actand implementing regulations from HUD.Ibid., 55–56.

409At least one court has stated, as a mat-ter of public policy, that insurers may notengage in witness intimidation. In L’Orange v.Medical Protective Company, 394 F.2d 57 (6thCir. 1968), an insurance company cancelled adentist’s malpractice policy after he testifiedagainst another dentist who was insured bythe same company. Ibid., 59. The courtacknowledged that insurance policies aretreated as voluntary contracts, but noted thatthey are also subject to public policy con-cerns. Ibid. In this diversity-of-citizenshipcase, the Court looked to the law of Ohio andfound that the defendant-insurer had violatedOhio’s public policy against witness intimida-tion. The Court noted the need for expert tes-timony, the existence of statutes againstintimidating witnesses from testifying, andthe potential chilling effect of the defendant’sbehavior, and awarded judgment to the den-tist. Ibid., 61–63.

410Of course, if a person can purchase thehouse through cash on hand, then securing amortgage is unnecessary and obtaining home-owners’ insurance is “optional.” A prudenthomebuyer would nevertheless purchasehomeowners’ insurance to protect his or herinvestment in the event of catastrophic loss.

411978 F.2d 287 (7th Cir. 1992).412Ibid., 297.413Ibid., 290.414Ibid, 298.415Ibid., 297–298.416Ibid., 290–291.417David H. Harris, Jr., “Using the Law to

Break Discriminatory Barriers to Fair Lendingfor Home Ownership,” North Carolina CentralLaw Journal 22 (1996), 101.

418Ibid.419Ibid.420Unless, of course, he is independently

wealthy and does not need a mortgage.421Kansas physicians are required to carry

malpractice insurance. Kansas Statute Anno-tated §§ 40-3402 (2000); State ex rel. Schnei-der v. Liggett, 557 P.2d 221 (Kan. 1978)(upholding Kansas’s mandatory, state-run mal-practice insurance program). Oregon requiresattorneys to carry malpractice insurance. Ore-gon Revised Statutes § 9.080(2)(a) (2003).

422See, for example, New York InsuranceLaw § 5303 (McKinney 2003) (New York’sassigned risk plan).

423Hospitals, Medicare and Medicaid, andhealth insurers would thus bear the cost fortreatment. See Stone, supra note 380, 394(making a similar argument regardinghuman diseases).

424See ibid., 392.425Baker, supra note 389, 376–378.426See Lockwood and Rindy, supra note 83.427In this respect, the use of the term “dis-

crimination” is a bit of a misnomer. All formsof risk classification are acts of discriminationin a literal sense. See Hellman, supra note376, 378 (“Because all insurer classificationsare ‘discrimination,’ understood non-pejora-tively, one must ask why use of this classifica-tion is ‘plain, old fashioned,’ ‘profoundlyunjust[,] and wrong[ful]’ discrimination”).

428Baker, supra note 389, 377.429These are alternatives that could be col-

laborative endeavors among animal groups,governments, and insurers. At least oneinsurer, State Farm, has publicly stated itswillingness to find proactive solutions to theproblem of dog bites. See Hattaway, supranote 70.

430See, for example, Sacks et al., supranote 150, 53; CDC 1979–1998, supra note94, 840; Winkler, supra note 194, 425.

431Some scientists have suggested thatsome breeds are more dangerous because,when they do bite, their jaw structure andother physical characteristics cause them toinflict more pain and physical injury. SeeLockwood and Rindy, supra note 83, 36.

432Brand, supra note 230, 60.433Sacks et al., supra note 150, 53 (sup-

porting enforcement of existing laws to regu-late dangerous dogs and dog fighting); CDC1979–1998, supra note 94, 840 (urging law-makers to focus on problem owners, not dogsor breeds); CDC 1989–1994, supra note 17,894 (same).

434American Kennel Club, “American Ken-nel Club Statement on Dangerous Dogs,”http:// www.akc.org/love/dip/legislate/dan-gerous.cfm (accessed Nov. 6, 2004).

435CDC 1979–1998, supra note 94, 840.436See CDC 1989–1994, supra note 17,

894; CDC 1979–1998, supra note 94, 840.437CDC 1979–1998, supra note 94, 840

(“[P]roblem behaviors [of dogs and owners]have preceded attacks in a great many casesand should be sufficient evidence for preemp-tive action”).

438I would exclude from this category “justcause” bites, discussed supra at Part II.F.

439CDC 1979–1998, supra note 94, 840.440Michigan Compiled Laws §§ 287.321,

287.322 (2004).441Ibid. § 287.322.442Ibid. § 287.321(a). The law provides a

number of exceptions to the definition of dan-gerous. An animal is not dangerous if: (1) the“victim” was a trespasser or provoked or tor-mented the dog; or (2) the animal was pro-tecting a person or livestock.

443Ibid. § 287.321.444Oklahoma Statutes title 4, §§ 44, 47

(2004).445Ibid. § 46(b).446CDC 1995-1996, supra note 93, 466.447See CDC 1989–1994, supra note 17,

894-8995 (“[I]t is important to recognizethat most of the 52 million dogs in this coun-try never bite or kill anyone”).

448Frederick Schauer argues that a systemof individualized determination of dangerous-ness—that myself and others propose—isunsound. Schauer, supra note 403, 69–72. Hestates that such a system comes at a highsocial cost, since the State responds only aftera dangerous dog attacks. He cites severalexamples, including the speed limit on high-ways, to show that BSL and breed discrimina-tion are simply forms of acceptable, forward-looking regulation. While some drivers mightbe better than others, the State has set a max-imum speed limit regardless of driver ability.This is a forward-looking or prophylacticattempt to prevent accidents, death, andinjury before they happen. He also points tothe regulation of doctors and lawyers as anexample of how society legitimately controlsbehavior in advance in order to prevent dan-

gerous occurrences from happening in thefirst place. Schauer’s analysis, however, failsto recognize that dogs are different. Speedlimits come at a small social cost—driverswho can drive safely at faster speed limits areforced to get to their destinations later thanthey would have had the speed limits notexisted. This is a small social cost that comeswith a more significant, larger societal benefit(saving many lives). See Philip Shuchman, “ItIsn’t That the Tort Lawyers Are So Right, It’sJust That the Tort Reformers Are So Wrong,”Rutgers Law Review 49 (1997), 485, 523 (20percent increase in highway fatalities afterstates given permission by the federal govern-ment to raise speed limits to 65). Similarly, anunlicensed or untrained person practicing lawor medicine has an almost 100 percent cer-tainty of causing damage to clients orpatients. The social cost of not regulatingthose practices would be significantly high.Although the dog population is around fiftymillion, see supra notes 288–289 and accom-panying text, only a handful of those dogs willbite a person. Fatalities for dog bites have hov-ered around seven per hundred million peopleper year. See supra Part II.A.1. Yet, the socialcost of having forward-looking regulations—such as BSL and breed discrimination—comes at a very high cost to families that owndogs, particularly those that are seeking topurchase a home. See supra Part IV. Further,individualized, dangerous-dog prosecutionsdo have a prophylactic effect. Like tort law,dangerous-dog laws indirectly encourage own-ers to take reasonable steps to preventinjuries. In this respect, dangerous-dog lawscan serve as a deterrence against negligent orintentional misdeeds.

449See Benjamin L. Hart and Lynette A.Hart, “Selecting, Raising, and Caring for Dogsto Avoid Problem Aggression,” Journal of theAmerican Veterinary Medical Association 210(1997),1129, 1131.

450Ibid. (neutering can reduce aggression).451See Gershman et al., supra note 139,

914 (finding a statistically significant relation-ship between number of bites and intact dogs).

452Hart and Hart, supra note 449, 1130.453Ibid. 454Ibid. An aggressive dog, for example,

might do well with an assertive family. 455This conclusion has been supported by

several scientists. CDC 1989–1994, supranote 17, 894 (calling for education of bite vic-tims and children); Sacks et al., supra note150, 53 (education programs needed on dogbehavior); CDC 1979–1998, supra note 94,840 (education needed for children); Gersh-man, supra note 139, 916 (suggesting educa-tion programs for children as a method toreduce bites and other attacks); CDC1995–1996, supra note 93, 466 (discussingpublic education as a strategy towards reduc-ing bites); Task Force, supra note 5, 1739(education is key); Hunthausen, supra note306, 1135 (public education a necessary com-ponent of bite prevention). Education is alsosupported by the AKC. American Kennel Club,“American Kennel Club Statement on Danger-ous Dogs,” http://www.akc.org/love/dip/legislate/dangerous.cfm (accessed June 8,2004). Other groups support heightenedefforts to teach responsible dog ownershipand dog safety. OVDO, supra note 51. Educa-

53

tion is also supported by the insurance indus-try. Hattaway, supra note 70, 1144 (“I believethat the insurance industry has a role in pro-moting responsible pet ownership, includingeducation, to help reduce this national prob-lem”); III, Dog Bite Liability, supra note 66(recommending that homeowners educatetheir children not to approach a sleeping oreating dog).

456Simon Chapman, et al., “PreventingDog Bites in Children: Randomised Con-trolled Trial of an Educational Intervention,”British Medical Journal 320 (2000),1512–1513. Children are more likely thanadults to be the victims of dog bites. CDC1979–1988, supra note 91, 1490 (70 percentof fatal dog bites were in children under tenyears old); Sacks and Kresnow, supra note150, 52–53 (children account for approxi-mately half of all people who seek medicalattention for dog bites); CDC 1979–1998,supra note 94, 836 (most victims of fatal dogbites are children); CDC 1995–1996, supranote 93, 463 (80 percent of fatality victimswere children); CDC Nonfatal, supra note 123(children between five and nine years old aremost likely to be victims of nonfatal dogbites).

Some have speculated that children aremore likely to be victims than adults becauseof their small stature and inability to defendthemselves. See CDC 1979–1988, supra note91, 1492 (young and old are most at risk to bevictims of fatal dog attacks).

457Ibid., 1512.458Ibid. 459See ibid. 460Ibid., 1513.461Daniel M. Sosin et al., “Causes of Non-

fatal Injuries in the United States, 1986,”Accident Analysis and Prevention 24(6)(1992), 685, 686.

462Bonnie V. Beaver, “Human-CanineInteractions: A Summary of Perspectives,”Journal of the American Veterinary MedicalAssociation 210(8) (1997), 1148, 1148.

463That would, I suppose, lead more peo-ple to get cats as pets. However, there is evi-dence to suggest that cat bites are more dan-gerous than dog bites because of the high rateof infection associated with them. See LosAngeles County Animal Care & Control, “CatBites,” http://animalcontrol.co.la.ca.us/html/pages/poetownerinfo/Catbite.htm (accessedJune 9, 2004) (one million cat bites arereported each year in the United States; catbites can be especially dangerous for children,the elderly, or those with suppressed immunesystems); Cynthia B. Whitney, “Ouch!—MoreThan You Ever Wanted to Know About CatBites,” http://www.thecatsite.com/cat_snips/snips.php?a=bites (accessed June 8, 2004)(reporting that 80 percent of cat bites getinfected and that one out of 170 people willbe bitten by cats each year); NBC11.com,“Cat Bites Can Be Deadly: Woman Hospital-ized After Bite,” http://www.nbc11.com/print/2191468/detail.html? (accessed May 9,2003) (describing the ordeal of a woman whohad a “brush with death” after being bitten byher cat; she required hospitalization for aweek); Sound Medicine, “Dog versus catbites,” http://www.soundmedicine.iu.edu/archive/2002/quiz/animalBites.html (accessedJuly 27, 2002) (50 percent of cat bites are

infectious, compared to 20 percent of dogbites; cat teeth can penetrate more deeplyand transmit bacteria more easily). In spite ofthese data, however, the insurance industryhas not tried to outlaw cats as pets. Why not?

464Beaver, supra note 461, 1148.465Ibid.466Pennsylvania Consolidated Statutes

Annotated 3 § 459-507-A(d) (West 2004).467A.B. 6761, 2003-2004 Sess. (N.Y. 2003).468H.B. 174, 2003 Sess. (N.H. 2003).


The Case Against Dog Breed Discrimination 2 by Homeowners’ Insurance … · 2019-12-16 ·...

Documents

Transcript of The Case Against Dog Breed Discrimination 2 by Homeowners’ Insurance … · 2019-12-16 ·...