The Case Against Dog Breed Discrimination by Homeowners’ 2CHAPTER Insurance Companies

Larry Cunningham

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

n spring 2003 I moved from Vir- Thankfully, the story ended hap- The insurance industry has pre- ginia to Texas to begin work as a pily for my dogs and me. After weeks judged entire breeds of dogs as Itenure-track faculty member at of calling nearly every insurance being “too risky,” instead of taking Texas Tech University School of Law. agent in Lubbock, I was able to a more reasonable dog-by-dog I brought my two dogs with me: Saffy obtain insurance through the Texas approach to risk assessment. (a four-year-old mixed breed whose Farm Bureau, an organization that Major veterinary and breed reg- parents were a fluffy red Chow Chow advocates for farmers and farming istry organizations have strongly 1 and a big black Labrador retriever) issues. Had it not been for the Farm opposed breed discrimination in and Semona (a two-year-old rot- Bureau, I would have found myself insurance. Authors of scientific tweiler). Neither Semona nor Saffy on the horns of a horrible dilemma: studies on dog bites have even has ever bitten anyone. Neither has whether to buy a home or give up argued against the use of their shown any aggressive tendencies. my dogs. Anyone who knows me can data to support breed-based deci- Both are extremely playful and confirm that this dilemma would sion-making by insurers and legis- friendly animals. have been easy to resolve; I would latures. Dog owners across the After I placed a bid on a house in have chosen my furry family mem- country have spoken out about the 2 Lubbock, Texas, I began the search bers over home ownership. Sadly, horrible choice they have been for homeowners’ insurance—a however, many Americans are find- forced to make between obtaining process that I thought would be ing themselves in similar positions insurance and keeping their dogs. straightforward and easy. Much to and are opting to give up their dogs There has existed a historic ten- 3 my surprise, dozens of insurance to animal shelters. sion between risk classification and companies denied my application Breed discrimination by insur- social policy. Classification and outright. The reason? Semona is a ance companies is on the rise in the insurability decisions are usually rottweiler and Saffy is half-Chow. United States. Insurers are refusing “actuarially justified”—that is, the Rottweilers and Chow Chows are on to write homeowners’ policies for insurance company has identified a the “blacklist” of dog breeds. Some people who own breeds that the statistical correlation between a insurance companies believe they, insurance industry considers to be characteristic and increased risk. along with pit bulls, huskies, Dober- dangerous. Their decisions are Actuarial justification is frequently man pinschers, and other specified based solely on the breed of the ani- cited by insurers as a reason to breeds, are more likely to bite mal, not the individual characteris- avoid social regulation. Insurers humans and, in turn, cause liability tics of the particular dog. Dog bites exist to make a profit for their claims to be brought against their are certainly a public health con- shareholders. They do so by mini- owners. Even mixed breeds, like my cern. However, the insurance indus- mizing risk, which, in turn, mini- half-Chow, Saffy, are blacklisted. try’s approach to the problem is mizes claims paid out. This practice is known by many dog based on faulty assumptions and Actuarial justification is only the owners as “breed discrimination.” improper use of dog-bite statistics. first step in determining the social

25 propriety of a proposed underwrit- My argument is quite simple: shows how the insurance industry ing mechanism. Social utility of the decisions regarding the provision, is a highly regulated industry that risky conduct must also be consid- rating, termination, or renewal of a subjects itself to legislative control ered. Statutes across the United homeowners’ insurance policy where, as here, the public is being States are replete with examples of should not be based on ownership harmed by underwriting decisions legislatures overruling actuarially or possession of a particular breed not driven by actuarial justifica- justified practices in favor of com- of dog unless there is evidence of tion. I also offer a number of alter- peting social policies. “Red-lining” dog-specific risk. Insurers would natives to breed discrimination. is a classic example. Actuaries iden- concededly be actuarially justified tified statistical correlations be- in charging higher premiums or tween living in certain neighbor- declining coverage for people who hoods and increased risk for claims own dogs that have unjustly bitten I. Dog Breed against homeowners’ policies. As a in the past. After all, the best pre- Discrimination result, insurance companies began dictor of future behavior is past be- Breed discrimination in insurance to refuse to write policies in these havior. Breed discrimination, as it is a recent phenomenon that was high-risk neighborhoods. The currently stands, is not actuarially preceded by the enactment of neighborhoods in question were justified because scientists have “breed-specific legislation” (BSL) often economically depressed and not been able to accurately deter- by some state legislatures and occupied by members of racial or mine whether certain breeds are municipalities. Both breed dis- ethnic minorities. Legislatures and inherently more dangerous, or, crimination and BSL are a per- courts stepped in to prohibit red- instead, whether a breed’s high ceived response to highly publi- lining, despite the actuarial justifi- population is making it appear cized attacks by certain breeds, cation for the practice.4 that the breed is more dangerous. particularly pit bulls. Breed discrimination is a differ- The consequences of breed dis- ent animal altogether. Even with- crimination could not be greater. out considering the high social Homeowners’ insurance is the Highly Publicized utility of pet ownership, insurers gatekeeper to homeownership. have been unable to demonstrate Without homeowners’ insurance, a Attacks by an actuarial justification for dis- buyer cannot get a mortgage. For Pit Bulls criminating based on breed. As most Americans, if a person can- In the 1980s there were a number the multidisciplinary Task Force not obtain a mortgage, he cannot of high-profile attacks on humans on Canine Aggression and Human- buy a home. by pit bulls. These attacks led to a Canine Interactions concluded, In Part I of this article, I give an near-hysterical reaction by mem- “[D]og bite statistics are not overview of the problem: dog breed bers of the communities that really statistics, and they do not discrimination by insurers, as well were affected by the attacks and give an accurate picture of dogs as a related problem of breed-spe- by the legislators who repre- that bite.”5 The popular notion cific legislation by some states. In sented them. that pit bulls and rottweilers are Part II, I analyze the major scien- In March 1984, pit bulls inherently more likely to bite is tific studies on dog bites, showing attacked Angie Hands, a nine-year- simply not supported by the avail- that no one has adequately proven old girl in Tijeras, New Mexico.7 able statistics. that some breeds are more inher- The dogs bit her right leg to the When the social utility of pets is ently dangerous than others. In bone, ripped flesh from her arms, added to the equation, breed dis- Part III, I show that breed discrim- and tore her ear in half.8 The child crimination becomes even more ination and breed-specific legisla- survived but had to undergo years unreasonable. Dogs and other do- tion are opposed by most veteri- of reconstructive surgery.9 She had mesticated animals provide im- nary and animal protection been attacked by her uncle’s four measurable joy and happiness to groups. Part IV demonstrates that pit bulls in between her bus stop the families that own them. Even insurers have been ignoring the and her home.10 The small commu- some components of the legal sys- unique and special role that pets nity of Tijeras, located outside of tem itself have evolved to recog- play in millions of American Albuquerque, responded with an nize pets as being more than mere homes. I draw upon not only the outright ban on pit bull owner- chattel.6 In addition, the failure to profoundly personal arguments ship.11 Dog owners challenged the obtain homeowners’ insurance is a advanced by myself and others, but law in court, but the law was death knell for homeownership— also the way in which the law itself upheld as a constitutional exercise no insurance, no mortgage; no is evolving by recognizing pets as of the town’s police power.12 mortgage, no house. more than mere property. Part V

26 The State of the Animals IV: 2007 The attack on Angie Hands fol- a breed that is commonly known as tion, and the vagueness doctrine.34 lowed a number of other pit bull a pit bull dog”20 is automatically a Plaintiffs have challenged BSL on attacks around the country. A four- “vicious dog.”21 “Vicious dogs” due process grounds by arguing year-old girl was killed in Oregon must be penned or tied up when on that there was no “rational relation- City, Oregon, when she fell into a their owners’ premises.22 If off- ship” to a legitimate legislative goal yard where a pit bull was chained.13 premises, they must be tethered, or purpose.35 Courts have ruled that Two pit bulls mutilated their owner caged, or muzzled.23 Owners must BSL is a rational response to a per- in Edgemere, Maryland.14 A re- obtain liability insurance to provide ceived problem of dog bites by cer- cent, widely publicized attack in coverage in the event of a bite.24 tain breeds.36 They have also San Francisco has also brought the BSL has also occurred at the local rejected plaintiffs’ arguments that issue of aggressive dogs to the fore- municipal level. Denver passed an the statutes and ordinances do not front of public attention. In Janu- outright ban on the ownership, pos- provide dog owners with sufficient ary 2001, Diana Whipple was session, keeping, control, mainte- notice and an opportunity to be mauled to death by two Presa nance, harboring, transportation, or heard, which are the requirements Canario dogs. The dogs were sale of pit bulls.25 A “pit bull” is for procedural due process.37 The owned by a pair of lawyers. Evi- defined as an American pit bull ter- Tijeras ordinance, for example, pro- dence at the owners’ murder trials rier, American Staffordshire terrier, vides that a pit bull may be de- showed that the dogs had tried to Staffordshire bull terrier, or any dog stroyed by the village only after a attack other people and animals displaying the majority of physical hearing to determine whether the in the past. Both defendants were traits of one of those breeds.26 This dog is, in fact, a pit bull.38 Plaintiffs convicted and served prison ordinance is in addition to Denver’s have also contended that BSL time.15 A subsequent civil lawsuit “dangerous dog” ordinance that reg- amounts to a taking without just brought by Ms. Whipple’s mother ulates “[a]ny dog with a known compensation. Courts have rejected was settled out-of-court.16 propensity or disposition to attack un- this argument, noting that personal provoked, to cause injury or to other- property is subject to regulation wise endanger the safety of humans under the police power of a state.39 “Breed-Specific or other domestic animals.”27 “Dan- Challenges based on vagueness have gerous dogs” must be confined while argued that identifying a dog’s breed Legislation” at home and must be leashed and is difficult.40 Most courts have found Highly publicized pit bull attacks muzzled while traveling.28 BSL to be sufficiently specific to in the 1980s led to knee-jerk reac- Not all states have followed the enable a reasonable dog owner to tions by many communities.17 BSL trend. Some legislatures have determine if his or her dog is covered Attacks led to editorials, which led prohibited BSL enacted by munici- by the particular statute.41 Plaintiffs to public outrage, which led to palities. Florida enacted a statute have also alleged that BSL violates swift and spontaneous legislative that permits localities to regulate equal protection by singling out pit action that was based on neither dogs “provided that no such regu- bulls but not other breeds.42 Courts good science nor good law. BSL lation is specific to breed.”29 Some have noted that pit bull ownership is began to emerge in the 1980s and legislators have attempted, with- not a “suspect classification,” and, early 1990s. These laws targeted out success, to repeal this anti-BSL therefore, BSL need only have some specific breeds for regulation or, in statute in response to several reasonable basis to be constitutional. some cases, outright bans. BSL is highly publicized attacks.30 Min- Courts have concluded that sufficient on the rise in the United States. nesota also has the following pro- evidence exists to support a finding States and municipalities across hibition against BSL: that pit bulls can be regulated by leg- the country have considered—and, A statutory or home rule char- islatures and municipalities.43 in some cases, enacted—breed- ter city, or a county, may not One significant decision found specific legislation designed to adopt an ordinance regulating BSL to be unconstitutional. In protect the public against dog dangerous or potentially dan- American Dog Owners Association, bites.18 Commonly, these statutes gerous dogs based solely on Inc. v. City of Lynn,44 the Massachu- and ordinances have banned, or the specific breed of the dog.31 setts Supreme Judicial Court up- placed restrictions on, pit bulls, Court challenges to BSL have held a trial court’s finding that the rottweilers, Doberman pinschers, been largely unsuccessful.32 Oppo- City of Lynn’s attempt to regulate Chow Chows, German shepherds, nents of BSL have brought lawsuits pit bulls was unconstitutional.45 and shar-peis.19 claiming the legislation is unconsti- The Court noted that it is particu- Ohio has aggressively targeted tutional because it violates due pro- larly problematic to determine a pit bulls for regulation. Ohio law cess (substantive and procedural), dog’s breed. The Court held, declares any dog that “[b]elongs to the Takings Clause,33 equal protec- [T]here is no scientific means,

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 27 by blood, enzyme, or other- lovers have a term for what the in- that there are alternatives to the wise, to determine whether a surance company did. They call it current practice and that it must dog belongs to a particular ‘breed discrimination’—arbitrarily stop.63 To achieve their goal, The breed, regardless of whether punishing all dogs of certain HSUS and the American Society “breed” is used in a formal breeds because some are vicious.”53 for the Prevention of Cruelty to An- sense or not.46 In the months that followed, sev- imals (ASPCA) have created a joint The Court upheld the trial court’s eral newspaper stories discussed grass-roots campaign designed to finding that animal-control officers the prevalence of breed discrimina- educate the insurance industry.64 had no real standards to identify pit tion and documented the effects bulls, in part because they had no this practice has had on families.54 The Insurance training in breed identification.47 These news reports replicate the Industry’s Defense of The ordinance included a ban on experience I had in trying to get Breed Discrimination mixed-breed dogs that contained homeowners’ insurance. Multiple Homeowners’ insurance protects a “any mixture” of pit bull.48 This pro- insurers denied coverage because policyholder in the event of finan- vision was likewise found to be un- of the dogs I owned. I literally cial loss. Most policies include two constitutional since it is scientifi- could not find a carrier in the Lub- provisions, property damage and cally “impossible to ascertain” bock market willing to write a pol- liability. Property damage provi- whether a dog is part pit bull.49 The icy for me until I stumbled upon sions protect the policyholder in ordinance was also unconstitutional the Farm Bureau on the advice of the event of fire, lightning, wind, because it tried to define “pit bull” one insurance broker who sympa- water, or hail damage, theft, and as including any breed where “com- thized with my plight. vandalism. Liability provisions pro- mon understanding and usage” dic- The practice of breed discrimina- tect the policyholder in the event tated that the dog was, in fact, a pit tion produces absurd results. Con- that a claim is made against a bull.50 The combination of these sider the case of Chris and Norm homeowner for negligence. Liabil- facts led the court to conclude that Craanen of San Antonio, Texas.55 ity coverage typically pays for bod- the statute was too vague to pass They own a twelve-year-old dog ily injury, medical payments, and constitutional muster.51 named Bukarus. He is a rottweiler, property damage that are sus- a breed often targeted for discrim- tained because of the negligence of ination by insurance companies. the property owner.65 Absent breed The Reaction Yet, Bukarus does not pose much of discrimination, most homeowners’ a threat: he is deaf, partially blind, insurance policies would cover in- of Insurers and has arthritis.56 Despite his bite- juries due to dog bites on the pre- While some communities and free history, his owners lost their mises between the amounts of states have responded to dog bites homeowners’ insurance.57 $100,000 and $300,000.66 In 1995 with breed-specific legislation de- Some of the most well-known in- the average policyholder paid $418 signed to regulate or outlaw cer- surers are engaging in breed dis- in homeowners’ insurance premi- tain breeds, insurance companies crimination.58 Some insurers have ums.67 By 2004 the average pre- have also reacted to the problem of outright bans on specific breeds,59 mium climbed to $608.68 dog bites in a breed-specific man- while others take a more realistic “Insurance is a business.”69 Insur- ner. Dubbed “breed discrimina- and logical dog-by-dog approach. ers must make profits in order to tion” by dog owners, insurance These decisions are predicated on continue in existence.70 Companies companies have started making insurers’ assessment of relative survive by minimizing risk, which coverage and renewal decisions risk.60 The “usual suspects” for reduces the likelihood of claims. based on one’s ownership of cer- breed discrimination are pit bulls, Some companies have decided that tain breeds of dog. rottweilers, German shepherds, certain breeds of dog are simply Doberman pinschers, Chow Chows, “too much of a risk” to insure.71 An A Rise in Breed 61 Discrimination wolf hybrids, and Presa Canarios. industry representative claims that The Humane Society of the the issue of dog bites “is a major During 2003 and 2004, the media United States (HSUS) has docu- concern for insurers.”72 brought breed discrimination to mented an increase in the number The industry defends its posi- light. The CBS Evening News with of people being denied insurance tion, in part, on a series of studies Dan Rather aired a story in June because they own certain breeds of from the Centers for Disease Con- 2003 that featured a family that dog.62 As a result, The HSUS has trol and Prevention (CDC), which had difficulty obtaining insurance started collecting data through the the industry claims as support for because they owned a dalmatian.52 Internet, in the hopes of eventually the proposition that certain breeds The report stated, “[A]nimal convincing the insurance industry have a propensity to bite.73 As I

28 The State of the Animals IV: 2007 demonstrate in Part II, however, the III defended breed discrimina- the industry’s reliance on the CDC tion on the basis that certain II. The Lack of studies is misplaced. Even the breeds cause more damage when Scientific Evidence authors of the CDC studies have and if they do bite.83 Ultimately, a Numerous scientific studies have stated that breed discrimination spokesperson for the III conceded, attempted to identify the number is wrong and is not supported by “[t]he industry isn’t positioned to of annual dog bites, the dogs most scientific evidence.74 determine which dogs should be likely to bite, the people most The industry has also pointed to deemed vicious....[W]e’re certainly likely to be bitten, and the circum- the large amount of money that not dog experts or veterinari- stances under which bites are most has been paid out in recent years ans.”84 This, however, has not likely to occur. Such studies have for dog-bite claims.75 The Insur- stopped many insurers from engag- not reached a uniform consensus ance Information Institute (III), a ing in breed discrimination. and have left us with more ques- trade group of the insurance indus- tions than answers. Even the stud- try, stated that in 2002 $345.5 mil- Some Exceptions ies that have attempted to report lion was paid out in dog-bite liabil- to the Rule? on breeds’ proclivity to bite have ity claims, up from $250 million in It appears that not all insurers cautioned that their research is 1995.76 The group argues that dog- have followed the breed discrimi- incomplete and should not be used bite lawsuits are on the rise and nation trend. DVM reported that to justify breed discrimination by juries are awarding larger claims.77 Nationwide Insurance changed its legislatures or insurers.90 It claims, therefore, the need to breed discrimination policy in curtail its risk. October 2003. While Nationwide The industry’s cost statistics are now insures all dog owners, it CDC Statistics misleading, however. The III states, specifically excludes dog bites from The CDC commissioned a number “[D]og bites now account for its liability coverage.85 of studies during the 1980s and almost one quarter of all home- State Farm’s national represen- 1990s to determine the scope and owner’s insurance liability claims tatives have repeatedly stated that nature of the problem of dog bites costing $345.5 million.”78 Some the company does not practice in the United States. perspective is in order. For breed discrimination.86 However, every $100 in premiums, insurers when I searched for homeowners’ spend $77 paying claims. Of that insurance in 2003, a State Farm Fatality Studies $77, the overwhelming majority agent in Lubbock refused to even Four separate studies attempted to ($72, or 93.5 percent) is spent take my application because of the chronicle the number of fatal dog on paying property damage breeds I owned. bites during the periods of claims. Liability claims only 1979–1988,91 1989–1994,92 1995– amount to $5, or 6.5 percent, of Other Instances of 1996,93 and 1997–1998.94 The total claims.79 Even then, dog Breed Discrimination studies were specifically limited to bites only constitute a percentage There are other examples where a fatal dog attacks because fatality of that figure. Put into perspec- person’s ownership of a particular statistics are easier to track.95 Non- tive, the money paid out in dog- breed of dog can have negative fatal bites were excluded from the bite claims is negligible when consequences. Families seeking to studies, although other scientists compared to the overall amount of adopt children can face roadblocks have attempted to use emergency money paid out for other types of if they own dogs that belong to cer- department reports and other claims. Damage due to lightning, tain breeds. In Massachusetts the sources to determine the number fire, and mold all individually Adoption and Foster Care Unit of of nonfatal bites per year.96 account for more claims payouts the Department of Social Services The authors combed three sets than all liability claims combined.80 will not place children in homes of sources in an attempt to deter- The insurance industry has not with certain breeds of dog.87 The mine the number of fatal dog bites been consistent in the reasons for state relied upon data provided by per year. First, they searched its defense of breed discrimination. the insurance industry when it NEXIS for news reports of dog bite- One report from the III’s website made its decision to discriminate related fatalities.97 Second, they seems to defend breed-specific based on breed.88 Some airlines used the National Center for responses based on the aggregate also practice breed discrimination Health Statistics’ (NCHS) single- claims paid81 and stories of several by prohibiting some dogs from fly- cause mortality tapes (SCMTs) to high-profile and tragic bites.82 ing, even though they are stored in identify deaths where the underly- However, in a statement to a cargo and in a closed carrier.89 ing cause was listed as a dog bite.98 newsletter of veterinary medicine, Finally, the authors supplemented

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 29 these reports with information col- ing period, 1997—1998, shows lected by The HSUS to help iden- that rottweilers caused ten fatal Nonfatality tify the breed of dog involved in bites per two-year period, while pit Studies each incident.99 From these three bulls caused six, and Saint Ber- The CDC fatality studies acknowl- sources, the authors tried to piece nards caused three.111 During the edged that, while death rates for together the number of people twenty-year study, ninety deaths dog bites do not appear to have who died each year in the United were excluded because the breed increased over time,121 nonfatal States from dog bites. was “unavailable.”112 bites were becoming more of a The authors concluded that dog The authors of the CDC studies public health problem.122 The CDC bites caused approximately seven acknowledged that the methods conducted a study of nonfatal dog deaths per year per hundred mil- they used in their studies had a bites in 2001.123 The study used lion people.100 They discerned no number of limitations. NEXIS, they data from the National Electronic identifiable trend that would indi- pointed out, was not designed for Injury Surveillance System-All In- cate an increase in the incidence scientific research. News reports jury Program (NEISS-AIP) to iden- of fatal bites over the years of the would only be flagged if their text tify the number of nonfatal dog studies.101 During the first report- contained certain keywords.113 bites during the 2001 calendar ing period (1979–1988), approxi- Further, reliance on NEXIS year. NEISS-AIP collects data from mately 70 percent of victims were assumes that newspapers accu- initial visits to emergency depart- under the age of ten.102 Males, rately reported the breed of dog ments (EDs) across the country.124 under the age of twenty-nine, were involved in a particular attack.114 NEISS-AIP data are drawn from a more likely than females to be vic- SCMTs have a one- to two-year lag nationally representative sample of tims.103 These findings as to age time, which means that some fatal- NEISS hospitals.125 The CDC ana- and gender were consistent ities may have been missed.115 The lyzed every case where “dog bite” throughout the study periods. authors believed that, on average, was listed as the external cause of Many of the fatal bites of children their methods only uncovered injury.126 involved horrific attacks on the very approximately 74 percent of dog- In total, NEISS-AIP data revealed young. A three-week-old girl was bite-related fatalities.116 that hospital EDs treated 6,106 killed in her crib by the family’s Even if one accepts the CDC sta- patients for dog-bite-related in- Chow Chow.104 A two-year-old boy in tistics as definitive on the subject, juries during 2001.127 Since the South Dakota wandered into a they have a number of other limita- NEISS-AIP data did not include neighbor’s yard, where he was tions in answering the question every hospital in the nation, the attacked and killed by two German of whether certain breeds are more authors used these data to extrap- shepherd-wolf hybrids.105 The elderly dangerous than others. First, the olate to the general population.128 were also victims of several fatal studies were limited to fatal dog They estimated that 368,245 peo- attacks. In March 1996 two rottweil- attacks.117 Second, the breed of ple were treated for dog-bite- ers killed an eighty-six-year-old Ten- the dog could not be accurately de- related injuries in 2001.129 The nessee woman. One month before termined in every case.118 Finally, largest cohort of victims was chil- the assault, the dogs had attacked the number of fatal attacks per dren between the ages of five and and injured the same woman.106 year is so low that it is problematic nine.130 Boys, under the age of In the twenty-year period of the to statistically extrapolate conclu- fourteen, were more likely than CDC studies, the breed responsible sions from the data. For example, girls to be seen in EDs for dog-bite- for the most number of bites has in the first two years of the study related injuries.131 changed.107 From 1979 to 1980, (1979–1980), Great Danes ac- The NEISS-AIP data included Great Danes caused the most num- counted for the most number of narratives for many of the attacks. ber of fatalities, with three deaths fatal bites (three).119 Four breeds, One case involved a four-year-old for the period. However, four however, followed closely behind who was bitten by a dog guarding breeds were tied with two deaths with two fatal bites each (pit bull, her puppies.132 Another involved a each: pit bulls, rottweilers, huskies, German shepherd, husky, and mal- three-year-old girl who was bitten and malamutes.108 In 1981 pit bulls amute).120 It would be statistically when she tried to take away a took over as the breed with the questionable to conclude that dog’s food.133 A thirty-four-year-old most number of fatal bites.109 Pit Great Danes were inherently more man was bitten while trying to bulls remained in that position dangerous than the other breeds, break up a dogfight. Some victims until 1993, when rottweilers began based on a net difference of only were bitten by their own dogs. A causing approximately ten fatal one fatality. twenty-seven-year-old woman was bites per two-year reporting bitten by her dog after he had period.110 The last available report- been hit by a car and became dis-

30 The State of the Animals IV: 2007 oriented.134 A seventy-five-year-old The study created a control (4,494,083 people) had been bitten woman was attacked while trying group of dogs to try to determine in the previous twelve months, and to prevent her dog from biting an whether certain characteristics 0.3 percent had sought medical emergency medical technician (such as breed) made a dog more attention.155 This shows that nonfa- (EMT) who was attempting to put likely to bite.144 Using a multivari- tal bites are a public health problem the woman in an ambulance.135 ate statistical analysis, the study that “is five orders of magnitude The Morbidity and Mortality concluded that biting dogs were greater” than fatal dog bites.156 The report describing the study does more likely than control dogs to study concluded that several factors not document the number of be German shepherds or Chow had no statistical significance on attacks per breed. This is likely due Chows, male, intact (not neu- the likelihood of being bitten: cen- to the fact that the ED reports did tered), and reside in a house with sus region, urbanicity, race/ethnic not specify the breed of dog. An one or more children.145 Denver group, and household income.157 attempt to determine the number had (and still has) a ban on pit The study did not attempt to corre- of bites per breed would depend on bulls, so it is not surprising that no late between the number of bites victims accurately self-reporting cases involved that breed.146 and the breed of dog. The authors the breed of the attacking dog.136 The authors acknowledged that acknowledged that the study relied The study had a number of limi- their results had several problems. on the self-reporting of data, which tations. First, the authors excluded First, they were only able to speak were not validated, and that they fatal dog bites. Second, the study to owners of approximately half of received a poor response rate (only only examined cases where the vic- the biting dogs. They excluded 56 percent of people responded to tim sought treatment in an ED. cases in which the victim did not the survey).158 Victims may have gone to other seek medical attention. In this health care providers, such as pri- respect, the authors believed that vate physicians or urgent-care cen- seeking medical attention was a Other Studies ters. Third, 26 percent of reports “surrogate” for “real bites.”147 The Other studies have attempted to doc- were missing an injury diagnosis. authors did not verify the breeds of ument the total number of dog bites Many cases had limited data on the the dogs involved, but, instead, and the number of bites per breed. circumstances of the attack or the “identified predominant breed as A study of ED visits for dog-bite identity of the dog involved.137 whatever breed the owner consid- injuries159 confirmed many of the Thus, the CDC’s estimates may be ered the dog.”148 Because of the conclusions of the previously dis- both overinclusive (“just cause” small number of bites per breed, cussed CDC study of ED visits.160 bites may have been included)138 the authors could not assess the The study noted that a lack of a na- and underinclusive (insofar as vic- statistical significance of breeds tional reporting system for dog- tims may have sought treatment at other than German shepherd and bite injuries makes gathering and other facilities). Chow Chow.149 analyzing data on the subject diffi- Another CDC study attempted to Another CDC study attempted to cult.161 The authors, in reviewing identify the incidence of dog bites determine the frequency of dog the literature on the subject, found in a particular locality: Denver, bites by conducting a random tele- that previous studies concluded Colorado.139 The authors examined phone survey of households.150 The that between 0.3 percent and 1.1 reports from the Denver Municipal authors used the Injury Control and percent of all ED visits are due to Animal Shelter in 1991.140 There Risk Survey (ICARIS), a random- dog-bite-related injuries.162 To were a total of 991 bites during the digit-dialing telephone survey.151 determine the true percentage, study period.141 However, only 178 They asked each adult respondent they collected data from the Na- were eligible for the study,142 as whether he (or his children) had tional Hospital Ambulatory Med- the authors excluded several cate- been bitten by a dog in the previous ical Care Survey (NHAMCS), a ran- gories of bites: bites involving twelve months and whether the vic- dom surveying instrument that is household members, attacks invol- tim had sought medical atten- used to calculate the number of ving multiple dogs, attacks before tion.152 Out of 5,328 completed in- ED visits per year.163 They esti- 1991, dogs who had been owned terviews, ninety-four adults and mated that between 1992 and for less than six months, cases in ninety-two children reported being 1994, 333,687 annual visits were which the owner did not live in bitten in the previous twelve made to EDs seeking medical Denver County, attacks where the months.153 Of these, twelve adults treatment for dog-bite-related owner’s phone number was not and twenty-six children sought med- injuries.164 This amounted to 0.4 listed on the report, and cases in ical care.154 From these data, the percent of all ED visits nation- which the victim did not receive authors extrapolated that 1.8 per- wide.165 Looking at the monetary medical treatment.143 cent of the American population cost of dog bites, they found that

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 31 the average cost for a dog-bite- dence rate must be higher than “systematic” in its approach.188 related ED visit was $274, result- that found in the study.177 Indeed, Pinckney and Kennedy con- ing in an annual cost of $102.4 Another survey178 in Pennsylva- ceded that their database was million.166 The study, however, did nia polled children in order to “incomplete” and “may not be not address the question of determine an overall bite rate from entirely reliable.”189 Their data whether certain breeds are particu- the perspective of bite victims.179 depended on newspaper reports, larly more dangerous than others. The survey, conducted in 1981, which may themselves be incom- This is partly due to the unavail- found that 46.1 percent of children plete or inaccurate. Thus, the ability of data through NHAMCS. reported that they had been bitten authors said their data required Moreover, the study most likely by a dog during their lifetime.180 “cautious interpretation.”190 An undercounted the number of non- The study concluded that example of such “cautious interpre- fatal dog bites because victims may [B]eing bitten by a dog is a tation” is represented by the have sought treatment from places rather common occurrence for authors’ observation that even other than EDs.167 children, especially those be- though German shepherds were Other studies have attempted to tween the ages of seven and involved in more fatalities than any examine the problem at a more twelve years, and the event is other breed in the study, such large localized level. A July 1991 greatly underestimated by offi- frequency could be reflective of the study168 found that dog bites were cial bite statistics.181 fact that German shepherds had responsible for 0.3 percent of all Nevertheless, the authors did not the highest AKC registration of any ED visits at The Children’s Hospi- attempt to catalog bites per breed.182 large breed.191 Hence, the use of tal of Philadelphia.169 Of those vis- Unfortunately, not all scientists AKC data to draw comparisons its, 77 percent involved cases have used statistically sound meth- between breeds is problematic,192 where the victim knew the biting ods to draw conclusions about the as demonstrated by the high num- dog.170 The study found one statis- relative dangerousness of breeds. ber of registrations for breeds such tically significant conclusion: more Two physicians, Lee E. Pinckney as German shepherds, and low pit bull injuries were the result of and Leslie A. Kennedy, from the number of registrations for a popu- unprovoked attacks as compared Department of Radiology at the lar breed, such as the pit bull.193 to such attacks by other breeds.171 University of Texas Southwestern William Winkler’s study194 in “Unfortunately, the absence of reli- Medical School and Children’s 1977 has also been criticized for able dog breed-specific population Medical Center, sent letters to the its lack of scientific method.195 His figures prevent[ed] the calculation editors of 245 major newspapers “study” involved compiling news of breed-specific injury rates.”172 requesting copies of all stories reports from eleven dog-bite-re- An October 1997 study tried to about dog-bite-related fatalities.183 lated fatalities from January 1974 determine the number of dog bites The number of fatalities reported through December 1975.196 From in Alleghany County, Pennsylvania by the responding newspapers be- these data, he made various con- (Pittsburgh), by using the “cap- tween March 1966 and June 1980 clusions about the breeds responsi- ture-recapture” method of statisti- totaled seventy-four.184 Of the sev- ble, finding that, “not unexpect- cal analysis.173 The authors found enty-four fatalities, sixteen were edly,” German shepherds were the that 790 dog bites were reported caused by German shepherds, nine breed most often responsible for to the Alleghany County Health by huskies, eight by Saint fatal dog attacks.197 Because Saint Department in 1993.174 Using the Bernards, six each by bull terriers Bernards were responsible for two capture-recapture method, along and Great Danes, and five by mala- deaths during this twenty-four- with log-linear modeling, the study mutes.185 The remaining dog-bite month period, he concluded, concluded that the number of un- fatalities were caused by a variety “[t]his relatively uncommon breed reported dog bites was 1,388 (with of breeds, including ten attacks by may be a greater hazard than a 95 percent confidence interval of mixed breeds and five attacks by others.”198 between 1,010 and 1,925).175 The dogs of unknown breeds.186 In A common thread running authors cautioned, however, that addition to acquiring bite fatality through several studies is the the self-reporting sources are prob- statistics from newspapers, the attempt to extrapolate conclusions lematic in that “whether or not a authors used American Kennel about breeds based on limited case is reported depends largely on Club (AKC) registration data to data. For example, an April 2000 the severity of the event and the compare the relative number of epidemiological study in Philadel- attitude, knowledge, or education fatalities per breed.187 phia used reports from the Depart- level of the victim.”176 Accordingly, The CDC authors criticized the ment of Health to conclude that the authors suggested that the Pinckney/Kennedy study as being between 1995 and 1997 there were actual Pittsburgh dog-bite inci- “primarily anecdotal” rather than approximately 5,390 bites.199 The

32 The State of the Animals IV: 2007 authors concluded that pit bulls, to bite without any evidence of in- population sizes, Breed Y would German shepherds, and rottweilers dividual dangerousness. be perceived to be the more combined were responsible for 59 dangerous breed on the basis of percent of bites each year.200 The the number of fatalities.207 authors felt comfortable drawing Numerators and Using the RDR normalizes the this conclusion despite the fact Denominators effect of a breed’s popularity, or that they could not determine the lack thereof. Dogs of popular breed in 74 percent of cases.201 in Dog-Bite breeds are going to bite more Statistics often simply because there are To date, no scientific study has more of them.208 A January 1997 The Unknown been able to resolve what I term to article warned that, as dalmatians be the problem of “numerators become more popular, people Origin of and denominators.” A person wish- should expect to see more bites Aggressiveness ing to determine whether certain from that breed.209 This is not to Despite all of the research and stud- breeds are more likely to bite than say that dalmatians are inherently ies on the subject, scientists and others must first determine the more dangerous than other breeds. veterinarians cannot state with cer- number of bites per breed (the Rather, an increase in their popula- tainty or confidence why certain numerator) and then compare that tion should also result in a propor- dogs are more aggressive than oth- number to the total number of tional increase in bites from that ers.202 It seems that a particular dogs of that breed in the general breed.210 Similarly, the Pinckney/ dog may be aggressive because of a population (the denominator). Kennedy study211 cautioned that, variety of factors.203 According to This can be expressed as a ratio: despite the fact that German shep- the American Veterinary Medical herds accounted for the most num- Number of Bites Association’s multidisciplinary Task Relative by Breed ber of deaths, their finding must Force on Canine Aggression and Dangerousness = be read in conjunction with the Ratio Total Population Human-Canine Interactions, “A of Breed popularity of the breed, as evi- dog’s tendency to bite depends on denced by AKC registrations of the at least five interacting factors: This ratio (RDR) allows for a same time period.212 heredity, early experience, later so- comparison between breeds. The The problem of numerators and cialization and training, health higher the RDR, the greater pro- denominators is that it is diffi- (medical and behavioral), and vic- clivity a particular breed has to cult—if not impossible—to accu- tim behavior.”204 bite. It allows for a comparison of rately determine the number of While breed (as an inherited “oranges to oranges” and “apples bites per breed and the number of characteristic) is one component to apples.” Otherwise, it is likely dogs in a particular breed. Without of predicting a dog’s dangerous- that highly popular breeds will an accurate count for either the ness, it is not the only factor.205 appear to be more dangerous, numerator or denominator, one There is no way to scientifically when in fact the number of bites is runs the risk of stigmatizing an determine whether a dog is likely reflective of the overall population entire breed as “overly dangerous” to bite in the future, any more of the particular breed. based on the breed’s absolute num- than psychologists can predict A study that tried to extrapolate ber of bites, instead of examining whether certain people will com- breed data from the previously dis- the breed’s number of bites rela- mit crimes of violence. The excep- cussed CDC studies agreed that tive to its overall population. tion to this rule is the axiom that the proper method for determin- the best predictor of future behav- ing a breed’s dangerousness was The Numerator Problem ior is past behavior. For this rea- the use of a comparative ratio: The principal problem in determin- son, many veterinary and scientific Ideally, breed-specific bite rates ing the total number of bites by groups support “dangerous dog would be calculated to compare a particular breed is that there is no laws” that target individual dogs breed and quantify the relative national reporting system for dog who have demonstrated a propen- dangerousness of each breed. bites.213 The CDC studies214 demon- sity to bite or attack innocent For example, 10 fatal attacks by strate that, while fatal dog bites are victims.206 The problem with BSL Breed X relative to a population easier to track than nonfatal bites, and breed discrimination is that of 10,000 X’s (1/1,000) imply a even the methodology used to legislatures and insurers have greater risk than 100 attacks by uncover fatalities misses approxi- attempted to prophylactically deter- Breed Y relative to a population mately 26 percent of cases.215 mine which breeds are most likely of 1,000,000 Y’s (0.1/1,000). Further, news accounts—on which Without consideration of the the CDC relied, in part, to deter-

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 33 mine the number of fatal dog bites lem then becomes how to deter- pit bulls. “Husky” refers to a class and the breeds involved—may be mine how many of those dogs be- of dogs, not any one particular biased toward reporting attacks by long to each breed. Determining breed. Siberian huskies, Alaskan certain breeds.216 the breed of one dog is difficult malamutes, and Samoyeds are all The numerator may also be enough.226 To take a census of all considered to belong to the biased against dogs who cause dogs and identify their breeds “husky” family, yet they are all dif- more damage, while ignoring would be an impossible task. ferent breeds.240 Similarly, there is breeds that bite more often but do Some scientists have suggested no AKC-standard breed called “pit not cause victims to seek emer- using AKC or municipal registra- bull.” “Pit bull” is a collective clas- gency treatment.217 If a dog bite tion data to determine the number sification of the American Stafford- does not cause serious injury, it is of dogs in a particular breed in a shire terrier, Staffordshire pit bull not likely that the victim would particular community.227 However, terrier, and bull terrier.241 seek medical treatment.218 This one study concluded that city reg- Scientists have not been able to then skews the results of studies istrations account for only 29.1 determine if victims of dog bites that use emergency department percent of all dogs.228 Further, can accurately report the breeds of visits to track the incidence of dog owners of breeds considered “dan- dogs that attacked them. Many sci- bites.219 “The problem with self- gerous” may be reluctant to regis- entists, particularly the CDC reporting sources is that whether ter their animals.229 This may be authors, have stated that misiden- or not a case is reported depends particularly true of dogs used for tification is a likely problem, espe- largely on the severity of the event illicit purposes, such as those cially under the stress of a dog and the attitude, knowledge, or ed- owned by drug dealers, dogfight- attack.242 Part of the problem may ucation level of the victim.”220 ers, and gang members.230 be that as a particular breed gets Studies that have used random AKC registration data is also a reputation for dangerousness, sampling221 are equally problem- problematic because the AKC only some victims jump to the conclu- atic because they, too, depend on registers purebred dogs231 and sion that they were bitten by a dog accurate self-reporting of their depends on owners taking the ini- of that breed.243 sample groups. The low response tiative to register their dogs.232 Even under ordinary, low-stress rates of these studies also lead to Mixed breeds, for which there are conditions, many people have diffi- questions about the accuracy of numerous combinations, are not culty identifying a dog’s breed. the results that are extrapolated to eligible for registration.233 Pit bulls For the average person any- the general population.222 are often registered with organiza- thing with prick ears and blue tions other than the AKC. If owners eyes automatically becomes a The Denominator do register them, they register “husky”....Any smooth coated Problem with the United Kennel Club or the brown dog, medium sized, and No one knows how many dogs are American Dog Breeders Associa- muscular becomes a “pit present in the United States at any tion.234 If a breed is undercounted bull”....Any tall dog becomes a one time. This should not be sur- in the denominator of the ratio, it Great Dane, fuzzy or hairy, and prising, as even the constitution- will make a breed appear more it’s a Chow Chow. If it’s black ally mandated223 decennial census dangerous than it actually is.235 and tan and heavy, it’s a rot- of human beings is known to un- tweiler, etc.244 dercount people.224 One survey of bite reports found Determining the true number, or The Problem that medium-size black and tan ani- even an accurate estimate, of dogs of Breeds mals were likely to be recorded as can be problematic. While many Breed is a human construct that is German shepherds. Stocky, short- dogs are kept as household pets,225 used to conveniently group dogs haired dogs were listed as pit bulls. others are used as service animals based on similar physical character- Media reports of pit bull attacks are or guard dogs; kept in animal shel- istics.236 There is no scientific test often accompanied by pictures of ters or animal stores; or simply to determine a dog’s breed.237 The boxers or pugs instead of American allowed to wander the streets as only way to determine a dog’s breed Staffordshire terriers.245 One enter- strays. The dog population is con- is to examine its heredity. This task taining website, called “Find the stantly changing and moving, is made possible but is expensive pit bull,” displays twenty-one pic- which makes obtaining an accurate and time-consuming,238 if a dog is tures of purebred dogs and chal- count difficult and expensive. registered with the AKC.239 lenges the user to identify the pit Even if it was possible to deter- As examples of the problem of bull among them.246 mine how many dogs exist in the defining and identifying breed, Even veterinarians and other country at any one time, the prob- consider the case of huskies and experts have difficulty determining

34 The State of the Animals IV: 2007 whether a particular dog belongs to rottweiler to be praised for defend- breed, and that it could be scientif- a particular breed.247 This was a cen- ing its owner. This is not the type of ically determined that certain tral concern of the Massachusetts bite that we should be trying to pre- breeds are inherently more danger- Supreme Judicial Court in Ameri- vent. It is also not the type of bite ous than others. What about the can Dog Owners Association v. City that is likely to lead to an insurance owners? Does this not excuse them of Lynn.248 The Court declared the claim. Similarly, if a dog is being from the responsibility to properly city of Lynn’s pit bull ordinance to physically tormented by a neighbor- train and care for their pets? be unconstitutional in part because hood child who is poking it in the The reality is that there is a wide the animal-control officers desig- eye, we would not deny that the dog spectrum of responsible pet owner- nated to enforce the ordinance used has an inherent right to defend ship. For some people, occasionally conflicting and subjective standards itself by growling, snarling, barking, providing food and water for a dog to determine and identify breed.249 or biting back.254 These are “just is considered sufficient. On the The problem of mixed breed cause” bites, bites in which the dog opposite end of the spectrum, complicates the issue even further. has a legitimate reason to defend some people spend thousands of In determining a relative danger- itself or its owners. dollars on luxuries such as pet ousness ratio, it is unclear how to It is possible that the statistics spas, advanced dog agility classes, count mixed breeds.250 Should they are being skewed because property and elaborate beds. Somewhere in be counted once per breed? Not at owners who wish to purchase the middle of the spectrum are all? Create a new category for each “guard dogs” may be self-selecting people who actively ensure that possible combination of breeds? certain breeds based on the popular their pets have food, water, and Aside from how to use the raw data notions of relative dangerousness. shelter; get exercise; are well on attacks by mixed breed, there is Guard dogs are trained to protect trained; and receive adequate vet- the additional problem of misiden- property by scaring away would-be erinary care.256 tification by laypeople.251 Victims intruders and, if necessary, to bite Unfortunately, a small percent- sometimes inadvertently report an actual trespasser. Owners who age of pet owners breed and use mixed-breed dogs as purebreds252 desire to have guard dogs may their pets for illicit purposes. They due to the heat of the moment and rationalize the purchase of one intentionally seek out vicious dogs their lack of training in identifying breed over another based on the who will attack and maim humans subtle breed characteristics. degree to which they subjectively and other animals.257 Dogfighting There is good reason to believe believe that the dog will be “mean” enthusiasts, gang members, and that the raw data being used to cal- or “scary.” This creates a self-fulfill- drug dealers will purposely select, culate relative dangerousness ing prophecy. The “scarier” a breed breed, and train dogs to be vicious. ratios are incomplete and inaccu- is considered by a community, the The purpose may be to intimidate rate. If the data being input into more likely a dog of that breed will rivals (in the case of gangs and the calculation are flawed, the re- be purchased for protection, used drug dealers), to defend illegal sults (claiming to show some for protection, and actually bite an drugs (in the case of drug dealers), breeds are more dangerous than intruder. This will skew the statis- or to make money (in the case of others) are equally flawed.253 tics in a way that purports to show promoters of dogfights).258 For that the particular breed is, in fact, some, having a vicious dog is sim- inherently more dangerous. ply a status symbol.259 In order to The Problem of Despite these concerns, it make dogs into vicious weapons, appears that the studies to date they use “revolting and painful “Just Cause” Bites have not excluded this category of techniques to bring the animals to Even if an accurate count could be bites from their datasets.255 This is the verge of bloodlust.”260 Drug obtained of the number of bites per a fatal flaw in the statistics, for it dealers in Philadelphia during the breed, there is the additional prob- confuses the issue between inher- 1980s had pit bulls named “Mur- lem of how to handle “just cause” ent dangerousness (due to breed) der, Hitler, and Scarface.”261 They bites in the resulting statistics. If and legitimate animal behavior. wore collars that concealed crack, the purpose is to determine which cocaine, and money.262 In Chicago, breeds are inherently more danger- gang members “brandish[ed] their ous, just by virtue of the breeds fierce pit bulls just as they would a themselves, then the statistics Breed Switching switchblade or a gun.”263 should exclude bites by the dog that by Bad Owners Current statistics do not take in- were justified. If a rottweiler bites Assume for the moment that an to consideration the degree to an intruder who is attacking the accurate relative dangerousness which the source of a dog’s aggres- homeowner, we would expect the ratio could be determined for each siveness is the torturous upbringing

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 35 described above, as opposed to the associated with the CDC and non- governments focus on individual dog’s breed.264 In those situations, CDC studies are inherent to the dogs and dog owners.277 the problem is clearly with the dog problem of trying to determine the The very scientists who have owner—not the dog itself or its number of bites per breed and the authored studies trying to deter- breed. These problem owners are number of dogs per breed. mine a link between breed and dangerous with any breed of dog.265 aggressiveness oppose breed dis- One solution would be for insur- crimination and BSL. In many ers to write policies that exclude of the CDC studies, the scientists injuries related to dogfighting. This III. The Widespread cautioned against using their in- would limit the claims paid out for Opposition to Breed complete data on attacks to make these high-risk animals, yet it would knee-jerk legislative or policy deci- leave potential plaintiffs without an Discrimination sions based solely on breed.278 adequate source of compensation. Breed discrimination by insurance They pointed to the lack of reliable This result might be a socially companies and breed-specific leg- data on bites per breed (the acceptable solution because of the islation by state and local gov- “numerator problem”) and the unclean hands of the “victims.” If ernments have attracted natio- absence of a reliable count of dogs dogfighting exclusions are incorpo- nal attention and outrage by vet- per breed (the “denominator prob- rated into standard homeowners’ erinarians, animal groups, and lem”).279 insurance contracts, the language dog owners. Animal groups have also opposed should be narrowly written to The American Veterinary Med- BSL and breed discrimination. The exclude only those bad faith actors ical Association’s Task Force on AKC has taken a strong stance who, as a matter of social policy, Canine Aggression concluded that against breed discrimination by in- should not be rewarded or compen- BSL and other breed-specific surance companies: sated for injuries attendant to an actions are “inappropriate and The American Kennel Club illegal activity. The key would be to ineffective.”270 The Task Force con- believes that insurance compa- write language that would still pro- sisted of a diverse coalition of vet- nies should determine cover- tect innocent passersby. erinarians, academics, physicians, age of a dog-owning household One of the arguments against insurers, representatives from ani- based on the dog’s deeds, not BSL is that once a breed becomes mal rights advocates, CDC scien- the dog’s breeds. If a dog is a banned, problem owners will sim- tists, and lawyers.271 The Task well-behaved member of the ply switch to another breed.266 In Force agreed that to properly de- household and the community, the 1930s, pit bulls were far from termine the relative dangerousness there is no reason to deny or considered a “vicious breed.” In of breeds, one must first determine cancel coverage. In fact, insur- fact, a pit bull named “Pete” the number of bites per breed and ance companies should con- starred in the Our Gang films of the total population of each breed. sider a dog an asset, a natural the time.267 Fifty years ago the As noted above,272 the accurate alarm system whose bark may Doberman was considered the calculation of both numbers is an deter intruders and prevent most vicious dog.268 During the immense challenge.273 potential theft.280 1980s the focus turned to pit The Task Force rejected the no- The AKC also issued this state- bulls.269 In short, today’s public tion that a dog’s breed is the sole ment concerning BSL: target may be tomorrow’s favorite determinant of dangerousness. The American Kennel Club pet, and vice versa. “[A] dog’s tendency to bite de- (AKC) strongly supports dan- pends on at least five interacting gerous-dog control. Dog-con- factors: heredity, early experience, trol legislation must be reason- Do the Insurance later socialization and training, able, non-discriminatory, and health (medical and behavioral), enforceable as detailed in the Companies Have and victim behavior.”274 They AKC Position Statement. Better Data? also pointed to the problems of To provide communities It is quite possible that one or mixed breeds, misidentification of with the most effective danger- more insurance companies have breeds, and shifting popularity of ous-dog control possible, laws their own proprietary data purport- breeds.275 The Task Force also ex- must not be breed specific. ing to show that one breed or pressed concern about making de- Instead of holding all dog own- another is disproportionately re- cisions based solely on breed, since ers accountable for their be- sponsible for bites. I am skeptical there is a lack of scientific means havior, breed specific laws that their data would be any better to identify breed.276 The Task Force place restrictions only on the than the CDC’s. The problems recommended, instead, that local owners of certain breeds of

36 The State of the Animals IV: 2007 dogs. If specific breeds are pression, reducing high blood pres- banned, owners of these The Growing sure, and addressing obesity.300 On breeds intent on using their Popularity of Pets the other hand, these effects must dogs for malicious purposes, Population be balanced against the negative such as dog fighting or crimi- health effects of dogs, such as bites A study estimated that in 1998 nal activities, will simply and the transmission of zoonotic there would be 53.6 million dogs change to another breed of diseases.301 When the positives are in the United States, a 2.1 percent dog and continue to jeopard- weighed against the negatives, at increase since 1991.288 Approxi- ize public safety.281 least one physician has concluded mately 34.3 percent of homes have In response to a perceived rise in that dogs probably are beneficial to one or more dogs.289 Dog owners breed discrimination, The HSUS human health.302 Some owners will are thus a significant portion of and the ASPCA developed a grass- forgo their own health in order to the United States population. roots campaign to educate the in- care for their pets—a demonstra- They are also a significant pool of surance industry.282 Both groups tion of how much pets mean to customers (actual and potential) oppose breed discrimination.283 some owners. “Most physicians are for insurers. Other groups that have spoken out familiar with at least one example against breed discrimination in- Spending of a person refusing hospitaliza- clude the American Veterinary tion...because there was no one To understand the scope and Medical Association, the American else in the home to care for their power of the pet-owning popula- Dog Owners Association, the West- pet.”303 tion, consider the amount of minster Kennel Club, and the The loss of a pet can have pro- money that is spent on pets each American Humane Association.284 found effects on an owner. A num- year. In 1998, Americans spent ber of organizations provide $11.1 billion on veterinary care bereavement support for people alone, a 61 percent increase from whose pets have died,304 and at 1991.290 There are more than 35 IV. The Unique least three greeting card compa- “pet vacation resorts” where dogs nies make sympathy cards specifi- and Special Role and cats can go to be pampered.291 cally for the loss of a pet.305 There are also more than 650 of Pets in Society Breed discrimination forces pet pet cemeteries in the United For at least twelve thousand years, owners to choose between their States, indicating the extent to the history of the domestic dog, homes and their dogs. Forcing which owners will go to memorial- Canis familiaris, has been inter- owners to make this choice repre- ize their pets.292 twined with that of human sents a significant misunderstand- beings.285 The law has generally Dogs: Members of ing of the role of pets in our soci- treated dogs as mere prop- ety. For some pet owners, giving up 286 the American Family erty —or worse, as nonprop- a pet is like losing a child, sibling, 287 Breed discrimination ignores the erty. As the popularity of dogs or spouse. as pets has grown, the law has reality that most pet owners con- responded in kind by recognizing sider their pets to be members of the importance of dogs, cats, and their immediate family.293 Indeed, The Consequences other pets. The insurance indus- this “coexistence has contributed try, by practicing breed discrimi- substantially to humans’ quality of of Breed nation, has failed to appreciate life.”294 Dogs were initially domes- ticated to be work animals, assist- Discrimination the unique and special role of When a dog bites, it can have lasting ing humans with farming, herding dogs to their owners and to soci- consequences for both the dog and livestock, and providing security at ety. This section is offered to pro- its owner’s family. When an insur- night.295 In time, dogs became vide some context for the implica- ance company refuses to insure or “four-legged members of the fam- tions of breed discrimination. renew a household based on a par- ily.”296 Some dogs provide assis- This is a problem that has the ticular breed of dog, it, too, can have tance to humans with disabili- potential for affecting a large seg- far-reaching consequences. ties.297 Service dogs serve as a ment of the population and for Most people do not respond tangible resource for people, not having damaging effects on the appropriately if their dog bites just a source of companionship.298 mental, physical, and emotional someone. Most punishment is too Dogs can have positive effects on health of people. severe and too late to be of any the health of their owners,299 such value to the dog in preventing as alleviating loneliness and de- future occurrences.306 The dog is

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 37 usually isolated from the family ticated animals (such as horses, and visitors. By limiting inter- Pets: More than cattle, and sheep) on the other action with humans, the dog Mere Property end, and dogs somewhere in be- does not learn how to deal with The problem of breed discrimina- tween.328 To the Sentell Court, people appropriately.307 Isolation tion should be viewed in light of dogs hold “their lives at the will of may also lead to inadequate med- modern developments in animal the legislature, and properly fall- ical care, which may in turn lead law, which is beginning to recognize ing within the police powers of the to serious health problems for that animals are more than mere several states.”329 The Court con- the dog.308 property. Until recently, the legal cluded, “It is purely within the Some owners abandon their status of animals was governed by discretion of the legislature to say dogs or euthanize them either out an 1897 Supreme Court case, how far dogs shall be recognized of frustration at not being able to Sentell v. New Orleans & C.R. Co.317 as property, and under what re- correct aggressive behavior or be- The case involved a Newfoundland strictions they shall be permitted cause an insurance company tells named Countess Lona who was to roam the streets.”330 them to do so in order to get killed by a railroad car.318 Her owner The question of the legal status homeowners’ insurance.309 When brought suit against the railroad for of dogs and other pets has recently BSL goes into effect or insurance negligence. The railroad defended been addressed by courts in the companies discriminate, it causes by relying on a statute that prohib- context of family disputes. Bennett some owners to purposely assume ited an owner from recovering for v. Bennett331 and Arrington v. a sheltered and low profile in the more than the declared value on the Arrington332 typify the majority community to avoid being caught animal’s registration form.319 An rule with respect to the “custody” with an unauthorized pet.310 Shel- owner whose dog was not registered of pets upon their owners’ divorce. ter drop-offs are common after could not recover anything for the In both cases, divorcing couples BSL goes into effect or insurers loss of or damage to the animal.320 sought both custody and visitation begin to discriminate based on Countess Lona’s owner brought of their dogs. In Bennett the trial breed.311 The humane society in suit, challenging the constitutional- court awarded legal custody of the Atchison, Kansas, reported a 40 ity of the law.321 dog, Roddy, to the husband, with percent increase in drop-offs of rot- The Supreme Court held that the the wife receiving every-other- tweilers because of breed discrimi- statute was constitutional as a valid weekend and holiday visitation nation.312 This is unfortunate be- exercise of the state’s police rights.333 Subsequent squabbling cause many shelters can only keep power.322 The Court declared that between the parties led the Court dogs a certain number of days dogs are a form of quasi-property to modify its order to have the par- before euthanizing them. Breed that is “imperfect or qualified” in ties swap custody of the dog every discrimination can have a chilling nature.323 The Court relied on the month.334 The appellate court effect on ownership of certain common law rule that dogs could reversed the trial court’s order and breeds,313 which means certain not constitute stolen property for affirmed the Sentell doctrine: breeds are not likely to be adopted purposes of larceny statutes.324 The “While a dog may be considered by and will have to be euthanized. common law held that wild animals many to be a member of the family, Breed discrimination will likely had no property value until killed under Florida law, animals are con- have an effect on homeownership or subdued.325 Domesticated ani- sidered to be personal property.”335 in states that permit this practice. mals, such as horses, cattle, sheep, The court found that the trial Homeowners’ insurance is the and other “work” animals, were court lacked authority to order vis- “gatekeeper” to homeownership. considered “perfect and complete” itation rights in mere property.336 Without homeowners’ insurance, property.326 Dogs fell in a third cat- The court in Arrington reached a person cannot get a mortgage. egory, that of “cats, monkeys, par- a similar conclusion. Arrington in- Without a mortgage, most people rots, singing birds, and similar ani- volved a custody dispute over Bon- cannot buy a house.314 An insured mals, kept for pleasure, curiosity, nie Lou, “a very fortunate little who chooses to lie about a dog’s or caprice.”327 The Court saw no dog with two humans to shower breed or the existence of a dog useful, social value for dogs, except upon her attentions and genuine altogether is committing policy for companionship, which the love frequently not received by fraud, running the risk of criminal Court dismissed as unsatisfactory human children from their prosecution315 and the complete for the establishment of a property divorced parents.”337 The trial cancellation of his or her policy.316 interest. Thus, the Court held that court had awarded custody of Bon- property interests in animals are on nie Lou to Mrs. Arrington. Mr. a continuum: wild animals (animals Arrington appealed, claiming he ferae naturae) on one end, domes- should have been appointed “man-

38 The State of the Animals IV: 2007 aging conservator” (primary York City civil court judge awarded tent with Sentell. The majority guardian) of Bonnie Lou.338 The $700 in damages to the owner of a went on to hold that the plaintiffs Court held that managing conser- deceased poodle. The dog had could recover only for the loss of vatorships were designed for hu- been euthanized by the defendant, value of prospective puppies but mans, not animals.339 The Court on instructions from the plain- only in the context of how much held, “A dog, for all its admirable tiff.346 “The plaintiff had arranged the animal itself would be worth and unique qualities, is not a for an elaborate funeral...includ- as breeding stock.356 human being and is not treated in ing a headstone, an epitaph, and A concurring judge took a the law as such....A dog is personal attendance by plaintiff’s two sis- broader view of damages in the property, ownership of which is ters and a friend.”347 When the case. He said the award for dam- recognized under the law.”340 plaintiff opened the casket, how- ages should be based on “the There is an indication that the ever, she saw the body of a dead intrinsic or special value of domes- legal status of dogs and other pets cat.348 She brought suit, alleging tic animals as companions and be- may be beginning to change. In that she had suffered emotional loved pets.”357 The market value Raymond v. Lachmann, the court distress as a result of the inci- was inadequate to compensate the had to determine the custody of a dent.349 The Court held that the plaintiffs for the full extent of their cat named Merlin.341 The defen- plaintiff was entitled to sue not loss.358 “It is common knowledge dant originally owned Merlin, but just for the market value of the among pet owners that the death left him for one and a half years dog (for conversion of her prop- of a beloved dog or cat...can be a with a former roommate, the plain- erty) but also for her mental great loss.”359 He called for the tiff.342 During that time, the plain- anguish and suffering in seeing acknowledgment of pets as a spe- tiff renamed him “Lovey” and grew the cat instead of her dog. The cial form of property360 based on to be quite attached to him.343 The Court stated: the relationship between humans trial and appellate courts both This court now overrules prior and their pets: held that Lovey should remain in precedent and holds that a pet Many people who love and the custody of the plaintiff, who is not just a thing but occupies admire dogs as family mem- had taken care of him for a lengthy a special place somewhere in bers do so because of the traits period of time. What is remarkable between a person and a piece that dogs often embody. These about this case is that the court of personal property....A pet is represent some of the best of used a “best interests of the cat” not an inanimate thing that human traits, including loy- standard to decide the issue. The just receives affection it also alty, trust, courage, playful- court discarded strict application returns it....To say that [the ness, and love. This cannot be of property law and in its place poodle] is a piece of personal said of inanimate property. At adopted a version of the “best property and no more is a re- the same time, dogs typically interests of the child” standard pudiation of our humanness. lack the worst human traits, from (human) family law. The This I cannot accept.350 including avarice, apathy, pet- court held: Dicta in other cases demon- tiness, and hatred....Losing a Cognizant of the cherished strate that courts are beginning to beloved pet is not the same as status accorded to pets in our rethink the concept that pets are losing an inanimate object, society, the strong emotions mere property. In Bueckner v. however cherished it may be. engendered by disputes of this Hamel,351 the Texas Court of Even an heirloom of great sen- nature, and the limited ability Appeals had to decide the amount timental value, if lost, does not of the courts to resolve them of damages to be awarded the constitute a loss comparable satisfactorily, on the record owner of then-deceased dogs, a to that of a living being. This presented, we think it best for dalmatian and an Australian shep- distinction applies even all concerned that, given his herd.352 The defendant shot the though the deceased living be- limited life expectancy, Lovey, dogs while they were chasing a ing is a nonhuman.361 who is now almost ten years deer.353 The plaintiffs brought suit Juries have been following this old, remain where he has lived, to recover damages for the loss of trend. In cases where harm had prospered, loved, and been their property, which the trial been done to pets, juries have been loved for the past four years.344 court found “had special value to awarding damages as high as Some courts have also recog- the Plaintiffs and were loved as $35,000. In contrast, the average nized that pets are more than pets by the Plaintiffs.”354 The award in the early 1990s was only a mere property in the context of majority concluded that “Texas few hundred dollars.362 tort awards. In Corso v. Crawford law recognizes a dog as personal Dog & Cat Hospital, Inc.,345 a New property”355—a holding consis-

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 39 poor risks or that they are simply The question then becomes V. Breed uninsurable.375 which factors an insurer may con- Discrimination In 1997 D.S. Hellman evaluated sider in making its underwriting the widespread practice of the decisions. Insurance is a highly Should Be Ended time of insurers in denying health, regulated industry. It does not Through Legislation life, and disability coverage to vic- operate in a regulatory vacuum, tims of domestic abuse.376 She pre- free to let the give-and-take of the or Administrative sented a compelling and detailed marketplace decide who gets in- analysis of the philosophical and surance, how much coverage they Regulation legal implications of this practice, get, and how much it will cost A central principle of insurance ultimately concluding that state them. There is social utility in law is that insurance companies legislatures should intervene and making insurance available to the operate at the pleasure of the prevent underwriting decisions highest number of people possi- states.363 “Indeed, the organiza- based on a customer’s history of ble.383 Insurance allows people to tion of an insurance company and domestic abuse.377 buy homes, afford health care, and the conduct of the business of writ- Hellman’s analysis started with drive automobiles.384 The high ing insurance is not a right but a the premise that insurers had been stakes and high social utility of privilege granted by the State sub- able to draw an actuarially justified insurance have historically justi- ject to the conditions imposed by conclusion that domestic abuse fied strict government regulation it to promote the public wel- victims were, from a statistical of the industry.385 fare.”364 The power to regulate in- standpoint, more likely than oth- All states require underwriting surance is so strong that a state ers to be victimized in the future decisions to be based on actuari- may take over the entire business and, thus, to result in claims ally sound data.386 In Maryland, of insurance if it decides it is in the against their insurers.378 Domestic for example, public interest to do so.365 abuse victims were a higher risk— An insurer or insurance pro- States have the power to regulate so high, the insurers concluded, ducer may not cancel or refuse insurers as an exercise of their po- that the insurance pool could not to underwrite or renew a partic- lice power.366 Although insurance bear to have them as a risk, no ular insurance risk or class of law is governed in part by contract matter how high the premium.379 risk for a reason based wholly or law,367 it is also quasi-public in na- Breed discrimination is an entirely partly on race, color, creed, sex, ture.368 States have the power not different problem altogether. or blindness of an applicant or just to regulate insurance con- There is a lack of statistically and policyholder or for any arbi- tracts, but also to declare the scientifically sound data to show trary, capricious, or unfairly terms and conditions of those con- that certain breeds are more dan- discriminatory reason.387 tracts and to impose additional gerous than others. Even if such Maryland law also provides that duties and obligations.369 On the data existed, a plausible case could underwriting must be accom- other hand, when a state does not be made that the breed of a fam- plished “by the application of stan- regulate a particular practice of the ily’s dog should not be used as a dards that are reasonably related insurance industry, companies are factor in underwriting. to the insurer’s economic and busi- free to contract as they see fit.370 ness purposes.”388 States regulate and legislate Actuarially justified underwrit- insurance on behalf of the public ing is not only the law, it is good interest. Regulations counterbal- Insurers’ Duty In making underwriting decisions, business. By accurately separating ance free market forces to protect insurers decide which of many out risks into “not insurable” and the public at large.371 Some states risks to insure in order to protect “insurable” (and, then, in turn, prohibit unfair and deceptive trade their fiscal solvency and profitabil- separating out insurable risks into practices.372 Some administra- ity.380 When an underwriter de- various risk classifications), actu- tively set rates.373 In determining cides not to insure a particular arially justified underwriting pro- whether a rate is reasonable, states risk, the would-be insured is left to motes efficiency and profit. Con- will look to see if the rate is based find insurance elsewhere. If no sumers are not allowed into the on “legitimate cost factors.”374 insurer will underwrite or accept insurance pool when the likelihood Some states require insurers to the risk, the result may be a cost- of loss is so high that inclusion of write policies for particular risks, shifting to society381 or the loss their risks threatens the viability even though the marketplace may of an economic opportunity to a of the pool itself.389 For those in- have determined such insureds are consumer.382 sureds allowed in the pool, actuar- ially justified underwriting pro-

40 The State of the Animals IV: 2007 motes efficiency by assigning low high that every insurer except the costs. A spokesperson for the III premiums to low-risk insureds and Farm Bureau declined to provide recently conceded, “[T]he industry high premiums to insureds more coverage. The dozen or so insurers isn’t positioned to determine which likely to have a claim.390 This cre- that I contacted in Lubbock who dogs should be deemed vicious.... ates a market incentive for low-risk declined to provide coverage lost [W]e’re certainly not dog experts or insureds to participate in the pool out on what would otherwise be a veterinarians.”405 Unless and until as opposed to engaging in adverse low-risk insured, simply because the industry can demonstrate that selection.391 Accurate risk classifi- they adhered to a hypothesis (rott- different breeds have different rela- cation also maximizes profits for weilers and Chow Chows are more tive dangerousness ratios with some the insurer. By eliminating the dangerous than other dogs) that degree of accuracy, breed discrimi- highest-risk insureds from the has not been scientifically proven. nation should be opposed by the pool, an insurer keeps premiums In my case, the insurer who used a general public, insurers themselves, low for the low-risk insureds who more actuarially sound rate classi- and regulators. remain. An insurer that does not fication structure (the Farm maintain its “classification edge” Bureau) benefited by offering a faces the potential of having its low-risk consumer a low-risk pre- low-risk insureds leave to join mium, thus gaining a market ad- Arguments other companies that are able to vantage over its competition.401 to Support charge lower premiums due to bet- I do not believe there exist suffi- ter risk classification decisions.392 cient data for an insurer to even Regulation in the The insurer is stuck with its high- justify a weak correlation between Public Interest risk insureds as well as the high- breed and bite risk. Insurers should The law is full of examples where risk insureds who migrate over work to minimize the risk of false “actuarially fair” factors have never- from the insurer’s competition.393 positives so as to “fine tune” their theless been prohibited in under- This means that the insurer is not risk classifications to the greatest writing because of overriding public maximizing its profitability. extent possible.402 Risk classifica- interests. Statistical correlation How much statistical correlation tions should be sufficiently refined between behavior and risk, there- is required for a rating factor to be so as not to be overbroad. Exclud- fore, is only the first step in a much “actuarially fair”? How legitimate ing all dogs would clearly be over- bigger, public policy analysis. Drive- do “legitimate cost factors” have broad and would come with high through deliveries,406 preexisting to be?394 Certainly, perfect 1:1 cor- social costs. Excluding some breeds medical conditions,407 civil rights,408 relation is not required.395 Thus, I is also unsound, based on my review and witness intimidation409 are all do not suggest that insurers must of the scientific literature.403 What examples of where otherwise actuar- be able to demonstrate that every I propose—and what the Task Force ially justified practices were prohib- Chow Chow will have an unjust bite on Canine Aggression and Human- ited by state legislatures and courts in its lifetime. Risk classification Canine Interactions proposed404— due to overriding interests in equal- necessarily will involve some “false is the refinement of breed-specific ity, health, and fairness. positives.”396 Otherwise, insurers actions by legislatures and insurers Part IV demonstrated the impor- would be very limited in the classi- to control and regulate “dangerous tance of dogs and other pets in fications they could use, there dogs.” Dangerous dogs are those society. Pets provide physical and would be insufficient stratification who have demonstrated (on an emotional benefits to humans and of the rate pool, and the dangers of individual, dog-by-dog basis) a pro- are not mere property. Even if moral hazard397 and adverse selec- pensity for violence. This would be breed discrimination were actuari- tion398 would increase dramati- actuarially fair because adequate ally justified, I think a plausible cally. On the other end of the spec- evidence exists that a dog with a argument would exist that the trum is the insurers’ position, that history of unjustified bites is likely practice should be regulated be- any correlation is sufficient.399 to be dangerous in the future. cause of the public interest in pro- This is not an economically viable As demonstrated in Part II, there tecting animal-human bonds. position for an insurer, since low- is insufficient evidence to support There is an additional, and risk insureds may be incorrectly the insurance industry’s argument arguably more important, social classified as high-risk customers, that certain breeds bite more often. value that is compromised by and high-risk insureds might be In other words, the current risk breed discrimination: homeowner- priced out altogether.400 For exam- classification (by breed) is too gen- ship. Most home buyers require ple, my ownership of a rottweiler eral and is generating too many homeowners’ insurance in order to and a half-Chow put me in an irra- false positives while at the same purchase a home. This require- tionally high-risk classification—so time having unnecessary social ment comes from mortgagors, who

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 41 require some protection in the and community-wide improvement. not likely to bite someone. Insur- event their security (the home For many families, homeownership ers still turned me away. They re- itself) is destroyed, damaged, or is the way to accumulate wealth for fused to write a policy with a dog- otherwise made unavailable for the future.417 Home equity can be bite exclusion in it. collection.410 As the Seventh Cir- borrowed against for emergencies, There are several good reasons cuit stated in NAACP v. American higher education, or retirement.418 why exclusions may not be good Family Mutual Insurance Co.,411 The family home is often the most public policy or wise business “No insurance, no loan; no loan, no significant component in an estate sense. Exclusions operate to the house; lack of insurance thus after a parent dies.419 detriment of third parties, those makes housing unavailable.”412 Breed discrimination should, would-be plaintiffs who are injured The issue in American Family was thus, be viewed in a larger social and need compensation for their a practice known as “red-lining” context. There is a high social cost loss. Exclusions would create where homeowners’ insurance when someone is denied homeown- pockets of plaintiffs who would, in companies were charging higher ers’ insurance: he is unable to buy effect, have no way to satisfy a rates, or declining to write insur- a home.420 The social harm in pre- judgment if they could prove liabil- ance altogether, based on geo- venting the dream of homeowner- ity. This is not an insignificant graphic location of insureds.413 ship must be weighed against the public policy, for the same reason The boundaries (“redlines”) that small risk of a dog-bite claim. that states require certain profes- defined the no-insurance zones fre- There are over fifty million dogs in sionals to have liability insur- quently fell along racial and socioe- the United States, yet only a few ance421 and drivers to carry mini- conomic lines, and the NAACP dogs have been responsible for bit- mum limits on their automobile brought suit alleging that this ing people. policies,422 to provide a source of practice was discriminatory and This is not a simple matter of de- recovery for third parties in the illegal. The Seventh Circuit held ciding to throw away the family event of a legitimate claim. If we that red-lining violated the Fair trampoline or forgo the purchase exclude dog bites or even those Housing Act, a statute passed by of an in-ground pool. Pets are not dog bites from breeds we can Congress to prohibit discrimina- mere property. To make people prove are the most dangerous, we tion in the housing market.414 choose between the family pet and would run the risk of creating a It is quite possible that red-lin- homeownership is unfair, unneces- special class of plaintiffs who ing was actuarially justified; that is, sary, and goes against an impor- would have no source of recovery. it may have in fact cost insurance tant social value: homeownership. Plaintiffs would have to turn to companies more to write policies other sources in order to have in certain areas than others. This, their basic medical needs met.423 however, did not end the inquiry How Else Exclusions are also bad for busi- for Congress or the Court of ness because they make insurance Appeals. The Seventh Circuit held Can Insurers less attractive to consumers. A per- that homeowners’ insurance is a Control Risk? son with cancer is a much higher service that has the power to make Let me assume for the moment risk than a healthy individual. If a homeownership available.415 If a that insurers could demonstrate health insurer began excluding plaintiff can demonstrate that an with some degree of actuarial con- coverage for cancer treatment, few application for homeowners’ insur- fidence that some breeds are more employers or individual consumers ance was rejected or unfairly rated likely to bite than others. Could would purchase that company’s on the basis of race or another pro- there be other ways of controlling insurance. My decision to try to hibited factor, the practice consti- this risk, short of outright denial bargain my way into the insurance tutes discrimination in housing.416 of coverage? risk pool by excluding dog bites Homeownership is a worthwhile from coverage was, in reality, public interest. People who own Exclusions pretty stupid. In the rare event their homes develop roots in a When I was shopping for homeown- that I was found liable for one of given community. A homeowner is ers’ insurance, one of the first my dogs biting someone, I would less likely to leave than is someone questions I asked insurers was be solely responsible for the judg- who is in a year-to-year or month- whether they would write a policy ment against me. I would lose to-month lease. The homeowner, with an exclusion for dog bites. I whatever equity I had in my house, therefore, has a personal invest- did this because I was desperate— my car, my savings, and I could ment in the well-being of the com- I needed insurance and I was will- have my wages garnished. In retro- munity. Homeownership provides ing to assume the risk that my spect, an exclusion would not have an incentive for civic involvement dogs were not dangerous and were been a good choice for me.

42 The State of the Animals IV: 2007 Insure but Reclassify are more likely to have claims fact true and therefore see little Another option would be for insur- against them. Let’s be clear: this incentive to market themselves to ers to write policies for families is not what is going on right now. people they view as high-risk. For with “dangerous” breeds but The current practice of breed dis- these reasons, it is unlikely that charge them higher premiums. crimination is to differentiate the marketplace will correct itself Risk classification is an accepted among breeds, even though there to end breed discrimination. practice in the insurance indus- is no statistical evidence to prove try.424 By separating and grouping that certain breeds are more dan- Other Solutions people of similar risks, insurers gerous than others. This creates an Preventing law-abiding homeown- keep rates low for the desirable, artificial risk classification that ers from obtaining insurance is not low-risk insureds, and insure ade- charges owners of certain breeds the answer to the problem of dog quate resources in the event that more than others. bites. Better and more effective high-risk insureds cause a claim.425 If all dog owners were classified alternatives exist.429 I would have the same objection to at a higher rate than non-dog own- high-risk classification for owners ers, I think there would be a great Collect Better Data of certain breeds as I would for out- public outcry. Then the social value An initial first step would be to right refusals to insure, that is, the of dogs as pets—and as security improve surveillance and reporting lack of actuarial justification for alarms on four paws—would come of dog bites. Until accurate num- the practice of breed discrimina- to the forefront of the debate. bers for the numerator and denom- tion. Classifying certain dog own- inator in the relative dangerous- ers in a higher category is unfair Allow the Marketplace ness ratio can be ascertained, because it places those insureds in to Correct Itself insurers and governments will be an artificially higher rate bracket. If, as I conclude, there are no reli- without realistic data on which to This is economically inefficient, able data to support breed discrim- base meaningful decisions. The although perhaps more profitable ination, then there is a market of need for more accurate data col- for the insurer. consumers (owners of rottweilers, lection has been championed by Where I think risk classification pit bulls, etc.) being overcharged the very scientists who have tried could work is if insurers could or not served altogether. This cre- to calculate the scope of the dog- demonstrate—to the veterinary ates an economic inefficiency. An biting problem.430 In addition, and CDC communities with a suffi- insurer with good business judg- studies should be commissioned to cient degree of scientific cer- ment would seek to corner this determine if certain breeds, when tainty—that certain breeds, when underserved market by writing they do bite, cause more physical they do bite, cause more damage. policies with low-risk premiums. injury or damage.431 It is hypothesized, for example, There are a number of reasons that the jaw structure of pit bulls why the market is not correcting Enforce Existing Laws causes them to inflict more injury itself. The number and identity of Against Dogfighting than other breeds.426 This would people being affected by breed dis- and Dogs at Large 427 crimination is unknown. Without still be breed discrimination There are existing laws that, if these data it would be difficult for but, in my view, an acceptable form enforced more vigorously, could an insurer to market itself to those of risk classification...provided reduce the number of dog bites. consumers. Also at work is the fact there is a scientific/veterinary Dogfighting explains why some that insurers try to market them- basis for the conclusion. To date, dogs are vicious. This underground selves to the lowest-risk con- the studies in this area have fo- industry brings some dogs “to the sumers. Although these consumers cused on determining the number verge of bloodlust.”432 By shutting pay lower premiums, they are re- of bites per breed, not the amount down criminal organizations of ille- sponsible for fewer claims. Every of damage per bite. gitimate breeders, promoters, and insurer tries to maximize its num- I believe insurers would also be owners, local governments could ber of low-risk insureds while max- actuarially justified in classifying take a first step toward reducing imizing the number of high-risk homeowners based on whether or bites by dogs that have been pur- insureds who are serviced by its not they own a dog, period. One posely bred to be dangerous.433 The competitors.428 The insurance in- does not need to be an actuary to AKC and other groups support the dustry as a whole appears to be state that a dog owner is more use of existing laws to break up caught up in this breed discrimina- likely than a non-owner to have a dogfighting rings.434 tion hysteria. Individual companies bite claim against him. Insurers Many attacks appear to be caused may fear that the assumptions could simply classify all dog owners by strays or dogs who have been per- at a higher rate level because they behind breed discrimination are in

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 43 mitted to run off-leash.435 The provides due process protections Children must also be educated enforcement of existing laws against to the owner—requiring a hearing about dealing with dogs safely.455 At “dogs at-large” could reduce the by a judge and a finding of danger- least one study has demonstrated number of bites.436 While these laws ousness before a disposition is the effectiveness of public education exist in many places, they are not ordered.441 A dangerous animal is as a way to improve children’s behav- adequately enforced. one who, without just cause,442 ior around and toward dogs.456 The Owners are sometimes to blame bites or attacks a person, or a dog study, conducted in Australia, exam- for socializing a dog to be danger- who bites or attacks and causes ined the reactions of children, ages ous or for permitting it to get into serious injury or death to another seven to eight, to a dog that was tied situations where it can cause dog while the other dog is on the up in their playground.457 Half of the injury. Dogfighting, leash, and at- property or under the control of study group received a thirty-minute large laws address the root of the its owner.443 Oklahoma has a simi- classroom lesson seven to ten days problem, which is irresponsible dog lar statute that allows for height- before on how to safely approach ownership. A dog is just as good as ened regulation of animals and treat dogs.458 Researchers ob- his owner trains him to be. One declared dangerous by their served the reactions of the children problem dog can be seized and conduct,444 but prohibits local to the dog.459 The group that re- destroyed. One problem owner, governments from enacting breed- ceived the classroom instruction dis- however, can continually breed, specific legislation.445 played greater precautionary behav- adopt, or purchase dog after dog. “Most of the approximately 55 ior than did the control group. Replacing one dog with another, or million dogs in the United States While 79 percent of the control one breed with another, will not never bite or kill humans.”446 Dan- group hastily patted the dog and help to reduce the overall problem gerous-dog laws are narrowly tai- tried to excite it, only 9 percent of of owner irresponsibility.437 Exist- lored to address the real problem, the group that received instruction ing laws can and should be used to which is the small percentage of did so.460 address the behavior. the overall dog population that is responsible for bites, injuries, and Regulate Problem deaths.447 Dangerous-dog laws Conclusion Dogs with Existing exist in many states. Insurers could While dog bites are serious events “Dangerous Dog” Laws work with local governments to for those who are bitten, the dog- fund additional animal-control offi- bite problem is not the public Some dogs, as a result of socializa- cers or work with owners of dan- health crisis that the insurance tion (or lack thereof), bad tem- gerous dogs to help take steps to industry has made it out to be. perament, or genetics, demon- prevent future dangerous acts.448 Some perspective is in order. The strate that they are dangerous. number of fatalities due to dog They have a history of bites or Educate the Public, bites is very low when compared to attacks against people or other the number of people who die from animals.438 By regulating these Particularly Children, heart disease, cancer, accidents, individual dogs, municipalities can about Animal Behavior suicide, and diabetes. Likewise, focus their efforts on the specific Insurers and local governments nonfatal bites are responsible for a dogs likely to cause injuries in the could partner together to educate small number of injuries when com- future.439 Instead of targeting an the public about proper ways of pared to other accidental, uninten- entire breed, governments can socializing and approaching dogs. tional injuries. Falls (11.5 million), address the handful of dogs who Proper training is essential for a motor vehicle accidents (4.3 mil- are really the problem. family with a new dog.449 Public lion), drugs (3.3 million), sports There are existing laws that per- education about the importance (2.0 million), insect bites (1.7 mil- mit local governments to regulate, of neutering can reduce the inci- lion), bicycle accidents (1.4 mil- or in some cases seize and destroy, dence of dog bites450 because a lion), poisoning (.7 million), and dogs who have demonstrated a disproportionate number of bites knives (.6 million) all individually propensity to bite without just are caused by intact dogs.451 New outrank dog bites (.5 million) as cause. Michigan enacted a statute owners should also be educated public health problems.461 Simi- to permit local governments to about the steps in picking the larly, claims paid out by homeown- seize “dangerous animals” and right dog for a household.452 ers’ insurance companies for dog have them tattooed, insured, “[T]here is no all-around best bites are miniscule when compared fenced, sterilized, destroyed, “or breed.”453 Certain breeds will be to payouts for property damage. any other action appropriate to more compatible with certain Damage due to fire, water, wind, protect the public.”440 The statute types of families.454 and theft represents much larger

44 The State of the Animals IV: 2007 problems for homeowners’ insur- The science behind dog bites is lection that I have highlighted ance companies. inconclusive at best. Most of the here. The irony is that insurers who One way to eliminate the entire scientists authoring studies on dog are practicing breed discrimina- problem of dog bites would be to bites have acknowledged that their tion are turning away good cus- outlaw all dogs.462 Without dogs, data are incomplete and should not tomers who pay premiums. Leg- there would be no dog bites and be used to enact breed-specific leg- islative action to correct this no dog-bite-related insurance islation or to deny insurance to practice will benefit both families claims.463 While this would result in families with certain dogs. No study with dogs and the shareholders of an elimination of the perceived has accurately or completely deter- insurance companies. financial burden to insurers, it mined the number of bites per Legislative action in this area is would not be “practical, realistic, or breed, or the number of dogs per both appropriate and necessary. desirable” to the average layman, breed. Without these numbers, it is What happened to me is happening scientist, or dog owner.464 Unless we impossible to compare breeds on across the country to thousands of as a society are willing to disregard the basis of dangerousness. Insur- other families. To some insurers, the social and health benefits of ers who are making judgments dogs are mere property—like an dogs as pets, then we must be will- about certain breeds are doing so old can of paint that can be left ing to accept a certain number of without adequate scientific evi- behind when a family moves. The bites. While “[t]he dog bite problem dence. This is the Achilles’ heel of truth is that dogs are members of as a whole is not preventable, it is breed discrimination; by acting the American family and deserve to controllable.”465 Better alternatives without adequate evidence, the in- be treated as such. When families to breed discrimination exist, such surance industry has left itself open are forced to make the choice as education and enforcement of to regulation by the states. between owning a home and having existing dangerous dog laws. State regulation is necessary to a dog, some have no choice at all; With over 34 percent of house- correct this injustice in the mar- they must give up their beloved pet holds owning at least one dog as a ketplace. Insureds are being shut to an animal shelter. There are doc- pet, dogs have become valued four- out of entire markets because of umented increases in “shelter legged members of our society. To the near-hysteria that has gripped drop-offs” due to breed discrimina- the families that love them, pets the insurance industry. This is not tion. These animals cannot be are not mere chattel. Refusing to a new phenomenon for the indus- housed indefinitely, so many have write homeowners’ insurance poli- try. In the past insurers have cut to be euthanized. cies, therefore, should be a nar- benefits and denied applications The social cost to families is too rowly curtailed remedy, limited to for insurance based on fiscal cost- much to ride on incomplete statis- those families that own dogs who benefit analyses that have had col- tics and hunches by insurance have proven to be dangerous to life lateral social and health conse- executives. Legislative action is or property. The insurance industry quences. It was more costly to necessary. Luckily, many state leg- has chosen to paint with a very keep new mothers in the hospital islators have become aware of this broad brush. Breed discrimination for forty-eight hours. Our society problem and have taken steps to is an overreaction, an attempt to came to the recognition, however, end breed discrimination. Pennsyl- solve a small problem by prejudg- that discharging new mothers and vania enacted a statute prohibiting ing all dogs of certain breeds as their newborns within six hours of breed discrimination, which states likely to be dangerous in the future. delivery was against public policy. the following: When insurers develop under- Legislatures stepped in to correct No liability policy or surety writing standards and decide which the injustice in the marketplace, bond issued pursuant to this risks to insure, they have a respon- knowing full well that it would cost act or any other act may pro- sibility to the public interest. In- the industry more money. The hibit coverage from any spe- surers do not contract with con- same should be done here. cific breed of dog.466 sumers in a vacuum. A long history To the insurance industry, breed New York is considering legisla- of state regulation of the industry discrimination reflects a belief that tion that would outlaw breed dis- serves as a backdrop for this issue. denying coverage to families with crimination as well. Bill 6761 Underwriting decisions should be certain breeds of dogs will save would prohibit insurers from refus- the product of reason, not specula- them money. Insurers have not ing to issue or renew, canceling, or tion. In other words, if insurers are produced scientific proof that dogs charging or imposing an increased going to engage in breed discrimi- of certain breeds bite more often premium or rate for owning a dog nation, they better have hard sci- or cause more damage. The evi- of a specific breed.467 A New Hamp- ence to back up their practice. dence simply does not exist be- shire bill would prohibit nonre- cause of the problems of data col- newal or cancellation of a policy

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 45 “based solely on the insured own- breedspecific.html (accessed June, 2004). 38Ibid. 19 39 468 Ibid. See, for example, ibid., 361–63. ing a certain breed of dog.” 20Ohio Rev. Code Ann. § 955.11(a)(1)(B) 40See, for example, Hearn v. City of Over- Other states should follow suit and (4)(a)(iii) (West 2004). land Park, 772 P.2d 758, 762–65 (Kan. 1989). enact legislation or administrative 21Ibid. § 955.11. For a general discussion 41See, for example, ibid. of Ohio’s regulation of pit bulls, See Paula 42See, for example, ibid., 766. regulations to prohibit the prac- Lynn Wilson, Note, “A.M. Sub. H.B. 352: An 43See, for example, ibid. tice of breed discrimination. Overview—Dogs Under Control,” University 44533 N.E.2d 642 (Mass. 1989). of Dayton Law Review 13 (1988), 297. 45The ordinance prohibited the sale of pit 22Ohio Revised Code Annotated § bulls and required the registration of any pit Notes 955.22(C) (West 2004). bulls within city limits. Ibid., 644. Transporta- 1The Farm Bureau provides a number of 23Ibid. § 955.22(D)(2). tion of pit bulls was permitted only if they were services to its members, including insurance 24Ibid. § 955.22(E). muzzled and leashed. Even then, they could and banking. See Texas Farm Bureau, 25Denver, Colorado, Revised Municipal only be transported to a veterinarian. Ibid. http://www.txfb.org (accessed July 7, 2004). Code § 8-55(a) (2003). The ordinance pro- 46Ibid., 644. For further discussion of the 2On the other hand, renting would not vides several exceptions, including a grandfa- problem of identifying dogs by breed, see infra necessarily have been an easy task either. ther clause, possession by animal shelters or Part II. Many landlords prohibit certain breeds from humane societies, public exhibition, or trans- 47American Dog Owners Association, Inc., living on their property or forbid dogs alto- portation through the city. Ibid. § 8-55(c). 533 N.E.2d, 644. gether. Obtaining renters’ insurance would 26Ibid. § 8-55(b)(2). For an analysis of the 48Ibid., 646. also have been difficult because of breed dis- problem of defining “pit bull,” see infra notes 49Ibid. crimination by insurance companies. 240–241 and accompanying text. 50Ibid. 3See infra Part IV. for a discussion of the 27Ibid. § 8-52(a)(1). 51Ibid. The purpose of this article is not to effects of breed discrimination. 28Ibid. § 8-52(c)-(d). address the constitutionality of BSL, but rather 4See 42 U.S.C. §§ 3601-31 (2003); Mich. 29Fla. Stat. Ann. § 767.14 (West 2003). to focus on the insurance aspect of breed dis- Comp. Laws § 37.2302 (2001); N.J. Stat. Ann. 30See Grey, supra note 17, 418. crimination. For further discussion of BSL, see §§ 10:5-4, 10:5-9.1 (2002); see also NAACP v. 31Minn. Stat. Ann. § 347.51(8) (West 2004). Linda S. Weiss, “Breed-Specific Legislation in the American Family Mutual Insurance Company, 32See generally Russell G. Donaldson, United States,” Animal Legal and Historical Web 978 F.2d 287, 297-98 (7th Cir. 1992) (Fair Annotation, “Validity and Construction of Center (2001), http://www.animallaw.info/ Housing Act is applicable to allegations of Statute, Ordinance or Regulation Applying to articles/aruslweiss2001.htm (accessed June racially discriminatory red-lining). Specific Dog Breeds, Such as ‘pit bulls’ or 8, 2004); Randall Lockwood, “Humane Con- 5American Veterinary Medical Association, ‘Bull Terriers,’” 80 American Law Reports 4th cerns About Dangerous Dog Laws,” University Task Force on Canine Aggression and Human- 70 (2004). For discussion and background on of Dayton Law Review 13 (1988), 267; Grey, Canine Interactions, “A Community Approach challenges to BSL, see Thorne, supra note 17; supra note 17; Ohio Valley Dog Owners, Inc., to Dog Bite Prevention,” Journal of the Ameri- Lynn Marmer, Comment, “The New Breed of “OVDO Is Opposed to Breed-Specific Bans” can Veterinary Medicine Association 218 Municipal Dog Control Laws: Are They Consti- (2003), http://www.canismajor.com/orgs/ (2001), 1732, 1733 (hereinafter Task Force). tutional?” University of Cincinnati Law ovdo/bslho.html (hereinafter OVDO); Cindy 6See Richard Willing, (2000). “Under Law, Review 53 (1984), 1067; Heather K. Pratt, Andrist, Note, “Is There (and Should There Pets Are Becoming Almost Human,” USA Comment, “Canine Profiling: Does Breed-Spe- Be) Any ‘Bite’ Left in Georgia’s ‘First Bite’ Today (Sept. 13, 2000), 1A. cific Legislation Take a Bite Out of Canine Rule?” Georgia Law Review 34 (2000), 1343; 7Fred Bayles, “Pit Bullterriers: Too Fierce Crime?” Pennsylvania State Law Review 108 Marmer, supra note 32, at 1067; Pratt, supra to Live? Call for Ban Follow Maimings, (2004), 855. note 32, at 855; Sullivan, supra note 34, 279; Death,” The Record (Dec. 30, 1985), B16. Cases in which courts have upheld the con- Diane Blackman, “Practicality of Breed 8Ibid. stitutionality of BSL include: Vanater v. Vil- Specific Legislation in Reducing or Eliminat- 9Ibid. lage of South Point, 717 F. Supp. 1236 (S.D. ing Dog Attacks on Humans and Dogs,” 10Ibid. Ohio 1989); Colorado Dog Fanciers, Inc. v. http://www.dog-play.com/pitbull.html (accessed 11Ibid. City and County of Denver, 820 P.2d 644 July 14, 2003). 12Garcia v. Vill. of Tijeras, 767 P.2d 355, (Colo. 1991) (en banc); State v. Peters, 534 52“Insurance Industry Discriminates 362 (N.M. Ct. App. 1988). So. 2d 760 (Fla. Dist. Ct. App. 1988); Hearn Against Dogs,” CBS News: Evening News with 13Bayles, supra note 7, B16. v. City of Overland Park, 772 P.2d 758, 766 Dan Rather (CBS television broadcast, June 3, 14Ibid. (Kan. 1989); Garcia v. Village of Tijeras, 767 2003), available at 2003 WL 5212276. 15Bob Egelko, “Appeal for Murder Rap in P.2d 355 (N.M. Ct. App. 1988); State v. Ander- 53Ibid. Dog-Maul Case/Attorney General Says Judge son, 566 N.E.2d 1224 (Ohio 1991); State v. 54See, for example, Jeff Bertolucci, “Man’s Had No Right to Let Knoller Off So Easy,” San Robinson, 541 N.E.2d 1092 (Ohio Ct. App. Best Friend but Insurers’ Foe; Their Assembly Francisco Chronicle (April 12, 2003), A15. 1989); American Dog Owners Association v. Bill Has Failed, But Dog Lovers Continue to 16Kenneth Phillips, “Diane Whipple Case,” City of Yakima, 777 P.2d 1046 (Wash. 1989) Rail Against Breed Discrimination,” Los Ange- http://www.dogbitelaw.com/PAGES/ (en banc). les Times (June 6, 2004), K1 (discussing fail- Whipple.html (accessed Jan. 8, 2004). The Massachusetts Supreme Judicial ure of Bill AB 2399, prohibiting breed discrim- 17Karyn Grey, Note, “Breed-Specific Legis- Court, however, found a pit bull ban in Lynn, ination but allowing a higher premium and lation Revisited: Canine Racism or the Answer Massachusetts, to be unconstitutional. Ameri- requiring a discount for canines that pass an to Florida’s Dog Control Problems?,” Nova can Dog Owners Association, Inc. v. City of obedience test); Vincent J. Schodolski, “‘Bad’ Law Review 27 (2003), 415, 417 (describing Lynn, 533 N.E.2d 642, 646 (Mass. 1989). Dogs Put Costly Bite on Insurers, Homeown- a series of dog attacks in Florida); see Julie A. 33U.S. Constitution, amendment V. ers,” Chicago Tribune (May 17, 2004), 1; Jeff Thorne, Note, “If Spot Bites the Neighbor, 34Sallyanne K. Sullivan, “Banning the Pit Bertolucci, “Is Nothing Private? Home Insur- Should Dick and Jane Go to Jail?” Syracuse Bull: Why Breed-Specific Legislation Is Con- ers Ask About Everything from Rover to Law Review 39 (1988), 1445, 1445 (assert- stitutional,” University of Dayton Law Review Rolexes. And the Answers Matter,” Los Angeles ing that media attention fueled public out- 13 (1988), 279, 280–93. Times (May 9, 2004), K1; Allan Woods, “Rot- rage against pit bulls); see generally Jeffrey J. 35See, for example, Garcia, 767 P.2d at tweilers, Pit Bulls New Insurance Liability,” Sacks et al., “Fatal Dog Attacks, 1989–1994,” 358. Since dog ownership is not a “fundamen- National Post (March 26, 2004), A3; Michele Pediatrics 97 (June 1996), 891 (hereinafter tal right,” BSL need only meet the “rational Derus, “Dog Bites Giving Insurers Pause,” Mil- CDC 1989–1994). relationship” test to be constitutional. Sulli- waukee Journal and Sentinel (Feb. 29, 2004), 18For a list of communities that have van, supra note 34, 281. 1F; Gloria Campisi, “Beware of Dog When adopted or are considering BSL, see Jan 36See, for example, Garcia, 767 P.2d at Seeking Insurance; Some Firms Have ‘Bad Cooper, “Breed Specific Legislation,” 358–62. Breed’ Lists,” Philadelphia Daily News (Oct. 7, http://www.rott-n-chatter.com/rottweilers/laws/ 37See, for example, ibid., 361. 2003), 14; Ryan Slight, “Liability Factor Can

46 The State of the Animals IV: 2007 Hurt Homeowners,” Springfield News-Leader (accessed July 5, 2004). See Sodergren, supra troversy,” Anthrozoös 1 (1987), 2, 4 (damage (Sept. 28, 2003), 6A; William Sweet, “Insurers note 59. caused by pit bulls is generally more severe in Doghouse with Some Pet Owners,” Spring- 64Davis, supra note 58. due to the attack behavior of the breed). field Union-News (Aug. 19, 2003), A01; Purva 65Insurance Information Institute, “Home- 84Davis, supra note 58, 36 (quoting Ale- Patel, “A Bite to the Pocket, Home Insurers owners Insurance,” http://www.iii.org/media/ jandra Soto, spokesperson for the III). Often Charge Higher Premiums Because of facts/statsbyissue/homeowners/ (accessed 85Ibid., 38. This policy of exclusion is Dogs,” Sun-Sentinel (Ft. Lauderdale, Florida) June 11, 2004) (hereinafter III, Homeowners not without problems, as it leaves plaintiffs (Aug. 12, 2003), 1D; Charles Toutant, “Insur- Insurance). without a remedy if a dog bite does occur ers Attempt to Leash Dog-Bite Claims: Small- 66Insurance Information Institute, “Dog and subjects a homeowner to loss of the home Scale Nuisance Litigation Turning Into Big Bite Liability,” http://www.iii.org/media/ and bankruptcy. Kenneth Phillips, “Breed Business,” Connecticut Law Tribune 29 (Aug. hottopics/insurance/dogbite/ (accessed June Specific Laws, Regulations and Bans,” avail- 11, 2003), 8; Jim Spencer, “Homeowners 11, 2004) (hereinafter III, Dog Bite Liability). able at http://www.dogbitelaw.com/PAGES/ Insurance Rules Not for the Dogs,” Daily Press 67III, Homeowners Insurance, supra note breedlaws.html (accessed July 21, 2003) (Virginia) (Jan. 10, 2003), C1. 65. (Phillips is a lawyer and expert on dog-bite Breed discrimination is not confined to 68III, Dog Bite Liability, supra note 66. The law). Litigation over exclusions can be costly American insurers, either. Reports have also insurance industry points to rising construc- and lengthy and can lead to uncertainty in come from Canada describing the practice tion costs and natural disasters, such as wind the marketplace. there by Canadian insurance companies. “Cal- and hail, as principal causes for the rise in pre- 86See Davis, supra note 58, 36 (“We don’t gary Man Denied Home Insurance Renewal miums. Ibid. discriminate or deny coverage based on breed Due to the Type of Dogs He Has,” Canadian 69Davis, supra note 58, 36 (quoting Ale- of dog.”); McMahon, supra note 55; Hattaway, Press (March 25, 2004), available at 2004 WL jandra Soto, spokesperson for the Insurance supra note 70 (author works for State Farm). 74118306; “Some Insurance Companies Information Institute, a trade group of the 87Franco Ordonez, “Pet Peeve: A New State Won’t Cover You if You Own Certain Dogs,” insurance industry). Regulation Prohibiting Owners of Some Dog CTV News-PM (CTV television broadcast, 70Dan Hattaway, “Dogs and Insurance,” Breeds From Adopting Draws Fire,” Boston March 24, 2004), available at 2004 WL Journal of the American Veterinary Medical Globe (Jan. 15, 2003), B1. 60848476. Association 210 (April 15, 1997), 1143. 88Ibid. 55See Patrick McMahon, “Dog Owners’ 71Sodergren, supra note 59. 89Phillips, supra note 85. New Policy: Bite Back,” USA Today (May 20, 72III, Dog Bite Liability, supra note 66. 90Infra Part III. 2003), 03a. 73Nationwide Insurance stated to USA 91Jeffrey J. Sacks et al., “Dog Bite-Related 56Ibid. Today that it relies on CDC data to support its Fatalities From 1979 Through 1988,” Journal 57Ibid. I wonder if breed discrimination is breed-specific policies. McMahon, supra note of the American Medical Association 262 encouraging homeowners, who are in a situa- 55, 03a. (1989), 1489 (hereinafter CDC 1979–1988). tion like mine or the Craanens’, to engage in 74Infra Part II. 92CDC 1989–1994, supra note 17. policy fraud by lying about their pet’s breed or 75III, Dog Bite Liability, supra note 66. 93Centers for Disease Control, “Dog Bite by hiding the dog altogether. In Mutual Bene- 76Ibid. One insurance company, State Related Fatalities—United States, 1995–1996,” fit Insurance Co. v. Lorence, No. 02-1734, Farm Fire and Casualty Co., reported that in Morbidity and Mortality Weekly Report 46 2003 WL 1354845 (4th Cir. Mar. 20, 2003), 1995 it paid out $70 million on eleven thou- (1997), 463 (1997) (hereinafter CDC homeowners lied on their insurance applica- sand claims. That year, it insured a total of five 1995–1996). tion about owning dogs. One of their dogs (a million dog-owning homes. See Hattaway, 94Jeffrey J. Sacks et al., “Breeds of Dogs pit bull) subsequently attacked a mother and supra note 70, 1144. Involved in Fatal Human Attacks in the United her son, resulting in a claim against the 77III, Dog Bite Liability, supra note 66. States Between 1979 and 1998,” Journal of homeowners’ insurance. The insurer brought 78Ibid. (emphasis added). the American Medical Association 217 (2000), suit against the owners, seeking a declaratory 79III, Homeowners Insurance, supra 836 (hereinafter CDC 1979–1998). judgment that it was not required to defend note 65. 95Ibid., 838. the claim because of the misrepresentation. 80Ibid. A plausible argument could be 96See infra Part II. Ibid., *596. The courts, however, declined to made that dog bites are a preventable risk, in 97CDC 1979–1988, supra note 91, 1489. exercise jurisdiction and instead deferred to a way that damage due to fire, mold, and theft 98Ibid. The NCHS uses reports from hospi- ongoing proceedings before the Maryland are not. Property damage claims are the result tals and other sources to classify deaths. Insurance Administration. Ibid., 597. The of true accidents—unforeseeable acts for There is a category of deaths just for dog owners had filed a complaint with the MIA which insurers have little to no control over. bites: “E906.0.” National Center for Health alleging that the insurance company’s prac- Dog bites are one small way in which insurers Statistics et al., International Classification of tice of breed discrimination is against Mary- can try to minimize risk. Diseases, 9th Revision, Clinical Modification land law. Ibid., 596–597. In reality, big-ticket risks can be con- (2003), http://ww.eicd.com. 58Stephanie Davis, “ASPCA/HSUS Cam- trolled. Insurers could decide, for example, 99CDC 1979–1988, supra note 91, 1489. paign Tackles Insurance Industry,” DVM: The not to insure a home unless it has smoke The HSUS reports are particularly detailed Newspaper Magazine of Veterinarian Medicine detectors. This would provide an incentive for and are believed to be the most accurate in 34 (November 2003), 1. (“Some household people to install smoke detectors, which determining the true breed of a biting dog. name insurance companies either have can- would, in turn, reduce the number of fire- CDC 1989–1994, supra note 17, 891. HSUS celed or refused to write homeowners’ poli- related claims. Insurers could also justify staff reviewed police reports, animal control cies for individuals with certain dog breeds.”) refusing to write policies for homes where reports, statements by dog owners and vic- 59See Brian Sodergren, “Insurance Compa- there is a high risk of wind or hail damage. tims, and photographs. Ibid. nies Unfairly Target Specific Dog Breeds,” Any risk is controllable, but only to the extent 100CDC 1979–1988, supra note 91, 1490; http://www.hsus.org/ace/18624 (accessed that the insurance market can bear the loss of CDC 1989–1994, supra note 17, 892. June 11, 2004) (perceived increase in the this business. 101CDC 1979–1988, supra note 91, 1490. number of claims, or risk, exists for people 81As shown, supra, while the industry’s 102Ibid. Between 1989 and 1994, that fig- who own certain breeds of dogs). aggregate figures may sound “scary,” they ure had dropped to 56.9 percent. CDC 60See infra Part I. for the insurance indus- misstate the scope of the dog-bite problem in 1989–1994, supra note 17, 892. From 1995 try’s arguments in favor of breed discrimina- the larger context of total claims paid. through 1996, that number climbed to 80 tion. 82III, Dog Bite Liability, supra note 66. percent. CDC 1995–1996, supra note 93, 61Davis, supra note 58. 83Davis, supra note 58, 36 (“We just know 463. From 1997 through 1998, the number 62Sodergren, supra note 59. that certain breeds, when they do attack, tend settled at 70 percent. CDC 1979–1998, supra 63Dog owners wishing to report incidents to cause a lot more damage when they do bite, note 94, 837. of breed discrimination to The HSUS can do not because they bite most often.”); Randall Victims were often elderly, suggesting that, so at the following website: http://files.hsus.org/ Lockwood and Kate Rindy, “Are ‘Pit Bulls’ Dif- “The main victims of fatal dog bites were the web-files/PDF/insurance_incident_form.pdf ferent? An Analysis of the Pit Bull Terrier Con- very young and very old, those least able to

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 47 protect themselves.” CDC 1979–1988, supra the last two digits. They called numbers until 180Ibid., 317. The survey included children note 91, 1492. an eligible control dog was found. in preschool to twelfth grade. 103CDC 1979–1988, supra note 91, 1490. 145Ibid., 915. 181Ibid., 319. The study went on to find 104CDC 1995–1996, supra note 93, 463. 146Ibid., 914; see also supra notes 25–28 that children’s attitudes toward dogs were not 105Ibid. and accompanying text. affected by being bitten: “Children appear to 106Ibid. 147Gershman, supra note 139, 916. accept being bitten by dogs much as they do 107CDC 1979–1998, supra note 94, 839. 148Ibid., 914. other accidents such as falling off a bike. The Sacks/Sinclair study attempted to piece 149Ibid., 916. Being bitten had little influence on their lik- together the data on attacks by breed from 150Jeffrey J. Sacks, Marcie-jo Kresnow, and ing for dogs.” the previous three studies as well as new data Barbara Houston, “Dog Bites: How Big a Prob- 182See Ibid., 317–320. from 1997 and 1998. Ibid., 837. lem?” Injury Prevention 2 (1996), 52 (two of 183Lee E. Pinckney and Leslie A. Kennedy, 108CDC 1979–1988, supra note 91. the three scientists who conducted the survey “Traumatic Deaths From Dog Attacks in the 109Ibid. were with the CDC at the time). United States,” Pediatrics 69 (1982), 193. 110CDC 1979–1998, supra note 94, 837. 151Ibid., 52. Only 48 percent of the newspapers responded. 111Ibid. 152Ibid. 184Ibid. 112Ibid. This accounted for approximately 153Ibid. 185Ibid., 194. 28 percent of the deaths during the study 154Ibid. 186Ibid. period. 155Ibid., 53. 187Ibid. 113CDC 1979–1988, supra note 91, 1492. 156Ibid. 188CDC 1979–1988, supra note 91, 1489. 114Ibid., 1489. For further discussion of 157Ibid. 189Pinckney and Kennedy, supra note 183, the problem of breed misidentification, see 158Ibid. 195. infra Part II. 159Harold B. Weiss, Deborah L. Friedman, 190Ibid. 115CDC 1979–1988, supra note 91, 1492. and Jeffrey H. Coben, “Incidence of Dog Bite 191Ibid. 116CDC 1979–1998, supra note 94, 838. Injuries Treated in Emergency Departments,” 192See infra Part II. 117Ibid. (Fatal attacks are easier to track Journal of the American Medical Association 193See Pinckney and Kennedy, supra note than nonfatal bites because fatal attacks 279 (1998), 51. 183, 195. result in SCMT coding and are more likely to 160CDC Nonfatal, supra note 123. 194William G. Winkler, “Human Deaths be reported by the news media). 161Weiss, supra note 159, 51. Induced by Dog Bites, United States, 1974–75,” 118CDC 1979–1988, supra note 91, 1492. 162Ibid. Public Health Report 92 (1977), 425. 119Ibid., 1491. 163Ibid., 52. 195The CDC authors criticized the Winkler 120Ibid. 164Ibid. study as being “anecdotal” rather than “sys- 121See supra note 101 and accompanying 165Ibid. temic.” CDC 1979–1988, supra note 91, 1489. text. 166Ibid., 53. The conclusions about the 196Winkler, supra note 194. 122CDC 1989–1994, supra note 17, 894. monetary impact of dog bites were confirmed 197Ibid., 428. They estimated that, in 1994, 1.8 percent of by a CDC study in 1999. See Kyran P. Quinlan 198Ibid. the population was bitten by a dog and 0.3 and Jeffrey J. Sacks, “Hospitalization for Dog 199Jason W. Stull and Robert R. Hodge, “An percent of the U.S. population sought some Bite Injuries,” Journal of the American Med- Analysis of Reported Dog Bites: Reporting medical care for a dog bite. ical Association 281 (1999), 232. A sample of Issues and the Impact of Unowned Animals,” 123Centers for Disease Control, “Nonfatal 904 hospitals in seventeen states found that Journal of Environmental Health 62 (2000), 17. Dog Bites-Related Injuries Treated in Hospital 5,991 hospital discharges in 1994 were the 200Ibid., 19. Emergency Departments—United States, result of dog bite injuries. The average length 201Ibid. The authors were truthful when 2001,” Morbidity and Mortality Weekly Report of stay was 3.6 days, costing a total of $164.9 they stated, “[f]or cases in which breed was 52 (July 4, 2003) 605 (hereinafter CDC million in direct care per year. given, pit bulls, German shepherds, and rot- Nonfatal). 167In this respect, it suffers from many of tweilers combined were responsible for over 124Ibid. The NEISS-AIP is operated by the the same flaws as the CDC nonfatal bite study 59 percent of bites each year.” Nevertheless, U.S. Consumer Product Safety Commission. discussed supra in Part II.A.2. as a lawyer, I am concerned by scientists draw- 125Ibid. 168Jeffrey R. Avner and M. Douglas Baker, ing this conclusion. Such statements can be 126Ibid. Deaths were excluded from the “Dog Bites in Urban Children,” Pediatrics 88 taken out of context and used by the insur- study. (1991), 55. ance industry and legislatures to justify breed 127Ibid. 169Ibid., 56. discrimination and BSL when such actions are 128Ibid. 170Ibid. Consequently, the authors con- not supported by statistics. 129Ibid. cluded that in such cases the patient’s accu- 202Lockwood and Rindy, supra note 83, 7 130Ibid. racy of breed identification should be high. (“The genetics of canine aggression are still 131Ibid. Ibid., 57. I disagree with this conclusion. Just poorly understood...”). 132Ibid., 607. because Little Johnnie knows Spot from the 203Ibid. 133Ibid. neighborhood does not necessarily mean that 204Task Force, supra note 5, 1736, (citing 134Ibid. Johnnie knows with any certainty what breed J.C. Wright, “Canine Aggression Toward Peo- 135Ibid. Spot is. ple: Bite Scenarios and Prevention,” Veteri- 136See infra Part II. 171Ibid, 56. nary Clinics of North America: Small Animal 137CDC Nonfatal, supra note 123. 172Ibid., 57. Practice 21 (1991), 299. 138See infra Part II. for a discussion of what 173Yue-Fang Chang et al., “Dog Bite Inci- 205Lockwood and Rindy, supra note 83, 7 I call “just cause” bites. dence in the City of Pittsburgh: A Capture- (discussing the factors that can lead a dog to 139Kenneth A. Gershman, Jeffrey J. Sacks, Recapture Approach,” American Journal of be aggressive). and John C. Wright, “Which Dogs Bite? A Public Health 87 (1997), 1703. 206See infra Part V. for a discussion of dan- Case-Control Study of Risk Factors,” Pedi- 174Ibid., 1703. gerous dogs laws as an alternative to BSL and atrics 93 (1994), 913. Two of the three 175Ibid., 1703–1704. breed discrimination. authors, Gershman and Sacks, were with the 176Ibid., 1704. 207CDC 1979–1998, supra note 94, 838. CDC at the time. 177Ibid. At the time of the study, Sacks and Gilchrist 140Ibid., 913. 178Alan M. Beck and Barbara A. Jones, were with the CDC’s National Center for 141Ibid., 914. “Unreported Dog Bites in Children,” Public Injury and Control, Division of Unintentional 142Ibid. Health Report 100 (1985), 315. Injury Prevention; Sinclair and Lockwood 143Ibid., 913. 179The authors described previous studies were with The HSUS; and Golab was with the 144Ibid. To create a control group, the as relying principally on official reports of American Veterinary Medical Association. authors used the first five digits of a biting- bites. Ibid., 316. This resulted in significant 208Task Force, supra note 5, 1733. dog owner’s phone number and randomized underreporting of bites. 209“Dalmatian Popularity May Spur

48 The State of the Animals IV: 2007 Increase in Bites, HSUS Says,” DVM: The judge found that there is no scientific means, 276Ibid. Newsmagazine of Veterinary Medicine (Jan. by blood, enzyme, or otherwise, to determine 277Ibid., 1733. 1997), 40. whether a dog belongs to a particular breed, 278CDC 1989–1994, supra note 17, 894 210The CDC authors noted, “[c]onsidering regardless of whether ‘breed’ is used in a for- (“Although several breeds appear over-repre- American Kennel Club registration data for mal sense or not.”) Professor Templeton sented in the population of animals involved rottweilers in parallel with fatality data for described one study in which two American in fatal attacks, this representation fluctuates that breed indicates that as the breed has Staffordshire terriers were compared with a over time. Thus, it may be unproductive to soared in popularity, so have rottweiler-related whippet. The profile of one of the terriers view this as a problem that is unique to any deaths.” CDC 1979–1998, supra note 94, more closely matched the whippet than the one breed”); CDC 1995–1996, supra note 93 838–839. other terrier. (“Although some breeds were disproportion- 211Pinckney, supra note 183. 238CDC 1979–1998, supra note 94, 839. ately represented in the fatal attacks 212Id. at 195. 239However, only some dogs have their described in this report, the representation of 213Weiss, supra note 159, 51. “AKC papers.” See supra notes 231–233 and breeds changes over time. As a result, target- 214Supra Part II.A. accompanying text. ing a specific breed may be unproductive; a 215CDC 1979–1998, supra note 94, 838. 240Katharine Dokken, “Dog Bite Statistics: more effective approach may be to target 216Ibid. Bad Logic,” http://www.thedogplace.com/library/ chronically irresponsible dog owners”); CDC 217Task Force, supra note 5, 1733. articles156.htm (accessed June 1, 2004). Nonfatal, supra note 94 (arguing for regula- 218Ibid. 241OVDO, supra note 51; American Dog tion of individual dogs over BSL). 219See supra notes 137–138 and 167 and Owners Association, Inc., 533 N.E.2d at 644. 279CDC Nonfatal, supra note 123; Lock- accompanying text. 242Wright, supra note 228, 301; Task wood and Rindy, supra note 83, 2 (describing 220Chang, supra note 173, 1704. Force, supra note 5, 1733; Gershman, supra BSL as “controversial” and attributing the 221See supra notes 150–158 and accompa- note 139, 916. problem to a lack of “good data” for the nying text. 243CDC 1979—1988, supra note 91, numerator and denominator). 222Ibid. 1492. 280American Kennel Club, “Canine Legisla- 223U.S. Constitution, art. I, § 2, cl. 3. 244Dokken, supra note 240. tion Position Statements,” http://www.akc.org/ 224Department of Commerce v. United 245Lockwood and Rindy, supra note 83, 2. love/dip/legislate/canleg.cfm (accessed July States House of Representatives, 525 U.S. 316, 246“Find the Pitbull,” http://www.pitbulls 2003). 322 (1999). ontheweb.com/petbull/findpit.html (accessed 281American Kennel Club, “American Ken- 225One study estimated that as of 1998 June 1, 2004) (“Breed misidentification is a nel Club Statement on Dangerous Dogs,” http:// approximately 34.3 million homes had dogs scary thing in a time breed specific legislation www.akc.org/love/dip/legislate/dangerous.cf and a total of 53.6 million dogs in the United is growing....Pit [B]ull dogs are often blamed m (accessed June 8, 2004). States being kept as pets. See J. Karl Wise and for dog attacks that may very well have been 282Davis, supra note 58, 1. Jih-Jing Yang, “Dog and Cat Ownership, caused by another breed”). 283Ibid. 1991–1998,” Journal of the American Veteri- 247CDC 1979–1998, supra note 94, 838 284Weiss, supra note 51; Lockwood, supra nary Medical Association 204(8) (1994), (“even experts may disagree on the breed of a note 51, 276. 1166, 1166–67. particular dog”). 285Weiss et al., supra note 159, 51; Task 226Supra Part II. 248533 N.E.2d 642 (Mass. 1989). Force, supra note 5, 1733. In contrast, the 227See, for example, Pinckney, supra note 249Ibid., 644. domestic cat has been with us for only about 183; Winkler, supra note 194. The Winkler and 250CDC 1979–1998, supra note 94, 838 four thousand years. Gina Spadafori and Paul Pinckney/Kennedy studies were also criticized (stating that it is unclear how to count D. Pion, Cats for Dummies, 2d ed. Hoboken, by Sacks et al. for taking a more anecdotal, attacks by mixed-breed dogs). N.J.: For Dummies, 2000, 14. rather than systematic, approach to determin- 251Ibid., 839. 286See infra Part IV. ing the relative dangerousness of breeds. See 252Task Force, supra note 5, 1736. 287Ibid. CDC 1979–1988, supra note 91, 1489. 253Dokken, supra note 240 (“To go by sta- 288Wise and Yang, supra note 225. 228John C. Wright, “Canine Aggression tistics alone assumes that the majority of dog 289Ibid. toward People: Bite Scenarios and Preven- bites are reported and that the majority of 290Willing, supra note 6. tion,” Veterinary Clinics of North America: breeds identified are correct”). 291Ibid. Small Animal Practice 21(2) (1991), 299, 301. 254There is reason to believe that incidents 292Ibid. 229CDC 1979—1998, supra note 94, 839. of “just cause” bites are not uncommon. The 293Task Force, supra note 5, 1739. 230David Brand, “Time Bombs on Legs,” CDC described a series of incidents where 294Ibid., 1733. Time (July 27, 1987), 60. Call me cynical, but nonfatal dog bites that involved “just cause.” 295Ibid., 1739. I cannot envision drug dealers or gang mem- Stated conversely, these are bites where the 296Ibid. bers registering their dogs with the local human “victim” is to blame. CDC Nonfatal, 297Susan D. Semmel, Comment, “When municipality or the AKC. supra note 123. Pigs Fly, They Go First Class: Service Animals 231American Kennel Club, “Register a Dog,” 255Dokken, supra note 240. in the Twenty-First Century,” Barry Law http://www.akc.org/registration/registeradog.cfm 256Task Force, supra note 5, 1742. Review 3 (2002), 39. (accessed July 9, 2004). 257Brand, supra note 230. 298Ibid., 44. 232While there were approximately fifty- 258Ibid. 299CDC 1995–1996, supra note 93, 466 seven million dogs in the United States in 259See, for example, ibid. (“Dogs provide many health and social benefits”). 1990, only twelve million were registered with 260Ibid. 300David T. Allen, “Effects of Dogs on the AKC. Mark Derr, “The Politics of Dogs,” 261Ibid. Human Health,” Journal of the American Vet- Atlantic Monthly (March 1990), 49, 50. 262Ibid. erinary Medical Association 210(8) (1997), 233See ibid. 263Brand, supra note 230. 1136, 1138. 234Lockwood & Rindy, supra note 83, 3. 264Lockwood and Rindy, supra note 83, 3. 301Ibid. 235Ibid. 265Ibid. 302Ibid. 236Derr, supra note 232, 51. 266OVDO, supra note 51. 303Task Force, supra note 5, 1740. 237Ibid., 52 (discussing interview with Joe 267Brand, supra note 230, 60. 304See, for example, “In Memory of Pets,” W. Templeton, professor of veterinary pathol- 268Ibid. http://www.in-memory-of-pets.com/ (accessed ogy and genetics at Texas A&M University, in 269Ibid. July 21, 2004); Association for Pet Loss and which Professor Templeton stated that scien- 270Task Force, supra note 5, 1736. Bereavement, http://www.aplb.org/ (accessed tists cannot distinguish between breeds by 271Ibid., 1732. July 21, 2004). using genetic fingerprinting or examining 272See supra Part II. 305Willing, supra note 6. chromosomes); American Dog Owners Asso- 273Task Force, supra note 5, 1736. 306Wayne Hunthausen, “Effects of Aggres- ciation, Inc. v. City of Lynn, 533 N.E.2d 642, 274Ibid. sive Behavior on Canine Welfare,” Journal of 644 (Mass. 1989) (“After trial, the [trial] 275Ibid. the American Veterinary Medical Association

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 49 (1997), 1134. 354Ibid. One court has stated, “Such regulation, 307Ibid. 355Ibid. even down to ‘the minutest particular in the 308Ibid., 1135. 356Ibid., 371. interest of the public,’ has never been ques- 309Ibid. 357Ibid., 373 (J. Andell concurring). tioned.” Blue Cross & Blue Shield, 674 N.E.2d, 310Ibid. 358Ibid. 1126 (quoting People v. Formosa, 30 N.E. 492 311Sodergren, supra note 59; McMahon, 359Ibid., 376. [N.Y. 1892]). supra note 17, 03a. 360Ibid., 377. 370Greer v. Aetna Life Insurance, 142 So. 312Sodergren, supra note 59. 361Ibid., 377–378. 393, 395 (Ala. 1932) (“The generally recog- 313Ibid. 362Willing, supra note 290. nized rule is that ‘in the absence of statutory 314See infra Part V. 363Blue Cross & Blue Shield of Central New prohibitions discriminations or rebating as to 315Robert R. Googins, “Fraud and the York, Inc. v. McCall, 674 N.E.2d 1124, 1126 premiums is not illegal’”). In Greer, the state Incontestable Clause: A Modest Proposal for (N.Y. 1996) (hereinafter Blue Cross & Blue complained that Aetna’s policy of not charg- Change,” Connecticut Insurance Law Journal Shield). For further discussion of insurance as ing higher premiums for mortgage insurance 2 51 (1996), 69 n. 86. a “privilege” instead of a “right,” see 43, for older people (who pose a greater risk) in 3162 Couch on Insurance § 31:10 (3d ed. American Jurisprudence 2d ed., Title Insur- essence discriminated against the young. 2004). ance (2004) § 23. “The organization of an Ibid., 394–395. Since state law did not pro- 317Sentell v. New Orleans & C.R. Co., 166 insurance company and the conduct of the hibit discrimination based on age or life U.S. 693 (1897). business of writing insurance is not a right expectancy, the Alabama Supreme Court held 318Ibid. but a privilege granted by the state, subject to that the insurance company did not do any- 319Ibid. the conditions imposed by it.” Ibid. § 24. thing illegal. 320Ibid. Regulation of the insurance industry is pri- 371Blue Cross & Blue Shield, 674 N.E.2d, 321Ibid., 694. marily a power of states, not the federal gov- 1126 (stating that the conduct of insurance 322Ibid., 696. ernment. 43 American Jurisprudence 2d ed., companies is subject to conditions imposed 323Ibid., 694. (2004) §§ 29, 30. The McCarran-Ferguson by the state to promote public welfare). 324Ibid. The Court noted, however, the Act, 15 U.S.C. §§ 1011–1015, declared that 37244 C.J.S. Insurance § 45 (1993). dogs can be considered property for purposes regulation of insurers by the states was in the 373Ibid. § 66 (“The rates charged by insur- of the tort of conversion. public interest. 43 American Jurisprudence ance companies must not deviate from those 325Ibid. 2d ed., (2004) § 30. McCarran-Ferguson cre- established by state authority. Rates may not 326Ibid. ates a reverse preemption system where the be unreasonable, excessive, inadequate, or 327Ibid. power to regulate is vested in the states, not discriminatory”). In states where rates are 328Ibid., 694–695. the federal government. Ibid. However, plans regulated, insurers bear the burden of proving 329Ibid. The Court’s distinction among created under the Employee Retirement the reasonableness of their rates. Ibid. § full, incomplete, and no status is interesting, Income Security Act (ERISA) are still subject 69(b). Legislatures can prohibit discrimina- but ultimately irrelevant, as the Court held to joint federal-state regulation. John Han- tion against insureds of the same “class.” that even if dogs were considered complete cock Mutual Life Insurance Co. v. Harris Trust Ibid. § 64. Some states provide a private cause property, they (like all forms of property) are & Savings Bank, 510 U.S. 86 (1993) (“We are of action to enable insureds to recover for subject to the police power of the state. satisfied that Congress did not order the unfair trade practices or violations of rate reg- 330Ibid., 696 unqualified deferral to state law that Hancock ulations. Ibid. § 45. 331Bennett v. Bennett, 655 So. 2d 109 (Fla. both advocates and attributes to the federal 37443 American Jurisprudence 2d ed., § 43 Dist. Ct. App. 1995). lawmakers. Instead, we hold, ERISA leaves (2003). 332Arrington v. Arrington, 613 S.W.2d 565 room for complementary or dual federal and 375Ibid. § 25. Most states require insurers (Tex. App. 1981). state regulation, and calls for federal to contribute to an “assigned risk plan” that 333Bennett, 655 So. 2d at 110. supremacy when the two regimes cannot be writes “high-risk” policies for drivers who can- 334Ibid. harmonized or accommodated”). not get insurance in the free market, often 335Ibid. 364Blue Cross & Blue Shield, 674 N.E.2d, because of DWI convictions. See, for example, 336Ibid. 1126. Texas Insurance Code Annotated art. 21.81 337Arrington, 613 S.W.2d at 569. 36543 American Jurisprudence 2d ed., (2004–2005). 338Ibid. In Texas family law a “managing (2003) § 25. 376Deborah S. Hellman, “Is Actuarially Fair conservator has the right to establish the 366Bekken v. Equitable Life Assurance Soci- Insurance Pricing Actually Fair?: A Case Study child’s residence and has primary custody of ety, 293 N.W. 200, 211 (N.D. 1940); 43 Amer- in Insuring Battered Women,” 32 Harvard the child. A possessory conservator typically ican Jurisprudence 2d ed., (2003) §§ 24, 25. Civil Rights and Civil Liberties Law Review has visitation rights under terms and condi- 367Bekken, 293 N.W., 209. (1997), 355. tions set by the court.” In re V.L.K., 24 S.W.3d 368Ibid., 210. 377Ibid., 410. 338, 340 n.1 (Tex. 2000) (citations omitted); 369Ibid., 211. (“It will be noted that the 378Ibid., 356, 377–378. see generally Texas Family Code Annual State, under its police power, in the interest 379See ibid., 358. (2004) §§ 153.132, 153.371, 153.192. of the general welfare, may regulate not only 380Deborah A. Stone, “The Rhetoric of 339Arrington, 613 S.W.2d at 569. the business of insurance in a general way, but Insurance Law: The Debate over AIDS Testing,” 340Ibid. it may, also, regulate contracts of insurance, Law and Social Inquiry 15 (1990), 385, 388. 341Raymond v. Lachmann, 264 A.D.2d and in a large measure prescribe the terms 381If a person cannot obtain health insur- 340 (N.Y. App. Div. 1999). and conditions of such contracts, and it may ance, his or her health care costs shift to the 342Ibid. impose duties and obligations incident to the public (through Medicare or Medicaid) or to 343Ibid. relation created by the contract or the nego- hospitals, which may be required to treat 344Ibid., 341. tiations for a contract different from those every patient regardless of the ability to pay. 345Corso v. Crawford Dog & Cat Hospital, arising or existing under other contracts, and In the latter case, the costs are likewise Inc., 97 Misc.2d 530 (NY. Civ. Ct., Queens it may prohibit the parties from altering such shifted to the public because hospitals must County 1979). duties or obligations.”). In Bekken, the North increase costs for paying customers (and their 346Ibid. Dakota Supreme Court had to decide whether insurers) in order to make up for their losses 347Ibid. a beneficiary of a life insurance policy could from their nonpaying customers. Ibid., 390. 348Ibid., 530–31. enforce the policy even though the insurance 382In the case of breed discrimination, the 349Ibid., 531. company had not acted promptly in writing latter results. If I had not been able to secure 350Ibid. the policy before the deceased died. The insurance with the Farm Bureau, I would have 351Bueckner v. Hamel, 886 S.W.2d 368, Court held that the insurance company had lost out on a significant economic opportu- 370 (Tex. App. 1994). breached its duty to act promptly on the nity—the ability to purchase a house. See 352Ibid., 370. deceased’s application and ordered the com- infra Part V.B. for a discussion of insurance as 353Ibid. pany to pay the beneficiary. Ibid., 218. gatekeeper to homeownership. In the alterna-

50 The State of the Animals IV: 2007 tive, I could have given up my dogs, which pools altogether, unless administrators pro- insureds. They are people likely to need addi- likewise would have been a significant loss of vide financial incentives for them to remain in tional care in the future. The market happiness and opportunity. the pools. Ibid., 375–376. responded by charging higher rates for people 383See Stone, supra note 380, 386 (stating 399See Hellman, supra note 376, 380. with preexisting medical conditions or flat- that insurance is the primary mode people 400See Ibid., 380–381. out rejecting them for coverage. The result provide for needs not affordable through nor- 401Ibid. was that many people found themselves with- mal work income). 402Ibid., 380 (noting that “statistical dis- out insurance when they switched insurers. 384Brian J. Glenn, “The Shifting Rhetoric crimination is justified when the social utility, See Grey, supra note 17, 421 (insurers are of Insurance Denial,” Law and Society Review including the cost to the insurer of employing given wide latitude to refuse to cover certain 34 (2000), 779, 782. a particular classification, could not be high-risk customers, absent state legislation 385Lee R. Russ and Thomas F. Segalla, increased by further refinement of the classifi- or regulation to the contrary). Some states Couch on Insurance, 3d ed. (1995) § 2, 1. cation,” a view expressed by Steven Meitzen in responded by passing legislation prohibiting 386Sonia M. Suter, “Disentangling Privacy “The Ethics of Statistical Discrimination,” insurers from refusing to provide coverage From Property: Toward a Deeper Understand- Social Theory and Practice 17 [1991], 23, 26). based on preexisting medical conditions. Con- ing of Genetic Privacy,” George Washington 403Frederick Schauer argues that BSL and necticut requires health insurers to provide Law Review 72 (2004), 737, 795 (“All states breed discrimination are justified practices. coverage for preexisting conditions if the per- require underwriting decisions to be actuari- Frederick F. Schauer, Profiles, Probabilities, son was previously covered by his or her pre- ally, or rationally, based; they cannot be arbi- and Stereotypes (Cambridge, Mass.: Harvard ceding plan. Connecticut General Statutes trary. Insurers must engage in good-faith University Press, 2003), 55–78. Schauer starts Annotated §§ 38a-476 (West 2004). practices in deciding whether to underwrite, with the premise that there is a statistical cor- 408Legislatures have acted to protect con- at what rate, and for what conditions.”); 43 relation between pit bulls and unprovoked sumers from discrimination on the basis of American Jurisprudence 2d ed., § 43 (2003) bites. Ibid., 55 (“Looking at what evidence we race, color, national origin, gender, and other (underwriting and rate setting may take into have, it turns out that the generalizations protected classes. Maryland, for example, has account only legitimate cost factors). To some underlying pit bull restrictions do indeed have a relatively far-reaching statute prohibiting extent, this requirement is based on a prohi- the kind of empirical support that distin- discrimination in underwriting based on race, bition of unfair trade practices, as enumer- guishes them from purely spurious generaliza- color, creed, sex, blindness, or for any arbi- ated in the model Unfair Trade Practices Act, tions”). He then goes on to argue that BSL trary or capricious reason. Maryland Code developed by the National Association of and breed discrimination, while both underin- Annotated, Insurance § 27-501 (2004). Insurance Commissioners. See Unfair Trade clusive and overbroad, are justifiable from the States have enjoined insurers from charg- Practices Act (National Association of Insur- perspective of statistical correlation and the ing higher automobile insurance rates to men ance Commissioners, 2001) (hereinafter need to protect people from dangerous even though actuarial statistics show that UTPA). The NAIC, a nongovernmental organi- events. Ibid., 59. As I have demonstrated, how- women are a lower risk. For example, in Hart- zation of insurance regulators, drafted this ever, the correlation between breed and dan- ford Accident and Indemnity Company v. model act to regulate unfair trade practices in gerousness is weak at best. See supra Part II. Insurance Commissioner, 482 A.2d 542 (Pa. the insurance industry. Most states have Schauer’s citation for his proposition is sev- 1984), the Pennsylvania Supreme Court held adopted UTPA in some form or another. See eral of the CDC studies that have been dis- that the practice of charging different rates Stone, supra note 380, 392; Richard J. Wirth, cussed in this article. Schauer, supra, at 59 based on gender violated the state’s statute, “My Customer’s Keeper: The Search for a Uni- n.7; see supra Part II. Yet, these very studies which prohibited “unfairly discriminatory” versal Suitability Standard in the Sale of Life cautioned that their data were incomplete rates. Ibid., 549. The Court looked to the pur- Insurance,” New England Law Review 24 and that it would not be statistically sound to poses of the act and insurance regulation in (2002), 47, 79. make generalizations about breeds. See supra general: to promote the public welfare. Ibid., 387Maryland Code Annotated - Insurance Parts II & III. 547. The Court held, “[P]ublic policy consid- (2000)§ 27-501(a)(1) (emphasis added). 404Task Force, supra note 5, 1737. erations require more adequate justification 388Ibid. § 27-501(a)(2). 405Davis, supra note 58, 36 (quoting Ale- for rating factors than simple statistical corre- 389Tom Baker, “Containing the Promise of jandra Soto, spokesperson for the III). lation with loss.” Ibid., 548 (quoting National Insurance: Adverse Selection and Risk Classi- 406The practice of “drive-through deliver- Association of Insurance Commissioners, fication,” Connecticut Insurance Law Journal ies” required doctors to discharge new moth- Report of the Rates and Rating Procedures 9 (2002/2003), 371, 377. ers within hours after giving birth. Vicki Task Force of the Automobile Insurance (D3) 390Otherwise insurers would run a risk of Lawrence MacDougall, “Medical Gender Bias Subcommittee, November 1978, 5–6 [foot- low-risk consumers being priced out of the and Managed Care,” Oklahoma City Univer- notes omitted]). market and only high-risk insureds remaining, sity Law Review 27 (2002), 781, 882, A Michigan court declared unlawful the a danger known as the “death spiral.” See 892–894. States responded by passing laws practice of refusing to write automobile insur- Peter Siegelman, “Adverse Selection in Insur- requiring insurance companies to allow new ance for adults under twenty-one unless they ance Markets: An Exaggerated Threat,” Yale mothers to stay in the hospital for a minimum resided with parents. Detroit Automobile Inter- Law Journal 113 (2004), 1223, 1224. period of time. Elizabeth C. Price, “The Evolu- Insurance Exchange v. Commissioner of 391Baker, supra note 389, 373. tion of Health Care Decision-Making: The Insurance, 326 N.W.2d 444, 445 (Mich. Ct. 392Kenneth S. Abraham, Distributing Risk: Political Paradigm and Beyond,” Tennessee App. 1982). The Court found that this prac- Insurance, Legal Theory, and Public Policy Law Review 65 (1998), 619, 626. Connecti- tice violated the state’s unfair trade practices (New Haven, Conn.: Yale University Press, cut, for example, passed a law that required law, which permitted refusing to insure a 1986), 67–68. insurance companies to pay for forty-eight group “only if the cost is unreasonable.” Ibid., 393Baker, supra note 389, at 377. hours of inpatient care following vaginal deliv- 447. The Court affirmed a central notion of 394See, for example, 43 American Jurispru- eries and ninety-six hours following caesarian insurance law that the free market is not the dence 2d ed., (2003) § 43. deliveries. Connecticut General Statutes Anno- sole determinant of insurance rates. The 395Hellman, supra note 376, 380. tated §§ 38a-503c, 38a-530c (West 2004). court stated, “The mere fact that one group is 396Ibid., 378. Congress, using its power to regulate ERISA- more expensive to insure than another does 397Moral hazard is the change in incen- based insurance policies, followed suit by not preclude fixing a reasonable rate.” tives that can result because an individual is passing the Newborns’ and Mothers’ Health Plaintiffs have also relied on federal law for protected by an insurance contract. Baker, Protection Act of 1996, which included the relief against discrimination in the provision supra note 389, 373. Insureds without a suffi- 48-/96-hour rule that had been adopted by of homeowners’ insurance. In National Fair cient personal investment in a particular several states. Newborns’ and Mothers’ Health Housing Alliance, Inc. v. Prudential Insurance risk—through co-insurance, for example— Protection Act of 1996, Pub. L. No. 104-204, Company of America, 208 F. Supp. 2d 46 may be more likely to engage in the risky con- 110 Stat. 2874, 2935 (codified at 29 U.S.C. §§ (D.D.C. 2002), a U.S. district judge held that duct in a negligent manner. 1185[a][1][A][i]-[ii][2004]). discrimination by a homeowners’ insurance 398Adverse selection is the tendency for 407Customers with preexisting medical company on the basis of race, color, religion, low-risk insureds to drop out of insurance conditions represent a high-risk class of sex, handicap, familial status, or national ori-

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 51 gin was made illegal by the Fair Housing Act laborative endeavors among animal groups, gerous occurrences from happening in the and implementing regulations from HUD. governments, and insurers. At least one first place. Schauer’s analysis, however, fails Ibid., 55–56. insurer, State Farm, has publicly stated its to recognize that dogs are different. Speed 409At least one court has stated, as a mat- willingness to find proactive solutions to the limits come at a small social cost—drivers ter of public policy, that insurers may not problem of dog bites. See Hattaway, supra who can drive safely at faster speed limits are engage in witness intimidation. In L’Orange v. note 70. forced to get to their destinations later than Medical Protective Company, 394 F.2d 57 (6th 430See, for example, Sacks et al., supra they would have had the speed limits not Cir. 1968), an insurance company cancelled a note 150, 53; CDC 1979–1998, supra note existed. This is a small social cost that comes dentist’s malpractice policy after he testified 94, 840; Winkler, supra note 194, 425. with a more significant, larger societal benefit against another dentist who was insured by 431Some scientists have suggested that (saving many lives). See Philip Shuchman, “It the same company. Ibid., 59. The court some breeds are more dangerous because, Isn’t That the Tort Lawyers Are So Right, It’s acknowledged that insurance policies are when they do bite, their jaw structure and Just That the Tort Reformers Are So Wrong,” treated as voluntary contracts, but noted that other physical characteristics cause them to Rutgers Law Review 49 (1997), 485, 523 (20 they are also subject to public policy con- inflict more pain and physical injury. See percent increase in highway fatalities after cerns. Ibid. In this diversity-of-citizenship Lockwood and Rindy, supra note 83, 36. states given permission by the federal govern- case, the Court looked to the law of Ohio and 432Brand, supra note 230, 60. ment to raise speed limits to 65). Similarly, an found that the defendant-insurer had violated 433Sacks et al., supra note 150, 53 (sup- unlicensed or untrained person practicing law Ohio’s public policy against witness intimida- porting enforcement of existing laws to regu- or medicine has an almost 100 percent cer- tion. The Court noted the need for expert tes- late dangerous dogs and dog fighting); CDC tainty of causing damage to clients or timony, the existence of statutes against 1979–1998, supra note 94, 840 (urging law- patients. The social cost of not regulating intimidating witnesses from testifying, and makers to focus on problem owners, not dogs those practices would be significantly high. the potential chilling effect of the defendant’s or breeds); CDC 1989–1994, supra note 17, Although the dog population is around fifty behavior, and awarded judgment to the den- 894 (same). million, see supra notes 288–289 and accom- tist. Ibid., 61–63. 434American Kennel Club, “American Ken- panying text, only a handful of those dogs will 410Of course, if a person can purchase the nel Club Statement on Dangerous Dogs,” bite a person. Fatalities for dog bites have hov- house through cash on hand, then securing a http:// www.akc.org/love/dip/legislate/dan- ered around seven per hundred million people mortgage is unnecessary and obtaining home- gerous.cfm (accessed Nov. 6, 2004). per year. See supra Part II.A.1. Yet, the social owners’ insurance is “optional.” A prudent 435CDC 1979–1998, supra note 94, 840. cost of having forward-looking regulations— homebuyer would nevertheless purchase 436See CDC 1989–1994, supra note 17, such as BSL and breed discrimination— homeowners’ insurance to protect his or her 894; CDC 1979–1998, supra note 94, 840. comes at a very high cost to families that own investment in the event of catastrophic loss. 437CDC 1979–1998, supra note 94, 840 dogs, particularly those that are seeking to 411978 F.2d 287 (7th Cir. 1992). (“[P]roblem behaviors [of dogs and owners] purchase a home. See supra Part IV. Further, 412Ibid., 297. have preceded attacks in a great many cases individualized, dangerous-dog prosecutions 413Ibid., 290. and should be sufficient evidence for preemp- do have a prophylactic effect. Like tort law, 414Ibid, 298. tive action”). dangerous-dog laws indirectly encourage own- 415Ibid., 297–298. 438I would exclude from this category “just ers to take reasonable steps to prevent 416Ibid., 290–291. cause” bites, discussed supra at Part II.F. injuries. In this respect, dangerous-dog laws 417David H. Harris, Jr., “Using the Law to 439CDC 1979–1998, supra note 94, 840. can serve as a deterrence against negligent or Break Discriminatory Barriers to Fair Lending 440Michigan Compiled Laws §§ 287.321, intentional misdeeds. for Home Ownership,” North Carolina Central 287.322 (2004). 449See Benjamin L. Hart and Lynette A. Law Journal 22 (1996), 101. 441Ibid. § 287.322. Hart, “Selecting, Raising, and Caring for Dogs 418Ibid. 442Ibid. § 287.321(a). The law provides a to Avoid Problem Aggression,” Journal of the 419Ibid. number of exceptions to the definition of dan- American Veterinary Medical Association 210 420Unless, of course, he is independently gerous. An animal is not dangerous if: (1) the (1997),1129, 1131. wealthy and does not need a mortgage. “victim” was a trespasser or provoked or tor- 450Ibid. (neutering can reduce aggression). 421Kansas physicians are required to carry mented the dog; or (2) the animal was pro- 451See Gershman et al., supra note 139, malpractice insurance. Kansas Statute Anno- tecting a person or livestock. 914 (finding a statistically significant relation- tated §§ 40-3402 (2000); State ex rel. Schnei- 443Ibid. § 287.321. ship between number of bites and intact dogs). der v. Liggett, 557 P.2d 221 (Kan. 1978) 444Oklahoma Statutes title 4, §§ 44, 47 452Hart and Hart, supra note 449, 1130. (upholding Kansas’s mandatory, state-run mal- (2004). 453Ibid. practice insurance program). Oregon requires 445Ibid. § 46(b). 454Ibid. An aggressive dog, for example, attorneys to carry malpractice insurance. Ore- 446CDC 1995-1996, supra note 93, 466. might do well with an assertive family. gon Revised Statutes § 9.080(2)(a) (2003). 447See CDC 1989–1994, supra note 17, 455This conclusion has been supported by 422See, for example, New York Insurance 894-8995 (“[I]t is important to recognize several scientists. CDC 1989–1994, supra Law § 5303 (McKinney 2003) (New York’s that most of the 52 million dogs in this coun- note 17, 894 (calling for education of bite vic- assigned risk plan). try never bite or kill anyone”). tims and children); Sacks et al., supra note 423Hospitals, Medicare and Medicaid, and 448Frederick Schauer argues that a system 150, 53 (education programs needed on dog health insurers would thus bear the cost for of individualized determination of dangerous- behavior); CDC 1979–1998, supra note 94, treatment. See Stone, supra note 380, 394 ness—that myself and others propose—is 840 (education needed for children); Gersh- (making a similar argument regarding unsound. Schauer, supra note 403, 69–72. He man, supra note 139, 916 (suggesting educa- human diseases). states that such a system comes at a high tion programs for children as a method to 424See ibid., 392. social cost, since the State responds only after reduce bites and other attacks); CDC 425Baker, supra note 389, 376–378. a dangerous dog attacks. He cites several 1995–1996, supra note 93, 466 (discussing 426See Lockwood and Rindy, supra note 83. examples, including the speed limit on high- public education as a strategy towards reduc- 427In this respect, the use of the term “dis- ways, to show that BSL and breed discrimina- ing bites); Task Force, supra note 5, 1739 crimination” is a bit of a misnomer. All forms tion are simply forms of acceptable, forward- (education is key); Hunthausen, supra note of risk classification are acts of discrimination looking regulation. While some drivers might 306, 1135 (public education a necessary com- in a literal sense. See Hellman, supra note be better than others, the State has set a max- ponent of bite prevention). Education is also 376, 378 (“Because all insurer classifications imum speed limit regardless of driver ability. supported by the AKC. American Kennel Club, are ‘discrimination,’ understood non-pejora- This is a forward-looking or prophylactic “American Kennel Club Statement on Danger- tively, one must ask why use of this classifica- attempt to prevent accidents, death, and ous Dogs,” http://www.akc.org/love/dip/ tion is ‘plain, old fashioned,’ ‘profoundly injury before they happen. He also points to legislate/dangerous.cfm (accessed June 8, unjust[,] and wrong[ful]’ discrimination”). the regulation of doctors and lawyers as an 2004). Other groups support heightened 428Baker, supra note 389, 377. example of how society legitimately controls efforts to teach responsible dog ownership 429These are alternatives that could be col- behavior in advance in order to prevent dan- and dog safety. OVDO, supra note 51. Educa-

52 The State of the Animals IV: 2007 tion is also supported by the insurance indus- infectious, compared to 20 percent of dog try. Hattaway, supra note 70, 1144 (“I believe bites; cat teeth can penetrate more deeply that the insurance industry has a role in pro- and transmit bacteria more easily). In spite of moting responsible pet ownership, including these data, however, the insurance industry education, to help reduce this national prob- has not tried to outlaw cats as pets. Why not? lem”); III, Dog Bite Liability, supra note 66 464Beaver, supra note 461, 1148. (recommending that homeowners educate 465Ibid. their children not to approach a sleeping or 466Pennsylvania Consolidated Statutes eating dog). Annotated 3 § 459-507-A(d) (West 2004). 456Simon Chapman, et al., “Preventing 467A.B. 6761, 2003-2004 Sess. (N.Y. 2003). Dog Bites in Children: Randomised Con- 468H.B. 174, 2003 Sess. (N.H. 2003). trolled Trial of an Educational Intervention,” British Medical Journal 320 (2000), 1512–1513. Children are more likely than adults to be the victims of dog bites. CDC 1979–1988, supra note 91, 1490 (70 percent of fatal dog bites were in children under ten years old); Sacks and Kresnow, supra note 150, 52–53 (children account for approxi- mately half of all people who seek medical attention for dog bites); CDC 1979–1998, supra note 94, 836 (most victims of fatal dog bites are children); CDC 1995–1996, supra note 93, 463 (80 percent of fatality victims were children); CDC Nonfatal, supra note 123 (children between five and nine years old are most likely to be victims of nonfatal dog bites). Some have speculated that children are more likely to be victims than adults because of their small stature and inability to defend themselves. See CDC 1979–1988, supra note 91, 1492 (young and old are most at risk to be victims 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-Canine Interactions: A Summary of Perspectives,” Journal of the American Veterinary Medical Association 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 rate of infection associated with them. See Los Angeles County Animal Care & Control, “Cat Bites,” http://animalcontrol.co.la.ca.us/html/ pages/poetownerinfo/Catbite.htm (accessed June 9, 2004) (one million cat bites are reported each year in the United States; cat bites can be especially dangerous for children, the elderly, or those with suppressed immune systems); Cynthia B. Whitney, “Ouch!—More Than You Ever Wanted to Know About Cat Bites,” http://www.thecatsite.com/cat_snips/ snips.php?a=bites (accessed June 8, 2004) (reporting that 80 percent of cat bites get infected and that one out of 170 people will be 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 who had a “brush with death” after being bitten by her cat; she required hospitalization for a week); Sound Medicine, “Dog versus cat bites,” http://www.soundmedicine.iu.edu/ archive/2002/quiz/animalBites.html (accessed July 27, 2002) (50 percent of cat bites are

The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies 53