Monday, 23 January 2012


...and the availability heuristic kicks in again. Dogs have random draw chance of mauling a kid, with some breeds having somewhat higher chance than others and some owners having much higher chances than others. When a couple of kids get bitten within a short period of time, debate on dog regulation heats up.
Local Government Minister Nick Smith promised yesterday to kick-start a stalled pledge to investigate laws governing dangerous dogs.
The inquiry was supposed to take place last year.
Former local government minister Rodney Hide had pledged to reform dog control laws before the November election but that review had been delayed following the Christchurch earthquakes.
The new minister, Smith, said last night he was concerned about the seriousness of the latest attacks.
His office would investigate the incidents and identify whether there were any issues that had implications on dog control legislation in relation to public safety
Were I dog czar, and if it were possible to enforce dog-owner registration links, I'd probably push towards a liability and insurance regime. Get rid of current breed-specific regulations but mandate that dog owners get liability insurance. An old lady with a corgi or normal person with a golden lab would likely pay $2/year; a skinhead with a pitbull would likely find dog ownership beyond his means and might be invited to buy a menacing-looking breed of cat. Owners without insurance and whose dogs bit kids would be bankrupted and put into the workhouse to pay off their debts.

That's likely outside of the set of the politically feasible. And, if linking dogs to owners is too difficult, it's also impracticable. But I like it otherwise. A decent proportion of dog-bite events are just random-draw bad luck, but a similarly decent proportion likely comes from "redneck owning a mean-looking dog because he likes scaring his neighbours and looking tough." Insurers would have reasonable incentive to get the premiums right. Keep it with strict liability regime; the only fights in the courts would be determining who owned the dog. Some of that could be helped by police routinely asking folks walking pitbulls to show registration and insurance certificates.

Meanwhile, a few simple rules for reducing your kid's risk of being bitten - almost certainly redundant for anybody reading this blog:

  • If you're already a redneck, don't get a pitbull-cross dog;
  • Don't associate with rednecks with pitbull-cross dogs;
    • Most rednecks are good people. But a pitbull is usually evidence that the owner is trying to induce a separating equilibrium; take the hint.
  • Let your kids pat friendly-looking dogs whose middle-class owners are nearby and give permission;
  • Give rednecks with pitbulls a wide berth. And, when the owner's well out of earshot, explain to your kids why you've done so, so they know what to do in your absence.
I wonder if anybody's ever done a New Zealand edition of this one...


  1. My personal feeling is that the likelihood of dog attacks are about 75% influenced by the dog owner, 15% by the victim, and only 10% influenced by the breed of dog. Owners bear the weight of responsibility to ensure that their dogs are properly housed and restrained so that they can't bite people unless they are unlawfully trespassing on the dog's property. Similarly, parents bear some of the responsibility - kids need to be educated about when and how to approach dogs. The breed of dog will influence its tendency towards aggressiveness, but I tend towards blaming nurture rather than nature. My dog is a complete softy, but I'm sure in the wrong owner's hands even she could be turned into a biting angry monster, you just need to abuse your dog enough so that it hates people.

  2. I share most of Lat's appreciation. I was just thinking what are the chances of a redneck looking to get a pitbull-cross stumbling on this blog...

  3. @Lats: And that's why breed-specific regulation is worse than insurance premiums that could consider owner-breed interaction effects. And put in a strict liability regime.

  4. I'm going to take this post as an excuse to paste my favourite Malcolm Gladwell sentence:

    "If you look, in fact, at emergency room statistics, you'll see that more people are admitted every year for non-dog bites than dog-bites—which is to say that when you see a Pit Bull, you should worry as much about being bitten by the person holding the leash than the dog on the other end."


    1. Exactly. Insurance premiums would generally be low; pressures for further regulation would be low as rare instances of attack would be met with full compensation; high premiums would only obtain where there were decent reason. It's the owner-dog interaction effect that's critical; I'd taken it as read that I was at least as nervous about the owner as the dog - the dog is often the signal.

      And I'd note too that low bite rates are at least partially a function of folks following the very simple rules noted above....

    2. (Cough, cough. I was trying to point out that Mr. Gladwell doesn't understand the difference between frequencies and rates.)

    3. Oh, I took that as read. But I also often worry as much about the pitbull's owner as about the dog.

  5. We used to have the same hand-wringing over Rottweilers. Before they were popular Dobermans and German Shepherds were the dangerous dogs. Now the focus is on Pit-bulls. Interestingly there was a story on Stuff last year which itemised the number of dog bite incidents by breed. From memory the number one offender in NZ was the humble Labrador, mostly I assume because of the number of them out there. Nobody sane is calling for them to be placed on the banned list of dangerous dogs, partly I guess because they are quite cute, but mainly because they tend to be owned by nice white middle class folk. If rednecks and gang members started training Labs to be attack dogs how long would it be before the do-gooders started calling for them to be banned?

  6. I think most people are a bit unfair to the pit bulls. Pit Bulls are a group of breeds, by natural temperament most of them are actually very friendly dogs. The problem with Pit Bulls is that most of them where breed as fighting dogs so if someone overcomes their natural friendliness they are significantly more dangerous than naturally unfriendly dogs (like a lot of small terriers). In general I've found that smaller dogs are the most aggressive, they are just less likely to cause an injury.

    This of course illustrates the point that looking at owners would be much smarter, but Redneck is over generalized if you ask me. I'm not sure how I'd describe actual problem owners, but I'd know them if I saw them. I guess the reason for this is that I've grow up around "rednecks", like you said above most of them a good folks, but that doesn't change because they own a pit bull. As long as they aren't being used for dog fighting, they seem to be generally good dogs.

    Most neighborhoods I've lived in seem to have this one super-friendly English Staffordshire Terrier (often considered a type of pit bull) that is never locked up and will rush at anyone who walks into their street, wagging their tale, and for all appearances thinking "New friend, new friend, new friend; play with me, play with me!!!" From my own experience and from what I've read the biggest danger from a well treated pit bull (assuming you aren't trespassing and/or threatening it's owner) is that it will get over-excited meeting a new small child and knock them on their back.

    I guess that's the way I'd probably go. If I see a pit bull that looks unhappy or mistreated I'll give it a wide berth, but if I see a happy pit bull that looks like it's been well treated I'll generally assume it's harmless.

    To summarize; I like pit bulls (particularly the English Staffy), and have an immediate knee-jerk reaction to defend them whenever someone says something mean about them. ;-)

    1. I'm sure that there are very nice Staffies out there. And I agree "redneck" isn't the right term. The usual NZ term is bogun, but I'm not sure that's right either. We need an updated NZ version of the book linked in the initial post.

      Agree that owning that breed doesn't change somebody. But your best guess about the individual changes with the dog they've bought. Consider two people who are otherwise observationally equivalent to me, the guy walking down the street. The first has gang connections and enjoys violence; the second enjoys dressing in Holden-branded stuff but is otherwise harmless. If the first person is more likely to buy a pitbull because he knows other people find them menacing, then I learn something about the likelihood of the person's type by the observation of the pitbull.

      This is just conditional probabilities, right?

      Suppose that among the set of people who are observationally equivalent, 10% have a propensity for violence and like intimidating people and 90% don't. Among those who enjoy violence and intimidating people, 80% prefer pitbulls and 20% prefer other dogs; among the 90% who don't enjoy violence, 10% like pitbulls and 90% like other dogs.

      Imagine a population of 100 people who are observationally equivalent but drawn from the distribution above. 90 normal people own 9 pitbulls and 81 other dogs; 10 scary people own 8 pitbulls and 2 other dogs. Observing a pitbull raises my estimate of the person's being scary from 10% to almost 50%. Because almost half of the pitbulls are owned by scary people. Because people who like to scare people disproportionately buy pitbulls.

      None of this requires that the breed be inherently more dangerous; it just requires that scary people like buying dogs that scare people. It's self-fulfilling pretty quickly.

  7. Eric,

    A conviction for failing to control a dog causing injury can carry a three-year prison term and a $20,000 fine.

    The guy whose dog killed his aunt got 18 months, and that term was down from 27 months because he plead guilty as soon as he could, including on TV the day after the attack.

    many owners of large dogs are judgment proof for civil litigation because of a lack of assets and income.

    1. The maximum reduction for an early guilty plea is 25%. There would have been another reduction for some other mitigating factor relating to the defendant. A common one is lack of previous convictions.

  8. p.s. I should add that the guy whose dog killed his aunt regarded her as his second mother, and he did not contribute to the attack in any way. he did have a big dog,

    1. That's why I like the combination of strict liability AND insurance. Random draw bad stuff happens; that's what insurance is for. But premiums vary with owner quality.

  9. How politically unfeasible is it, really? Maybe a scheme for dangerous dog breeds, rather than all dog owners. A compromise law, rather than restricting pitbulls (or whatever other breed is in the headlines), mandate insurance for whatever breeds you would ban.

    @Jim Rose:
    "many owners of large dogs are judgment proof for civil litigation because of a lack of assets and income"

    Hence the desire for a mandatory insurance scheme.

    1. Owners very likely the larger problem. Yes, pit bulls disproportionately involved in attacks, but that could be because guys who want to look scary choose to buy pit bulls because those dogs look scary; they treat the dogs badly and they then are more likely to attack. Like how red cars are, allegedly, involved in more speeding incidents; people who like driving fast are more likely to buy red...

  10. Thanks Eric,
    if it is a random draw, deterring behavior is through people not buying a ticket in the lottery or buying smaller tickets.

    The Clifton study of attacks from 1982 through 2006, pit bulls, Rottweilers, Presa Canarios and their mixes were responsible for 65% of the canine homicides during that period of 24 years in the USA.

    Dog bites are the second most frequent cause of visits to emergency rooms from 9 activities common among children. are children to sue their parents?

    Three quarters of dogs involved in bite incidents belong to the victim's family or a friend. The majority of dog attacks (61%) happen at home or in a familiar place.

    The odds that a victim of a fatal dog attack will be a burglar are 1 in 177. The odds that it will be a child are 7 in 10.


    1. All dogs have random draw probability of biting. But some owners make their dogs more likely to bite, and some breeds are more likely to bite. Insurance premiums wouldn't punish folks who pose no particular risk and who have normal dogs. But folks with criminal records for mistreating animals would likely pay a fair bit more. And the insurance companies would have reasonable incentive to figure out which owner-breed combinations really posed heightened risk.