30 Million Dangerous Dogs?

Here's an excerpt from a recent paper that came my way, published by the Animals and Society Institute:*

"Dog Bites: Problems and Solutions - a Policy Paper"
Existing and Proposed Legal Remedies - Attempts to Identify and Remove high risk animals: Dangerous Dog Laws"

The second common legislative approach (the first is breed bans) to remove dangerous dogs from the population targets the behaviour of the individual dog, designating dogs with labels such as "potentially dangerous", "dangerous", or in some cases "vicious" based on actual incidents, and then either eliminating the dogs or limiting the conditions under which they may be kept (such as requiring sterilization, micro-chipping, training, behavioural consultation, muzzling, etc.) Such laws increasingly also specify civil and criminal liability incurred by people whose dogs injure someone after receiving such a designation.

There is some evidence that a "prior behaviour" approach to the "dangerous dog" designation may decrease injurious bite incidence. This has only been demonstrated, however, where the dangerous dog label is limited to dogs who have already bitten and injured someone. A program in Oregon showed a decrease from 25 percent to 7 percent in repeat injurious bites (of people) after the implementation of a program restricting conditions of ownership of dogs who had injuries.

Many dangerous dog laws try not only to control dogs who have already injured people, but to predict which ones will do so in the future and attempt to prevent this. Typical legal descriptions of "dangerous" dog behaviour include "approaches in a vicious or terrorizing manner", "in a menacing fashion", having a "known disposition, tendency, or propensity", or "engages in any behaviour that requires a defensive action by any person to prevent bodily injury".

Aside from the subjectivity of these descriptions, the main difficulty with such an approach is that the best research to date indicates the likelihood that a majority of dogs engage in such behaviour without hurting anyone.

One groundbreaking study found that 41 percent of the dogs studied had growled, snarled or snapped at a familiar person at some time, but only 15 percent of those dogs actually bit anyone. Of those who bit someone, only 10 percent of the bites were considered injurious, making the total incidents of injurious dog bites only 1.5 percent of the total dogs studied.

This means that a hypothetical net cast to identify the 1.5 percent of dogs actually captures at least 41 percent of the dog population. And since this study only included behaviour toward family members and other people well known to the dog, and only included guardians responsible and caring enough to provide veterinary care for their companions, the percentage of potential problems within the entire dog population must certainly be considerably higher.

A history of threatening behaviour has not been shown to predict that a dog will bite, much less that she will injure if she bites.
...

Other Outcomes

With regard to dangerous dog laws based on behaviour, as discussed above, definitions of dangerous are so varied and subject to interpretation that most dogs' behaviour could be interpreted to qualify.

A conservative estimate would be about 30 million dogs would likely meet the criteria (this estimate is based on a study that found 41 percent of dogs growl, snarl, or snap at a familiar person, and thus does not include dogs that only threaten strangers, so the real percentage is almost certainly considerably greater).

Some statutes require only that the dog "endangered" a person in some way, leaving the way open for complaints by anyone who simply felt (but was not really) endangered.

All this creates a serious danger of abuse in any system that attempts to weed out "potentially dangerous" animals who have not bitten anyone. It casts a net far too wide to be enforceable.

When laws exist without the practical means to widely enforce them, the result is selective enforcement based on grudge complaints, and widespread non-compliance.

*Boldfacing, underlining, a few parenthesized notes, and paragraph separations, added by me).