The Insurance Corporation of British Columbia (ICBC) is in trouble. Not only are car insurance rates in BC among the highest in Canada, but costs have risen so dramatically in recent years that the crown corporation is in desperate need to pass on the higher costs to its customers through higher premiums (or decrease payouts). Some have also argued that it's time to break up ICBC and privatize car insurance in BC. Competition usually helps keeps prices low, but in BC's case the lack of competition is not the main problem that needs solving. So what's ailing car insurance in British Columbia?

Why is Quebec's average insurance rate only about two-thirds of the average rate in British Columbia? Only Ontario, which has a private auto insurance market, has higher average premiums than BC. In principle, ICBC runs a break-even profit system, although stricter reserve requirements first led to an increase in reserves that were subsequently tapped by the provincial government to fix holes in the budget. But this unfortunate forth-and-back by the provincial government does not explain the rapid rise in costs. What has spiraled up and up is the frequency and cost of accidents and, even more so, the cost of injury claims. Collisions are up, in particular at intersections (23% between 2011 an 2015, with a 17% increase in injured victims). Bodily injury claims are up 28% since 2012. Disturbingly, the cost average bodily injury claim paid out for minor injuries has risen from $8,220 in 2000 to $30,038 in 2016, an almost four-fold increase. Similarly increases have been seen in pain and suffering awards.

The basic idea of insurance is to pool risk, as we cannot predict precisely who will have the misfortune of getting into an accident. Instead, we can use actuarial principles to determine who is more likely to be involved in an accident. Thus insurance may discriminate between new and experienced drivers, or between young and the old, or between urban and rural drivers, or between frequent drivers (commuters) and occasional drivers. Demographic information such as gender and martial status may also be used, but at least the use of gender is quite controversial. ICBC is in fact prohibited by provincial law from using personal characteristics such as age and gender in setting its rates. It is also possible to discriminate between different makes and models of vehicles, and this is widely used in practice. But these groups are still quite broad, and public insurers may be inclined to squeeze the differentials in order to keep insurance rates "affordable". Excluding personal characteristics to vary rates leads to broad pooling, because what is left is simply the preference (or statistical correlation) of certain demographics for certain brands and models. But group information and personal characteristics will always remain inferior to direct measurement of driving behaviour—the "holy grail" of fairness for setting car insurance rates.

British Columbia, like all other provinces, makes use of a no-fault principle for basic insurance coverage. Those injured in a collision receive compensation and benefits from their insurer regardless of fault. This does not mean that there are no adverse consequences to being at fault in a collision; a driver's insurance premium is usually affected by an at-fault accident. Adversarial legal ("tort") systems still surround this no-fault scheme in some provinces. Only Quebec has a pure no-fault system. BC offers a hybrid system that retains the right to sue. The no-fault "Part 7 Claim" in BC stipulates maximum damage amounts (e.g., $150,000 for medical expenses), while a motorist at fault may still be sued by the injured party for pain and suffering, cost of future care, lost wages, and lost future earning capacity. For the at-fault motorist extended third-party liability will cover the claim up to the insured limit. Basic Autoplan only covers the first $200,000; but many motorists buy liability coverage of up to $5 million. While it has been claimed that a tort system deters poor driving, it turns out that BC has higher rates of accidents and injuries than other provinces. Injury lawyers abound in BC's hybrid tort-based system. Some province have moved entirely to a no-fault system where injury compensation is determined similar to the Workers' Compensation Board (no WorkSafeBC). The problem with tort is that it encourages fraudulent claims.

In a press release, ICBC pointed to its efforts to reduce insurance fraud (false claims, exaggerated claims, and organized fraud). They estimate that the cost of fraud amounts to the equivalent of $100 per year per policyholder, but as fraud is difficult to detect, these estimates may be at the low end. Stepping up fraud detection and enforcement must be one of the highest priorities for ICBC to keep costs in check.

Ultimately, detecting who is a good driver and who is not will provide better and fairer insurance rates. This will mean that insurance rates for bad drivers will soar, but so be it. If high insurance rates keep bad drivers off the road altogether, so much the better, as long as they can't drive someone else's vehicle. And here is the underlying problem. Car insurance insures vehicles, not drivers. Attached to a vehicle is a "principal driver", but if this is selectively a driver with a good record and the vehicle is in fact driven by a higher-risk driver, the insurance premium is mispriced. For example, a vehicle insured by a parent but driven by a teenage son or daughter will have a higher collision probability but be insured at the low rate of the experienced driver. Of course, if a vehicle is involved in an accident, it doesn't matter who drove the vehicle; the insurance rate of the experienced driver may go up.

Insurance risk depends both on the driver and the vehicle that is driven. Perhaps insurance should be two-fold: a separate insurance for each driver and each vehicle. Insurance that simply correlates that younger drivers prefer driving certain types of makes and models is not sufficient. Why not put an insurance rate straight on each licensed driver? If you have a license and want to drive a vehicle anywhere in your country, then why not require drivers to buy appropriate insurance regardless of when and where you actually drive a particular vehicle? Perhaps such an insurance system will discourage risky drivers, as young people will postpone obtaining a license. Without a financial penalty on risky drivers, risky drivers will continue to drive. A separate insurance on vehicles will also be better able to attribute the risk to particular models and makes. The current principal-operator principle used in BC is a big loophole as it is impossible to ascertain who is actually driving the vehicle most of the time.

Making the insurance cost of particular vehicles more transparent could also help. Expensive vehicles are more expensive to repair, and providing upfront (perhaps mandatory) vehicle insurance cost information at the point of sale could encourage buyers of motor vehicles to make better-informed choices. Some vehicles are safer than others, and lower insurance rates for such vehicles may steer customers to these better (but more expensive) choices. ICBC recently excluded vehicles that cost more than $150,000 from its coverage. Clearly, then, the existing insurance premium for such vehicles was much too low—especially when these Ferraris and Maseratis have "N" stickers on their back.

Private insurance could help, but only if this brings up needed innovation. Private insurers tend to engage in "cherry-picking" low-risk customers and spurning crash-prone drivers. Ultimately, this will leave some drivers unable to obtain insurance as no company will be willing to underwrite their risk. Proactive insurers are able to monitor their drivers' behaviour better as well as incentivize and influence their behaviour. For example, insurance companies could provide a discount to drivers willing to allow a GPS-enabled device to be installed that tracks driving behaviour and risk ("telematics"). Such a voluntary measure, targeted at higher-risk groups such as new drivers, could be useful to encourage these drivers to pay more attention and improve driving behaviour—and receive a lower insurance premium quicker. The goal is not to exclude people from obtaining insurance, but to make our roads safer. Incentives need to be adjusted so that the riskiest drivers are encouraged to improve their behaviour and lower their premiums.

Fixing ICBC is entirely possible even without privatization. Innovation is needed, and lots of it. If ICBC can bring about this innovation and restructure its premium system, that would be the simplest course of action. If ICBC and its political masters are unwilling to embrace painful reform, ICBC's performance will continue to deteriorate and privatization of the auto insurance market will become unavoidable. It is not too late to turn ICBC around. Is there the political will for reform, even when such reform will create both winners and losers?

On July 24, 2017 the Times Colonist reported that the new NDP government rejects a 30% ICBC rate hike, photo radar, and no-fault insurance. (To be clear, BC has a hybrid system that combines tort claims with basic no-fault coverage. The minister probably meant a pure no-fault system as in Quebec.) A recent report from consultancy Ernst & Young points to a number of measures that should be pursued to fix ICBC's problems:

• Road Safety: double the number of intersection cameras and use them both for red-light and speed-on-green infractions; also use such cameras at high-risk sites; use average-speed enforcement and variable speed limits where appropriate.
• Safer Driving: combat distracted and impaired driving with more road safety officers.
• Insurance caps: limit payouts on minor injury claims and move BC's hybrid system a little bit more towards no-fault adjustment.

These are very sensible recommendations from Ernst & Young. However, what the report fails to address is the incentive system for drivers beyond penalties for risky behaviour. Insurance rates must reflect risk more closely, and this means improved differentiation. Current legislation prohibits the use of personal characteristics, but it does not prohibit insuring drivers and vehicles separately. This would be a start. Ignoring age and gender for young drivers is a folly that belies actuarial risk—young male drivers are notoriously more risky drivers than their female counterparts, and young drivers are more risky than middle-agedd drivers, while very old drivers are again more risky than middle-aged drivers. If such differentiation is politically problematic, the remedy is to move to observing actual driving behaviour. Drivers willing to trade a little privacy for lower insurance rates should be given that option. Allow higher-risk drivers to prove their road-worthiness by offering voluntary, usage-based insurance. In Quebec and Ontario, Desjardins and Industrial Alliance have already pioneered such programs. It is time that BC followed and introduced more technological innovation along with better rate differentiation. Such measures will create winners and losers, but ultimately will be much fairer.

The new government in Victoria would be well advised to focus on the solutions rather than blaming its predecessor, worthy of blame as they may be for failing to fix ICBC earlier when they had a chance. It will involve making some hard, unpopular choices that the previous government failed to make. Kicking the ball forward is popular in politics, but it's bad policy. Car insurance in BC can become more affordable again. It takes more carrot (a more-differentiated premium system) and more stick (stricter enforcement and penalties).