The history of the American administrative state has been filled with attempts to introduce “rationality” and “discipline” to circumstances in which markets, left to their own devices, allegedly lead to socially or politically undesirable outcomes.

Brimming with good will and well-meaning intention, champions of state intervention seek to invent systems to accommodate and serve those least able to care for themselves. Californians can even directly participate in such efforts through the initiative process.

In 1988, a slim majority of California voters passed Proposition 103, a landmark initiative they were told would reform the state’s auto insurance market for the better. What they did not count on was that, more than 25 years later, Prop 103’s architects would still be standing protectively, to thwart any attempt to update the initiative to reflect changes in the market, technology and the world at-large.

Within a few years, experts now predict, self-driving and autonomous vehicles will radically change the risk profile of vehicles on California’s roads. A recent study by the Casualty Actuary Society found that many of the risks associated with automated vehicles can not only be identified, but they can also be quantified, and therefore eliminated. But while that’s encouraging from a health and safety perspective, thanks in part to Prop 103, reducing accidents by such an overwhelming margin could have perverse impacts on drivers who can’t afford to upgrade to cars with more advanced safety technology.

Prop 103’s process for approving auto insurance rates is rigid. There’s a statutory hierarchy, with three mandatory rating factors that insurers must take into account above all others, the first of which is accident history. Self-driving and autonomous vehicles will experience far fewer accidents as a result of driver error, which currently is and historically always has been the cause of most accidents.

Since operators of self-driving and autonomous vehicles will experience fewer accidents, their premiums will decrease. But the benefits conferred to early adopters will be balanced out by higher rates for those who continue to drive conventional vehicles. Under Prop 103, insurance rates are a zero-sum game.

The law requires insurers to increase or decrease the significance of cost factors to comply with the rating-factor hierarchy. This “pumping and tempering” occurs independent of the actual cost associated with a rating factor. Thus, even though other rating factors will be of greater significance to self-driving and autonomous vehicles, Prop 103’s inflexibility will continue to require accident history to be the single most important determinant in the rate-making process.

The result will be that other members of the “class plan” will have to make up the reduced cost of early adopters. In a state split between vehicle operators who bear transportation risk and others who do not, Prop 103 will burden those least able to foot the bill.

If traditional drivers are made to cross-subsidize operators of self-driving and autonomous vehicles, it is likely Prop 103 will run afoul of its own judicially constructed purpose. In a 2005 state appeals court decision, Foundation for Taxpayer and Consumer Rights v. Garamendi, a discussion of the appropriateness of a legislatively enacted change to Prop 103 led the court to find that the initiative’s purpose was to protect the uninsured from arbitrary rates, so that insurance could be “fair, available and affordable.” In the context of a persistency rating factor that the court decided subsidized the insured at the expense of the uninsured, that purpose was deemed violated.

But what happens when existing language, faced with novel circumstances, leads to an identical result? Left unchanged, Prop 103 will violate its own purpose.

Perhaps, in the name of “fairness,” and to forestall the inevitable dismantling of their ugly creation, the drafters of Prop 103 will seek a ban on autonomous vehicles. For now, fortunately, they are satisfied to ignore reality and claim that “California is a long, long way from the so-called ‘autonomous vehicle.'”

To avoid problems like these in the future, the Prop 103 model should be scrapped entirely. Failing that, to properly address the risk presented by self-driving and autonomous vehicles, different rating factors will need to be introduced and given hierarchical flexibility. Contextual factors will be of particular importance, since they likely will present the greatest hurdles to self-guidance programs. For instance, a region’s weather and the quality of its road network could both predictively relate to an autonomous vehicle’s risk profile.

Increasingly and inevitably, Prop 103 is proof that the regulatory rationale of one moment can grow stale in the next. Prop 103 has swollen useless and counterproductive government intervention. More distressing still, Prop 103 now may injure the very same economically deprived people it was passed to help.

Featured Publications