Case Study 2: When the State Rewrites the Rating Plan — California's Proposition 103 and Michigan's No-Fault Reform

Two real, public, legislatively-driven reshapings of personal-auto rating. The statutes, ballot measures, and reforms named here are real and documented; this study treats their design and consequences qualitatively, per the book's citation rules, and attaches no fabricated statistic to either. The point is not to score the policy debates but to show the underwriter what it means, operationally and ethically, when the state — not the desk and not the actuary — decides which factors a rating plan may use and how heavily.

Background

Chapter 14 (§14.5) makes an abstract claim: personal auto is the line where society negotiates, in public and factor by factor, the boundary between actuarial fairness (price reflects risk) and social fairness (access to affordable coverage). Two real episodes make that abstraction concrete from opposite directions — one constrains the factors, the other re-engineered the coverage mandate.

California's Proposition 103 (1988). Passed by ballot initiative, Proposition 103 restructured insurance regulation in the nation's largest auto market. Among its provisions: it moved California to prior approval of rates (the regulator must approve a rate before it is used); it created an elected Insurance Commissioner; and — most consequential for this chapter — it mandated that automobile rates give the greatest weight, in order, to the driving record, the annual miles driven, and the years of driving experience, with other rating factors permitted only in a subordinate role and subject to regulatory approval. Over the decades since, Proposition 103's implementation has included long and contested rate proceedings, restrictions on the use of credit in auto rating, and sustained legal and regulatory fights over the role of territory and other factors. It is the clearest example in the country of a state legislating that a rating plan prioritize the factors a driver controls over the territorial and demographic factors the unconstrained actuarial data might weight more heavily.

Michigan's no-fault system and the 2019–2020 reform. Michigan ran, for decades, a uniquely generous no-fault system: it required unlimited lifetime personal-injury-protection (PIP) medical benefits — the most generous auto medical coverage in the country. That mandate did exactly what Chapter 14 says a coverage mandate does: it drove price. Michigan was widely reported to have among the highest auto-insurance premiums in the nation, and the cost of the unlimited-medical benefit, together with the medical-billing and litigation dynamics around it, was central to why. In 2019, the legislature enacted a reform (effective 2020) that, for the first time, let drivers choose lower PIP medical limits (down to capped options, with the unlimited option retained for those who wanted it), introduced a medical fee schedule for auto- injury care, and adjusted some rating practices. It is the standing example of how the coverage the state requires the policy to pay drives the premium as powerfully as any rating factor in the plan.

The insurance / underwriting issue

Set side by side, the two episodes frame the chapter's regulatory theme from both ends.

Proposition 103 constrains the factors. It is, in effect, a legislative judgment that some predictive factors are too disconnected from the driver's own conduct (territory) or too entangled with income and demographics (credit) to be given free rein — even though they predict loss. The actuary's instinct, as §14.5 explains, is to use every factor that predicts, because that is what keeps the price matched to the risk and the pool free of adverse selection. Proposition 103 overrides that instinct in the name of social fairness and a particular notion of desert: that a driver's price should turn first on what they do (record, miles, experience), not on where they live or what their credit looks like. Whether that improves fairness or merely moves the cross-subsidies around is exactly the kind of question the book refuses to resolve glibly — but the operational consequence for an underwriter is unambiguous: in California, the rating plan you apply is a legally re-ordered one, and the strongest data-driven factors are subordinated by law.

Michigan re-engineered the mandate. Here the lever was not which factors you may use but what the policy must cover. An unlimited lifetime medical benefit is, in pure insurability terms (Chapter 1), an extraordinary promise: a low-frequency but potentially unbounded severity exposure on every policy. That mandate had to be priced, and pricing an unbounded tail is expensive — which is a large part of why Michigan premiums ran so high. The 2019–2020 reform let consumers trade some of that protection for a lower price, and it attacked the cost side directly with a medical fee schedule. The underwriting lesson is the mirror image of Proposition 103's: sometimes the most powerful determinant of a line's price is not in the rating plan at all — it is in the coverage architecture the state has mandated, and changing the price means changing the promise.

What it shows

  • The rating plan is not the underwriter's (or even the carrier's) to design alone. In personal auto the state holds a pen over the plan — sometimes re-ordering the factors (California), sometimes re-shaping the coverage itself (Michigan). The underwriter operates inside whatever boundary the law has drawn, and "what may I use here, and as filed?" (§14.5) is a live, state-specific question, not a settled one.
  • Actuarial fairness and social fairness genuinely conflict, and the conflict is resolved politically. Proposition 103 is a deliberate choice to sacrifice some actuarial precision for a fairness goal — to subordinate predictive-but-contested factors to driver-conduct factors. That is a defensible social choice and it has actuarial costs (the subordinated factors predicted something real; suppressing them moves the cost onto other drivers). Both halves are true, which is the whole point of Chapter 35.
  • The coverage mandate is a rating factor in disguise. Michigan shows that you cannot understand a line's price by studying the rating plan alone; the required coverage sets a floor under the loss costs. An unbounded medical benefit is an unbounded exposure, and no rating sophistication makes an unbounded tail cheap.

Outcome

Both reforms reshaped their markets durably. Proposition 103 remains the framework for California auto rating decades later, and its prior-approval regime, factor-ordering mandate, and restrictions on credit and territory continue to generate contested rate proceedings — the regulatory lag of §14.7 in an especially pronounced form. Michigan's 2019–2020 reform changed the product consumers buy (choice of PIP medical limits) and attacked auto-medical cost directly; its full long-run effects on price and on the catastrophic- injury care system have continued to be debated and litigated. In neither case did the reform "solve" personal auto — both states still grapple with affordability and availability — but both permanently changed what an underwriter in that state is allowed to do and must price for.

Lesson

  • Know your state, and re-ask the question. The single most state-dependent line in the book is personal auto. A factor that is mandatory-to-prioritize in California, a coverage that is mandated-then-made-optional in Michigan, a credit rule that flips at the border — these are not trivia; they are the boundary of your job. Memorizing one state's plan is not knowing the line.
  • Separate "predictive" from "permitted," and respect the gap. Proposition 103's subordinated factors did not stop predicting; the law decided their social cost outweighed their actuarial benefit in California. The disciplined underwriter holds both facts (§14.3's trap) and operates inside the boundary without arguing with it from the desk.
  • Look past the rating plan to the mandate. When a line's price seems inexplicably high, check what the state requires the policy to cover. Michigan's unlimited PIP shows that the coverage architecture can dominate every factor in the plan.

Discussion questions

  1. Proposition 103 mandates that driving record, miles, and experience get the greatest weight. Using §14.2, evaluate this ordering on pure predictive grounds: are these in fact the three most predictive factors? If not, what is the state trading away, and for what? (§14.2, §14.5; Ch. 35)
  2. Michigan's unlimited-PIP mandate created an unbounded severity exposure on every policy. Connect this to the insurability criteria of Chapter 1 (especially "definite and measurable" and "not catastrophic to the insurer"). Why is an unbounded benefit so expensive to price? (Ch. 1; §14.1)
  3. A carrier argues that Proposition 103's subordination of credit and territory forces good drivers in high-cost areas to be over-charged and bad drivers in low-cost areas to be under-charged — i.e., that the reform creates its own unfairness. State the strongest version of this argument and the strongest rebuttal. (§14.3, §14.5; Ch. 35)
  4. Both episodes show the state re-drawing the boundary between actuarial and social fairness. As an underwriter, what is your professional duty when you believe a state's rule subordinates a factor that genuinely predicts loss? Is "comply, and price within the boundary" a sufficient ethical posture? (§14.5; Ch. 35)
  5. Contrast the mechanism of these two reforms: California constrained the factors; Michigan re-shaped the mandate. If your goal were to make auto insurance more affordable without simply hiding the cost, which lever has fewer hidden cross-subsidies — and what would you watch to find out? (§14.5, §14.7)