Case Study 1: The HPR Tradition — How Loss Control Became an Underwriting Strategy

What this case is. A look at a real and durable strand of property-insurance history: the highly protected risk (HPR) model pioneered by the mutual fire insurers that grew out of the New England textile industry — most prominently the Factory Mutual organizations now known as FM (FM Global). It is the clearest large-scale demonstration of this chapter's central claim: that a risk is not fixed — controls change it, and that loss control can be the whole underwriting strategy rather than an afterthought. The broad historical and engineering facts below are matters of public record; consistent with the book's rule, no proprietary or precise loss statistics are asserted, and any illustrative figure is labeled.

Background

In the nineteenth century, the textile mills of New England were among the most fire-prone industrial properties in the country: large, multi-story, full of combustible cotton fiber, lit and powered by dangerous early technology, and packed close together. Commercial fire insurers priced these mills as the hazardous risks they were — and many mill owners, convinced their own well-built and well-run factories were far better than the average mill the rates were based on, balked at paying the class price. This is the adverse-selection problem (Chapter 1) seen from the good risk's side: when one price covers a whole class, the best risks in the class are overpaying and have every incentive to leave.

So a group of mill owners did something that reshaped property insurance: they formed their own mutual insurance companies (Chapter 3 owns the mutual structure) to insure each other — but only if each member's factory met a rising standard of construction and protection. The bargain was explicit and is the heart of this case: you may join the pool, and pay a lower price, if and only if you make your factory a better risk. Membership required things that were novel at the time and are now universal — heavy "slow-burning" mill construction, fire walls and fire doors that compartmentalized a building so a fire in one bay could not consume the whole mill, organized water supplies, and, eventually, the technology that changed everything: the automatic sprinkler.

The underwriting / insurance issue

The HPR model inverts the ordinary relationship between underwriting and loss control. In standard commercial underwriting, loss control is something you assess — you inspect the risk, grade what you find, and price it. In the HPR model, loss control is something the insurer engineers and then requires, on a continuing basis, as the price of coverage. The defining features:

• Engineering, not just inspection. The HPR insurer's loss-control arm employs engineers who do not merely tick a checklist but study how fires actually start and spread, test protection equipment, and develop standards. The inspection is a partnership with the insured to reduce the risk, not a one-time grading exercise.
• Protection as a condition of membership. A property that lets its protection lapse — a sprinkler system impaired, a fire pump out of service — is not simply debited; it can lose its standing. This is the §9.5 idea, controls as a condition of coverage, taken to its logical end and applied continuously rather than once at binding.
• The two COPE letters that the model attacks. The whole strategy is a concentrated assault on the Protection letter of COPE and on the severity side of the frequency × severity equation. Sprinklers, fire walls, water supply, and detection do not stop every fire from starting (frequency), but they overwhelmingly determine whether a fire that starts is a contained, survivable event or a total loss (severity). The HPR insurer's bet is that a portfolio of protected risks will run at a dramatically better loss ratio than the unprotected market — good enough to justify the lower price the good risks wanted in the first place.

A note on the sprinkler itself, because it is the single most consequential loss-control device in this chapter. The automatic sprinkler is a severity control of almost unique cost-effectiveness: it activates locally, at the fire, without human intervention, and turns the overwhelming majority of structure fires into small events. The HPR insurers were central to its development, testing, and standardization — which is why, when this chapter tells you to verify the sprinkler system actually works, it is invoking more than a century of evidence that a working sprinkler is the difference between a \$40,000 loss and a total one. (The dollar figures here are illustrative.)

What it shows

This case is the living proof of three of the chapter's claims. First, a risk is not a fixed quantity. The same mill, with the same cotton and the same location, is a fundamentally different risk protected than unprotected — and the underwriter's job is to write the protected version, with the protection required and verified. Second, loss control serves the social function (theme six) at the same time as it serves the combined ratio: the controls the HPR model required did not just lower loss ratios, they saved factories, jobs, and lives, and the engineering standards developed for insurance purposes became general industrial safety practice. Third, the model is the answer to adverse selection from the good risk's perspective: by creating a pool that only the genuinely well-protected risks could join, the mutuals let the best risks escape the average price — and made themselves better risks in the process.

Outcome

The HPR approach endured because it worked. The mutual fire insurers that grew out of this tradition became, over more than a century and a half, among the most respected property carriers in the world, built around engineering and loss prevention rather than around chasing premium. The construction and protection standards they developed — slow-burning then fire-resistive construction, compartmentation, automatic sprinklers, organized water supply, the discipline of keeping protection in service — diffused far beyond their own insureds and became the backbone of modern property protection generally. The "highly protected risk" remains a recognized category in commercial property underwriting to this day (Chapter 19 returns to HPR vs. non-HPR in depth), and the idea at its core — that the insurer and the insured are partners in reducing the risk, not merely transacting over it — is exactly the posture this chapter asks you to bring to every loss-control recommendation you write.

Lesson

The best underwriting strategy may be to change the risk, not just to price it. The HPR tradition is the large-scale, century-tested version of what you do on a single account when you require a hot-work permit program and a sprinkler certification before you bind Harbor Steel: you write the risk as it will be under the controls you require, not the risk that walked in the door. Harbor Steel is, in fact, a non-HPR risk — original sprinklers of unverified function, an aging roof, fire protection class 4 — and that is precisely why the HPR contrast is instructive. You will not turn Harbor Steel into a highly protected risk overnight. But you will apply the same logic in miniature: identify the protection gaps, require the controls that close the most dangerous of them, verify them, and price what remains. The HPR insurers proved that this logic, applied with discipline across a whole book, is not charity to the insured — it is one of the most profitable ideas in the history of the business.

Discussion questions

1. The HPR model lets good risks escape the average class price by joining a pool only well-protected risks can enter. Explain, in adverse-selection terms (Chapter 1), why this is good for both the good risks and the insurer — and what happens to the unprotected mills left behind in the standard market.
2. The model attacks the Protection letter of COPE and the severity side of frequency × severity. Why is severity the right axis for a sprinkler-and-compartmentation strategy, and what would a complementary frequency strategy for the same mills have looked like?
3. HPR loss control is continuous — protection must stay in service, not just be present at binding. How does this differ from the single-account move of attaching a control as a condition of binding (§9.5, §9.7)? What morale-hazard problem (§9.4) does the continuous version solve that a one-time condition does not?
4. Harbor Steel is explicitly a non-HPR risk. Does the HPR tradition tell you to decline it, or does it tell you something more useful? Translate the HPR lesson into the specific loss-control recommendation you would write for Harbor Steel.
5. The engineering standards developed for insurance purposes became general industrial safety practice. What does that say about the relationship between actuarial self-interest (a better loss ratio) and the social function of insurance (fewer fires, jobs and lives saved)? Where, if anywhere, do the two come apart?