Insurance: the purest reuse

Almost nothing new

If tax is the domain that falls out for free because two cross-cutting primitives happen to apply, insurance goes one step further. Insurance is the domain that is nearly all reuse, top to bottom, built from primitives that already exist in other fields and recombined into a new shape. It contributes essentially no new mechanism of its own.

That absence is not a disappointment. It is the entire point of writing about insurance at all. A whole regulated industry, with its own regulators, its own vocabulary, its own century of accumulated rules, turns out to require almost nothing the engine did not already have. Insurance is the clearest evidence that the leverage at the center of this engine is real and not a slogan, because insurance is where you would most expect to need new machinery and you end up needing almost none.

Look at what insurance actually requires, piece by piece, and notice where each piece already lives.

Solvency adequacy, the question of whether an insurer holds enough capital against its requirement, is a coverage ratio. Eligible funds over required capital, compared against a threshold. That is the same numeric comparison that polices a bank's capital ratio. Banking built it, and insurance uses it unchanged.

Product governance and suitability, the question of whether a product is sold to a defined target market and genuinely suits the customer's demands and needs, is a conduct check. That is the same shape as the suitability rules in securities, where an instrument has to fit the investor it is sold to. Securities built that evaluator, and insurance does not just imitate it. It shares the very same one.

Claims handling, the question of whether a claim was acknowledged, decided, and settled within the required time, is a deadline. The clock from letters of credit, from securities filings, from breach notifications. Built three domains ago, reused here without a modification.

Three of the core obligations in a major regulated industry, and every one of them is a primitive from somewhere else.

What is data, and what is a rule

The reference layer is the parameter tables, and they follow the pattern every domain follows. Solvency factors by risk module. Product and disclosure templates. Claims-handling service levels by jurisdiction. And alongside the regulatory tables, the insurer's own product catalog, which is customer-supplied, the way a legal playbook or an AI-system inventory is customer-supplied, because only the insurer knows the full shape of what it sells.

The rules are the obligations on top, and they read like obligations from any other domain. Coverage below the required ratio. A product distributed without a defined target market. A claims milestone missed. A disclosure document left incomplete.

{
  "rule_id": "INS-SCR-EU-001",
  "title": "Solvency capital coverage ratio below 100%",
  "jurisdiction": "eu",
  "source": "Solvency II",
  "severity": "block",
  "expected_outcome": {
    "action": "review",
    "message": "Eligible own funds do not fully cover the Solvency Capital Requirement. Under Solvency II this triggers supervisory action and a recovery plan. Confirm the figures and escalate."
  },
  "conditions": [
    { "type": "numeric_comparison", "path": "solvency.scr_coverage_ratio", "operator": "<", "value": 1.0 }
  ],
  "deterministic": true,
  "validation_status": "expert_reviewed"
}

There is no exotic evaluator in this rule. The single condition is a numeric_comparison, the same one banking uses to check a capital ratio, pointed now at a solvency figure instead of a bank's. A regulator on one side of the financial system and an insurer on the other, checked by the identical operator, because underneath the different vocabulary it is the identical question: is the ratio below the line.

Fail closed

The discipline does not relax just because the domain is mostly reuse. A missing solvency factor, an unresolved product that the catalog cannot identify, and the engine says it could not certify compliance, rather than presenting an unverified result as a clean one. Reuse changes where the parts came from. It does not change the standard they are held to. An insurance verdict built entirely from borrowed primitives is still required to admit when it could not actually check.

The point

As a demonstration, insurance is close to invaluable, because of how little it asks for. A whole regulated industry, prudential capital regulation on one side and market conduct regulation on the other, two halves that insurers themselves often run as separate departments, delivered almost entirely by recombining primitives that banking and securities had already shipped for their own reasons.

When a new domain costs this little to stand up, the claim at the center of everything stops sounding ambitious and starts sounding like arithmetic. Each domain pays a primitive forward. The next domain spends primitives that are already there. By the time you reach a domain like insurance, late in the order, so much has been built that the domain is mostly assembly. One engine. Every regulated domain. And the genuine surprise, the thing that is easy to say and hard to believe until you have watched it happen six times, is that the more you build, the cheaper the next one gets.