The onion: A through F#

Open any American homeowners policy and the coverages are lettered A to F. The letters look bureaucratic. They’re actually concentric — each one sits further out from the thing at the centre, which is the building itself.

Coverage A — the dwelling. The structure. Walls, roof, floors, built-in appliances, and anything physically attached to it: the attached garage, the deck, the front porch. This is the core of the onion.

Coverage B — other structures. The stuff on your land that isn’t attached: the detached garage, the shed, the fence, the gazebo, the pool house. One layer out. Conventionally set at 10% of Coverage A, which tells you how the industry sizes the risk.

Coverage C — personal property. Everything inside that isn’t nailed down. Furniture, clothes, the TV, the espresso machine. Usually 50–70% of Coverage A on an owner-occupied policy — the insurer’s rough guess at how much stuff fits inside a house of that value.

So far the onion is physical: attached, detached, loose. Then it stops being about objects at all.

Coverage D — loss of use. The house is standing but unlivable. Coverage D pays for the hotel, the restaurant meals, the extra costs of not being home. Notice the shift: A, B and C insure things, D insures the consequence of losing them.

Coverage E — personal liability. Now we’ve left your property entirely. Your dog bites the mailman. Your kid puts a baseball through a neighbour’s window. Someone slips on your steps and sues. This pays what you owe them, plus your legal defence. Typically $100,000 as a standard limit, with$300,000 common.

Coverage F — medical payments to others. The outermost skin, and the strangest layer in the whole onion.

Coverage F is the weird one, and it’s weird on purpose#

Coverage F pays a guest’s medical bills when they’re hurt at your place, regardless of fault. No lawsuit required. No finding of negligence. Nobody has to admit anything. The limits are small — $1,000 to$5,000 per person.

The first time I read that, it looked like a rounding error. Why would an insurance company hand out money without establishing that anyone was liable?

Because it’s cheaper than the alternative. Think of it as the insurance equivalent of a restaurant comping your dessert when the kitchen is slow. Nobody’s admitting the kitchen failed. They’re spending five dollars to prevent a one-star review.

Your neighbour trips on your front step and needs eight stitches. If there’s no mechanism to just handle it, her next move is to call a lawyer, and now you’re in Coverage E territory: a liability claim, a fight about whether the step was up to code, legal fees on both sides, and a neighbour you’ll be avoiding at the mailbox for a decade. Coverage F short-circuits all of that for two thousand dollars.

Which is why it has a telling exclusion: Coverage F does not cover you or your household. It isn’t health insurance. It’s a mechanism for keeping other people from becoming plaintiffs. Once you know that, the design is obvious — and once you notice the pattern, you start seeing it everywhere in insurance. Many of the odder-looking coverages exist to prevent a more expensive claim rather than to pay for a loss.

The onion has a seam#

Before we leave the letters, one structural fact that turns out to explain more than anything else in this article.

Those six coverages aren’t a flat list of six. They’re two blocks, and the policy is physically organised that way:

1
Section I   —  A, B, C, D   —  property
2
Section II  —  E, F         —  liability

Section I is about your stuff and your use of it. Section II is about other people. The break falls exactly where the onion stops being about your property and starts being about the world outside it.

And here’s the part worth remembering: Section II is the same text in every homeowners form. It isn’t tuned per form — HO-2’s liability language is HO-5’s liability language. A form either carries Section II or it doesn’t.

That sounds like trivia until you notice what the industry does with it. The seam is where products get cut.

(Added later: I eventually worked out what this seam is actually made of, and it turned out to be the single most useful thing I’ve learned doing this — it’s the first-party / third-party division, and the same seam runs through auto insurance under completely different vocabulary. It also explains why Section II never varies, and why Coverage F sits on the liability side despite paying medical bills. Written up here.) Later in this article we’ll meet a whole family of policies that has Section I and simply no Section II at all — no liability, no medical payments, buy it separately — and the reason is not that landlords don’t get sued. It’s that the seam was there to cut along.

Why the forms have numbers at all#

Here’s the part that surprised me most: HO-3 isn’t a product. It’s a template.

An organization called ISO — the Insurance Services Office — writes standardized policy forms, and most American carriers build their products on top of them. So HO-3 is a document that thousands of unrelated insurers have all adopted as their starting point.

The relationship works like generic drugs and brand names. HO-3 is the molecule. The carrier’s marketing department supplies the brand: “Quantum Home”, “Executive Capstone”, “Optimum Protection”. Peel back the branding and you find the same base form underneath.

This is enormously useful if you’re writing software that reads these documents, because it means the space of legal answers is small. It’s also a trap, because carriers absolutely do file their own numbering — and then your neat little whitelist starts rejecting real policies. More on that at the end.

The one switch that matters: named perils vs open perils#

Every difference between the homeowners forms comes down to one question, asked separately for the building and for your stuff:

Named perils means the contract lists the disasters it covers. Fire, lightning, hail, explosion, riot, aircraft, vandalism — a specific enumerated list. If your loss isn’t on the list, it isn’t covered.

Open perils — also called all-risk — inverts it. Everything is covered unless the contract specifically excludes it.

It reads like a subtle distinction. It isn’t. It flips who has to prove what.

Named perils is a guest list. You’re standing at the door with your loss, and you have to point to a line on the list and say “this one, item 13, accidental discharge of water.” No line, no entry.

Open perils is a bouncer with a banned list. You walk in. If the insurer wants to stop you, it has to point at the list and show that your loss is on it.

Same fire, same water damage, same roof — but the burden of proof has changed hands, and in a contested claim that’s frequently the whole ballgame.

The ladder: HO-1, HO-2, HO-3, HO-5#

Now the numbering makes sense. Four forms, all for owner-occupied homes, and each step up flips exactly one switch:

HO-1 → HO-2 swaps how you get paid. HO-2 → HO-3 flips the building to open perils. HO-3 → HO-5 flips your belongings too. That’s the entire ladder.

That settlement column deserves a line of its own. Actual cash value means depreciated: your ten-year-old roof burns down and you’re paid for a ten-year-old roof. Replacement cost pays for a new one. It’s the difference between a car’s trade-in value and its sticker price, and on a total loss it can be tens of thousands of dollars.

Why HO-3 won#

HO-3 is the most common homeowners policy in the United States, and it’s not a compromise between the cheap one and the fancy one. It’s each risk set to its correct position.

Losses to the building are large and rare. You might have one in a lifetime, and it could be six figures. That’s exactly the risk worth writing on open perils — partly because the frequency is low, and partly because arguing over which peril caused a burnt-out house costs more than the argument is worth.

Claims on your belongings are small and frequent. Someone’s bike gets stolen, a TV dies, a laptop goes swimming. Here a named list earns its keep: it filters out the endless marginal claims, and the premium saving is big enough that customers feel it.

So HO-3 is the building on all-risk, the contents on a list. Not the middle of the ladder — the right rung.

HO-5 is then simply HO-3 with the contents flipped to open perils as well. One cell in the table, and it becomes the broadest standard homeowners policy sold. It’s positioned as the flagship: stricter underwriting, higher premium, and a habit of turning the optional extras into standard ones.

Why HO-1 has nearly vanished#

HO-1 is technically still a homeowners policy. In practice you’ll almost never see one — ISO’s own reference works describe it as discontinued in nearly all states. Two reasons, and they compound.

The first: it pays actual cash value. Mortgage lenders won’t accept that. The bank’s interest isn’t in your house being worth something, it’s in your house being rebuildable, so lenders require replacement cost. Most owner-occupied homes carry a mortgage, so most owner-occupied homes can’t use HO-1.

The second is simply that it’s thin, and buyers noticed. HO-1 runs on a short named-perils list — roughly ten causes of loss, against the sixteen-plus on the forms above it — and ACV settlement on both the structure and the contents. The industry’s own account of why it disappeared is exactly that: buyers demanded the broader coverages available in the other ISO forms, so carriers stopped filing it.

That’s the whole story, and it’s a market story rather than a regulatory one. Nobody banned HO-1; people just stopped buying a policy that pays for a ten-year-old roof with ten-year-old money. If I see one in a document today, my first instinct is that something got copied out of a comparison table by mistake.

The forms that aren’t for homeowners at all#

The HO prefix covers more than owner-occupied houses, and the giveaway is always the same question: what do you actually own?

HO-4 — renters. Coverage A and B are simply not there, because the walls aren’t yours. Your landlord insures the structure; you insure your things. So Coverage C stops being a percentage of anything and becomes the primary coverage, with a limit you pick yourself.

HO-6 — condo. Coverage A exists but only covers walls-in: your finishes, fixtures, and the improvements you made. Coverage B is gone entirely, because everything outside your unit belongs to the association’s master policy. And HO-6 is where loss assessment earns its keep — your share of a special assessment when the association’s own coverage falls short. The $1,000 base limit is actually built into every HO form, but on a condo it’s the one that gets used, and residents of an association are usually advised to endorse it upward. That risk only exists because of how condo ownership works.

HO-7 — mobile and manufactured homes. Modelled on HO-3, adapted for how those homes are built.

Look at HO-4 and HO-6 side by side and you can watch the ownership boundary move. Coverage D on an owner’s policy is a percentage of Coverage A. On a renter’s or condo policy it’s a percentage of Coverage C, because there’s no meaningful Coverage A to hang it off. Same coverage, different base — because the property line moved.

The DP forms: when you own the house but don’t live in it#

Everything so far has been HO — owner-occupied, renters, condo. There’s a second family entirely, and it’s for landlords.

DP-1, DP-2 and DP-3 are the dwelling fire forms: rental properties, vacant houses, seasonal homes. They climb the same peril ladder — DP-1 named, DP-2 expanded named, DP-3 open — and DP-3 is the common landlord policy for the same reason HO-3 is the common homeowners policy.

But three things change, and each one follows from the fact that the owner doesn’t live there.

Coverage D means something different. On an HO form it’s additional living expenses — the hotel bill while your home is repaired. On a DP form it’s fair rental value: the rent you aren’t collecting while the property sits unrepaired. Same letter, opposite side of the transaction.

Coverage C is optional. A rental usually doesn’t contain the landlord’s furniture, so personal property isn’t automatic.

Coverages E and F aren’t there at all. This is the seam from earlier, cut. The standard dwelling fire form has Section I and no Section II — no liability, no medical payments. A landlord buys it back separately, either as a supplement bolted onto the dwelling policy or as a freestanding personal liability policy.

This one genuinely surprised me. The coverage I’d assumed was the floor of any property policy is absent from an entire family of forms — and not because landlords don’t get sued. A tenant or a visitor hurt on the property is among the most typical claims a landlord will ever have. It’s also a trap in a specific way: the liability on the landlord’s own homeowners policy does not extend to a house they rent out, because that house isn’t an insured location under it. Owning insurance on the home you live in does nothing for the one you let.

The whole thing on one grid#

Here’s the payoff. Forms across the top, coverages down the side. ● included as standard, ○ available but not automatic, — not part of the form at all.

• ¹ walls-in only — the association’s master policy covers the exterior
• ² a percentage of Coverage C, not Coverage A — there’s no meaningful Coverage A to hang it off
• ³ fair rental value, not additional living expenses
• ⁴ contents are ACV by default on most forms; replacement cost is an endorsement
• ⁵ 1,000 is built into every HO form. It matters most on HO-6, but it is not exclusive to it; the endorsement buys a *higher* limit, up to \50,000 under ISO rules
• ⁶ HO-1 has been discontinued in nearly all states, and secondary sources disagree on its details. Liability and medical payments (Section II) are part of the form per ISO; loss of use is the one I’d treat as uncertain

Read the dashes and you can see the whole market at a glance. Two dashes at the top of HO-4 — a renter owns no building. One dash on HO-6 — a condo owner owns the inside of the walls and nothing beyond them. Three dashes at the bottom of the DP column — a landlord’s standard form carries no liability and no loss assessment.

Every gap in that grid is a property line drawn somewhere in the real world.

HO-8, the one nobody mentions#

HO-8 is the modified coverage form, and it exists for a problem the rest of the ladder can’t solve: houses where rebuilding costs far more than the house is worth.

Picture a century-old home with plaster mouldings, solid wood panelling, and joinery nobody produces anymore. Restoring it exactly could run $800,000. Its market value is$300,000. No insurer will write replacement cost on that, and actual cash value leaves the owner unable to repair anything.

HO-8 splits the difference: it pays for functionally equivalent repair using modern materials. Your plaster comes back as drywall. It’s the vintage-car problem — the original parts aren’t manufactured any longer, so the insurer pays to make it work, not to make it identical.

It’s uncommon precisely because the situation is uncommon — old and historic houses, mostly. But it’s a real form, and it really is a homeowners policy, which matters more than it sounds: it doesn’t appear on every reference table of the HO family, and that’s exactly the kind of fact a tidy whitelist gets wrong.

What the numbering doesn’t tell you#

Here’s the lesson I actually took away, and it generalizes well past insurance.

The ISO forms give you a beautifully small answer space. Homeowners is HO-1, HO-2, HO-3, HO-5, HO-8. Renters is HO-4. Condo is HO-6. Mobile home is HO-7. Dwelling fire is DP-1, DP-2, DP-3. It’s tempting — genuinely tempting — to turn that into a validation rule and reject anything else.

Don’t. Carriers file their own form numbers all the time, especially on the smaller lines. Some products are sold as packages that cover two lines at once. Regional insurers use house numbering that looks like nothing on the list. Every one of those is a real policy, and a whitelist would call every one of them a mistake.

The standard is a baseline, not a boundary. When you find something outside it, the useful question isn’t “is this on my list” — it’s “does this thing behave like a real base policy?” A form carrying one lonely coverage next to another form carrying two hundred is suspicious no matter how legitimate its number looks. A form with an unrecognizable number carrying a full, coherent set of coverages is probably just a carrier doing its own thing.

Which is, in the end, the same lesson the onion and the ladder both teach: the structure is real and it’s worth learning, but it’s a description of how the market usually organizes itself — not a law it has agreed to obey.

Credit where it’s due#

Almost none of the above is something I worked out from the documents. I understand it because Brigitte Viola, an Industry Solutions Consultant on the insurance side of our team, sat with a stream of questions that must have looked absurdly basic from where she stands — why does a coverage and an endorsement have the same name?, is a blank form number a bug?, what even is HO-8? — and answered every one of them with a real-world example instead of a definition.

The onion, the ladder, and every “wait, that follows from who owns what” moment in this article came out of those conversations. Any part of it I’ve got wrong is mine, not hers.

She writes about this work too, from the insurance side of it: AI for Insurance Renewal Review Automation.