This week, people in Indianapolis paddled kayaks down their own streets. The White River rose higher than it has in more than thirty years; at least seven people died. And in the coming days, thousands of families will discover a second disaster — this one on paper. The insurance on their home does not cover floods. It never did: a standard American homeowners policy excludes flood damage entirely. For that, you need a separate policy. In the county that just went underwater, one home in a hundred and thirty-six has one. The other ninety-nine percent have nothing.

Here's the design behind that number. In 1973, Congress tied mandatory flood insurance to a line on a federal map. Inside the line, your bank will not close a mortgage without a flood policy. Outside the line — silence. And everyone reads silence as safety. Since then, the rain has changed. The line, mostly, has not. Today, about a third of America's flood claims come from outside it.

Starting on the eleventh of August, parts of Indiana took more than eleven inches of rain in three days. The White River crested at a record twenty-four point nine feet in Anderson — the worst flooding the state has seen in over thirty years. At least seven people died, among them a four-year-old boy and an eighteen-year-old student on his way to his first semester. More than three hundred and fifty people were evacuated; at the peak, a third of a million homes had no power; in Indianapolis, a section of road simply collapsed. The president signed an emergency declaration the same weekend.

Now the second wave, the paper one. Fewer than one percent of Indiana households carry flood insurance — about sixteen thousand policies for two and a half million homes. Everyone else is left with federal disaster aid, and that aid has a ceiling: this year, roughly forty-four thousand dollars per household. The average flood insurance claim, for comparison, pays out about eighty-three thousand. Federal help is real. It is also a fraction of a ruined house.

The usual story writes itself: freak weather, terrible luck. But hold that against the record. Houston, 2017 — hurricane Harvey floods the city, and about eighty percent of the damaged homes turn out to be uninsured for flood. Kentucky, 2022: same discovery. Vermont, 2023: same again. Indiana, this week: same. When the same surprise repeats in state after state, year after year, it stops being a surprise. It's a design. And this particular design lives in two laws and one map.

Start with why flood insurance is separate at all. A fire burns one house; a flood takes the whole street in one night. Losses that arrive all together are the kind private insurers cannot carry, and by the 1960s they had largely walked away from flood. So in 1968, Congress built a public insurer for it — the National Flood Insurance Program. One problem remained: almost nobody bought the policies voluntarily. So in 1973, Congress added the push. Mandatory purchase — but a duty needs a boundary, and Congress pinned it to the federal flood map: the Special Flood Hazard Area, the zone with a one-percent chance of flooding in any year. Inside that zone, any mortgage touching the federal system — which is most mortgages — requires a flood policy, checked at closing, enforced for the life of the loan. Outside the zone: no duty, no check, nothing.

On paper, that's elegant: force insurance exactly where the risk is. But watch what the line does to people. Inside it — and the zone holds only a small slice of a state's homes to begin with — people buy, because they must. Outside it, the absence of the requirement becomes information. The government looked at my street and didn't make me insure — so my street must be safe. The map stops being a measurement and becomes a verdict. House prices read it. Builders read it. Families read it.

Which would work — if the line were true. It isn't, in two ways. First, it's old. Many of the maps were drawn decades ago, built mostly around rivers and coasts; the violent short rainstorms that drown a street in an hour are barely in them, and a changing climate isn't in them at all. When researchers redrew the maps with modern rainfall data, the count of high-risk American properties went from FEMA's seven point nine million to seventeen point seven million — more than double. Around ten million homes carry real flood risk and sit outside the official line. FEMA's own claims history says it simply: over the last decade, twenty-nine percent of flood claims came from outside the high-risk zones. Roughly a third of the water ignores the map.

Second, the insurer behind the line is itself underwater. The program owes the U.S. Treasury more than twenty-two billion dollars; it collects about four billion a year and pays out nearly six. Congress forgave sixteen billion of its debt in 2017 — the hole simply started refilling. The whole machine lives on temporary extensions — the thirty-fifth expires this September — and during a government shutdown last year it lapsed for forty-three days: in the entire United States, no new flood policy could be written at all. So: the duty is pinned to a line, the line has lost sight of millions of at-risk homes, and the fund behind the line runs at a loss nobody wants to price honestly.

Design one: France. Since 1982, every property policy carries a mandatory catastrophe surcharge — flood, drought, earthquake — backed, in the end, by the French treasury. Nobody is uninsured, because nobody can opt out. The currency is the risk signal: a dry hilltop subsidises a riverside villa, building in a floodplain costs nothing extra, and the climate bill arrives through the back door — the surcharge went from 5.5% to 12% to 20%, after eight straight years of deficit. Design two: Britain. A levy on every home policy funds Flood Re, which keeps flood-prone homes insurable — with two fuses: homes built after 2009 are deliberately excluded, and the whole scheme self-destructs in 2039. The currency is the holes and the countdown — and 11% of new English homes are still built in flood-risk areas anyway.

Design three: America's experiment on itself — tell the truth. Since 2021, every policy is being repriced to the house's actual risk, not its zone. The currency is what truth costs: riverside premiums multiply, Congress files bills to cancel it, owners drop coverage so the gap widens — and the price itself quietly says of some streets, do not rebuild here. Notice the triangle: everyone a little with the signal muted; the risky few subsidised, with fuses and a timer; or each house carrying its own truth, and the revolt that follows. Every corner has a defender. The one design with no defender is the current one: a duty pinned to an aging line that a third of the water no longer reads.

Come back to the kayaks on an Indianapolis street. The water did not check the map on its way in. The families outside the line weren't reckless — they were told, by the silence of the system itself, that they were safe. That's what a line does when a law is pinned to it: inside, it protects; outside, it reassures. And reassurance, unlike protection, costs nothing until the day it costs everything.

So the next time a flood is in the headlines — and there will be a next time — the revealing questions aren't about the rain. They're about the line. Where is it drawn? When was it last redrawn? And who pays on the day the water crosses it? Your country has lines like this too, whatever it calls them — seismic zones, fire maps, coastal setbacks. Every one of them is a rule wearing the costume of a fact. And rules, unlike facts, have an age.