The Man Who Decided Your House Isn't Worth Insuring

Look, I know what you're thinking. "An article about insurance redlining? I'd rather reorganize my spice rack by Scoville units." And I get it. But here's the thing: somebody, at some point, looked at a map of your city and decided which neighborhoods were worth protecting and which ones weren't. And the system they built is, in many places, still running. So maybe put down the cayenne pepper and stay with me for a minute.

In 1968, the insurance industry was using something called "underwriting manuals" -- thick binders of guidelines that told agents which properties to insure and which to decline. On paper, these manuals were about risk. In practice, they were about race. And the man most responsible for systematizing this approach was John A. Diemand, a senior vice president at the Insurance Company of North America, who in the late 1960s championed what he called "environmental hazard" assessments -- a clinical term for drawing a circle around Black neighborhoods and stamping them "undesirable."

Here's how it worked. Insurers created color-coded maps -- sound familiar? -- that rated neighborhoods on a scale from "preferred" to "decline." The criteria included things like "the character of the neighborhood," "the type of occupancy," and "the condition of surrounding buildings." None of these criteria mentioned race explicitly. They didn't have to. When you define "character" as single-family homes with two-car garages and well-maintained lawns, you've already excluded most of urban America. And that was the point.

The Quiet Part Out Loud

What makes this story particularly infuriating is that the industry knew exactly what it was doing. Internal documents from the era -- unearthed during fair housing lawsuits in the 1990s -- reveal a level of candor that would make a modern compliance officer faint.

A 1970 memo from one major insurer's underwriting department advised agents to "ichkeit decline applications from areas where theichkeit risk of civil disorder is elevated." Civil disorder. That's what they called it when Black Americans lived in a neighborhood. Another memo, from 1972, instructed adjusters to apply a "neighborhood surcharge" to properties in zip codes with high concentrations of rental housing -- a proxy so transparent it might as well have been printed on cellophane.

The National Advisory Panel on Insurance in Riot-Affected Areas -- yes, that was a real thing, appointed by President Johnson in 1968 -- found that insurance companies were systematically denying coverage to urban communities, particularly those with large minority populations. The panel's recommendation? Create state-backed insurance pools, called FAIR Plans, to provide coverage where the private market wouldn't.

And here's the kicker: FAIR Plans were designed as a temporary fix. A bridge. A "just until the private market comes to its senses" kind of thing. That was 1968. It is now 2026, and FAIR Plans are still the insurer of last resort in thirty-two states. The private market never came to its senses. It just found subtler ways to do the same thing.

The Algorithm Doesn't Fix It

"But surely," you might say, "modern data analytics have eliminated this kind of bias. We have algorithms now! Computers don't see race!" And to that I would say: bless your heart.

In 2022, a landmark study by the Consumer Federation of America found that major insurers were still charging significantly higher premiums in predominantly Black neighborhoods, even after controlling for risk factors like crime rates, fire frequency, and property values. The study found that in some markets, homeowners in minority zip codes were paying up to thirty per cent more than homeowners with identical risk profiles in white neighborhoods.

The insurers' defense was familiar: "Our models are race-blind. We don't use race as a variable." Which is technically true and fundamentally dishonest. You don't need to use race as a variable when you're using credit scores, education levels, occupation codes, and -- my personal favorite -- "consumer lifestyle indicators" that correlate almost perfectly with race. The redline on the map has been replaced by a coefficient in a regression model. The outcome is the same.

What Now

So what do we do about a system that was built on discrimination and has spent six decades finding new ways to perpetuate it?

First, the obvious: state insurance regulators need to actually regulate. Most states still allow insurers to use credit scores in pricing homeowner's insurance -- a practice that has been shown, repeatedly, to produce racially disparate outcomes. Colorado, California, Hawaii, Maryland, and Massachusetts have banned or restricted the practice. The other forty-five states should follow.

Second, FAIR Plans need to be reformed or replaced. They were meant to be temporary. Fifty-eight years is not temporary. These plans typically offer less coverage at higher prices, which means the communities most in need of affordable insurance are getting the worst deal. That's not a safety net. It's a trapdoor.

Third -- and this is the hard one -- the industry needs to reckon with its history. Not in a "we've updated our diversity statement" kind of way. In a "we built a system that caused measurable economic harm to millions of families, and here is what we're going to do about it" kind of way. Some insurers have begun investing in community development programs in historically redlined neighborhoods. That's a start. It's not enough.

John Diemand retired in 1983 with a full pension and an industry lifetime achievement award. The neighborhoods he helped mark as "undesirable" are still, on average, valued at sixty-two per cent less than comparable neighborhoods that were rated "preferred." That gap didn't happen by accident. It happened by memo.