Examining the Appraisers' Role in Redlining through Historical Appraisal Narratives
This study utilizes Natural Language Processing and Geographic Information Systems to demonstrate how subjective racial biases in 1930s HOLC appraisal narratives institutionalized redlining, creating enduring economic disparities and wealth inequality in formerly redlined communities that persist in contemporary real estate assessments.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
The Invisible Ink of the Past
Imagine you are trying to figure out why some neighborhoods in a city are thriving with new coffee shops and parks, while others just a few miles away seem stuck in time, struggling with empty lots and lower home values. Is it just bad luck? Or is there a hidden history written into the very streets? This question sits at the intersection of history, economics, and computer science. To understand it, we need to know about three things: Redlining, which was a practice from the 1930s where banks drew lines on maps to mark certain neighborhoods as "risky" for loans, often based on who lived there rather than the quality of the houses; Appraisals, which are the official reports experts write to decide how much a house or neighborhood is worth; and Natural Language Processing (NLP), a type of artificial intelligence that lets computers read and understand human text, spotting patterns in words that humans might miss. Why does this matter? Because the way we value our homes today builds our future wealth. If the rules for valuing homes were written with bias in the past, that bias might still be hiding in the system, quietly shaping who gets rich and who gets left behind.
The Detective Work: Uncovering Hidden Words
In this study, a researcher named Linda Loubert acts like a digital detective, hunting for clues in thousands of dusty, old documents from the 1930s. She is investigating a specific question: Did the words appraisers used to describe neighborhoods actually help create the unfair economic gaps we see today?
Back in the 1930s, a government group called the Home Owners' Loan Corporation (HOLC) sent out appraisers to grade neighborhoods from A (the best, "green" zones) to D (the worst, "red" zones). These grades decided who could get a mortgage and who couldn't. But the appraisers didn't just give a letter grade; they had to write a story explaining why. They filled out forms with a section called "Detrimental Influences," where they listed what they thought was wrong with a neighborhood.
Loubert gathered over 7,000 of these old forms. She couldn't read them all one by one, so she used a super-smart computer program called RoBERTa (a type of AI) to read them for her. Think of RoBERTa as a robot librarian that can scan thousands of books in a second and tell you, "Hey, in these 500 books, the word 'race' shows up a lot, and in those 200 books, they keep talking about 'old buildings'."
The computer sorted the appraisers' comments into categories like "Race and class concerns," "Proximity to industry," "Aging infrastructure," and "Environmental hazards." Then, Loubert compared these old words to what those neighborhoods look like today. She created a "Redlining Resilience Score" (RRS) for each area, which is like a report card measuring how well a neighborhood is doing now in terms of income, education, and home values.
What the Clues Reveal
The results were striking. The study found that the words the appraisers wrote back in the 1930s are still linked to how neighborhoods are doing today.
First, the "grades" themselves matter. Neighborhoods that were graded "D" (red) back then are still doing worse today than those graded "A" (green). But here is the twist: the words the appraisers wrote mattered significantly, though not quite as much as the lowest grades.
When the computer looked at the "Detrimental Influences" section, it found that neighborhoods described with negative language about race and class (like mentioning "Negro infiltration" or "inharmonious racial groups") or proximity to industry (being near factories) ended up with much lower resilience scores today. In fact, the negative impact of writing about "proximity to industry" was nearly as consequential as the penalty for getting a "B" grade, though the penalty for a "D" grade was still about 2.7 times larger than the penalty for those specific words.
The study suggests that the appraisers weren't just describing the neighborhood; they were building a trap. They used language to turn subjective feelings about race and class into official government rules. For example, if an appraiser wrote that a neighborhood had "undesirable elements," that phrase became a reason to deny loans. Over decades, this lack of money meant those neighborhoods couldn't fix their roofs or build new schools, leading to the poverty we see today.
The research also looked at whether this was just about the official grades or if the story the appraisers told was the real culprit. The answer is complex. The grades set the rules, but the appraisers' words provided the "scaffolding" that held those rules up. When the study looked at the words without considering the grades, the negative language strongly predicted lower wealth. However, when the study included the official grades in the same analysis, the independent effect of the words became much smaller for some categories. This suggests that the bias wasn't just in the final letter grade; it was baked into the very language used to justify it, but the grades themselves carried the heaviest weight in determining long-term outcomes.
The Big Picture
This paper doesn't claim that redlining is the only reason for today's inequality, but it proves that the way appraisers wrote about neighborhoods in the 1930s helped lock in those inequalities. It's like if a teacher wrote a note on a student's report card saying, "This student is not smart because of their background," and then, for the next 90 years, every teacher after them refused to give that student a chance to learn, simply because of that old note.
The study shows that the "red" lines on the old maps weren't just lines; they were stories written in biased ink. And unfortunately, those stories are still echoing in our cities today, keeping some neighborhoods from growing while others flourish. The research suggests that to fix these problems, we can't just look at the numbers; we have to understand the words we use to describe our communities, because those words have the power to shape our future.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.