4 ms·

I thought redlining was solved solved. Can you share an example of implicit or explicit redlining (which I interpret to mean realtors, banks, and other institu

I thought redlining was solved solved. Can you share an example of implicit or explicit redlining (which I interpret to mean realtors, banks, and other institutions enacting policies that restrict black people from purchasing homes in certain areas or neighborhoods. Sorry to ask this sort of question, it’s frankly a pet peeve of mine in internet discourse. In this case though I truly am looking for whatever answer(s) you have handy, cheers

acdha 5y ago
Redlining doesn't have a simple solution — after 1980 or so it was probably not the case that someone would outright be turned down for a mortgage but that bakes in a lot of inherited inequality (your ability to qualify for a mortgage in a pricey part of town, or feel welcome in that neighborhood, will vary depending on your race and your parents' economic status) and there were a lot of dubious practices which followed boundaries based on those historic redlining practices. For example, during the previous mortgage bubble the people targeted by predatory lenders and foreclosed on rather than negotiated with were disproportionately non-white. Just recently there was a lawsuit against Wells Fargo alleging unfair lending practices and given how many examples there have been over the years of different credit scores or loan terms for people who appear quite similar financially, I'd be surprised if there wasn't a fair amount of truth to it: https://www.bloomberg.com/news/articles/2022-03-19/wells-fargo-sued-by-black-borrower-for-refinance-redlining https://www.bloomberg.com/news/articles/2022-03-19/wells-far... The interesting question is whether this is intentional human discrimination or more subtle things like algorithms laundering old historical data. For example, if someone was training a model on borrowers' applications and credit histories versus loan default rates it would not surprise me at all to learn that it picked up on details like traditionally Black neighborhoods, colleges, societies, etc. and used that to “objectively” give a higher risk score because the algorithm doesn't have a way to recognize that many old outcomes were due to things like discrimination at work (“last hired, first fired”), being offered less money for the same job, higher likelihood of bad judicial outcomes (e.g. disparate prosecution rates & sentencing for things people of all races do like use marijuana), etc.

acdha 5y ago

Redlining doesn't have a simple solution — after 1980 or so it was probably not the case that someone would outright be turned down for a mortgage but that bakes in a lot of inherited inequality (your ability to qualify for a mortgage in a pricey part of town, or feel welcome in that neighborhood, will vary depending on your race and your parents' economic status) and there were a lot of dubious practices which followed boundaries based on those historic redlining practices. For example, during the previous mortgage bubble the people targeted by predatory lenders and foreclosed on rather than negotiated with were disproportionately non-white. Just recently there was a lawsuit against Wells Fargo alleging unfair lending practices and given how many examples there have been over the years of different credit scores or loan terms for people who appear quite similar financially, I'd be surprised if there wasn't a fair amount of truth to it: https://www.bloomberg.com/news/articles/2022-03-19/wells-fargo-sued-by-black-borrower-for-refinance-redlining https://www.bloomberg.com/news/articles/2022-03-19/wells-far... The interesting question is whether this is intentional human discrimination or more subtle things like algorithms laundering old historical data. For example, if someone was training a model on borrowers' applications and credit histories versus loan default rates it would not surprise me at all to learn that it picked up on details like traditionally Black neighborhoods, colleges, societies, etc. and used that to “objectively” give a higher risk score because the algorithm doesn't have a way to recognize that many old outcomes were due to things like discrimination at work (“last hired, first fired”), being offered less money for the same job, higher likelihood of bad judicial outcomes (e.g. disparate prosecution rates & sentencing for things people of all races do like use marijuana), etc.