At the time, the average rate for a used auto loan was about 5 percent. I had good income and exceptional credit scores. After pulling my credit report, the salesman said he could offer me 10 percent.
"That's awfully high," I said. (The salesman didn't know I had already been pre-approved through my credit union for a rate of about 4 percent.)
I suspected what was happening. I had reported on class-action lawsuits brought against auto-finance firms alleging they allowed dealers to charge interest-rate "markups" to black and Latino customers more frequently, and at significantly higher rates, than similarly situated white customers.
I kept pressing the salesman about what I felt was an interest-rate overcharge.
"I guess you have some blemishes on your credit," he said.
Oh, no, he didn't.
It took an amazing amount of self-control on my part not to cuss that man out. I was sure he was racially profiling me. It's happened many times before. He was counting on two things: I was unaware of my good credit standing, and I might not be inclined to shop around for a better rate. I only buy cars with cash now.
There is still a two-tier pricing system that penalizes blacks and Latinos who are creditworthy of better rates.
It happens with auto loans as well as mortgages.
But what if you took the human element out? Surely, minorities would get as good a deal as other borrowers with similar credit histories with a supposedly race-neutral automated application system, right?
No. Even with automation, mortgage bias exists, according to recent findings by researchers at University of California, Berkeley.
The Berkeley study looked at 30-year, fixed-rate, single-family residential loans issued from 2008 to 2015 and guaranteed by Fannie Mae and Freddie Mac. It found that -- whether they apply for a mortgage loan face-to-face or online -- blacks and Latinos are charged 5.6 to 8.6 basis points higher than white borrowers with comparable credit history. The extra mortgage interest is costing minority borrowers $250,000 to $500,000 more per year.
"Algorithms have not removed discrimination, but may have shifted the mode," the researchers wrote.
When refinancing, minorities pay 1 to 3 basis points more. Interest-rate discrimination can boost lender profits 11 percent to 17 percent per loan.
So, how can a machine discriminate?
It's all about the formulas fed into the algorithms, the researchers conclude.
"With algorithmic credit scoring, the nature of discrimination changes from being primarily concerned with human biases ... to being primarily concerned with illegitimate applications of statistical discrimination," the researchers wrote.
By law, lenders can't discriminate based on race. They can however price loans based on credit risk. With an online application, the algorithms may be categorizing borrowers based on behavioral ethnic or demographics profiling, the researchers said.
"Even if the people writing the algorithms intend to create a fair system, their programming is having a disparate impact on minority borrowers," said study co-author Adair Morse, a finance professor at UC Berkeley's Haas School of Business.
Here's a key point the study made: Statistical discrimination could be occurring because minority borrowers are less likely to rate shop.
However, it's not just minorities who don't shop around for mortgage rates. Nearly half of consumers don't comparison shop for better rates before taking out a mortgage to buy or refinance a home, according to a study earlier this year by Freddie Mac.
On average, borrowers could save almost $1,500 over the life of the loan by getting one additional rate quote and about $3,000 for five quotes, Freddie Mac said. For minorities who are up-charged, the dollar savings would be even greater.
By the way, the scoring models are better now at allowing consumers to rate-shop for a single loan. For its newest versions, FICO says all inquiries done within a 45-day window for auto, mortgage or student loans are treated as a single credit inquiry.
Since bias still exists, decrease the likelihood that you’ll be discriminated against by shopping around and being better informed about your creditworthiness.