Actually sitting down and talking to a mortgage officer has been replaced by online applications and algorithms but discrimination is still very much alive and well.
A new study by the University of California, Berkeley discovered that online and face-to-face lenders charge higher interest rates at 11 to 17% to minorities including African Americans and Latinos.
Those home buyers are paying half a billion dollars more in interest each year than white borrowers with the same credit scores.
The researchers believe these findings have raised serious legal questions about the statistical discrimination in the financial sector and point to widespread violations of the U.S fair lending laws.
While, historically, discrimination has been caused by human prejudice, pricing discrepancies are increasingly the result of algorithms that use machine learning to target applicants who might not shop around, due to higher-priced loans.
In other words, human prejudices are being replaced by algorithm prejudices that seem to practice lending discrimination.
Adair Morse, study co-author and professor at UC Berkeley's Haas School of Business, said that even if people writing the algorithms intended to create a fair system, their programming is having a huge impact on minority borrowers.
In other words, it's discriminating under the law.
The Challenges Facing Researchers:
The leading challenge for studying discrimination has been the only large data source that covers race and ethnicity which just happens to be the Home Mortgage Disclosure Act (HDMA) which covers approximately 90% of residential mortgagees but does not have information on loan structure and the types of property.
Using machine learning, researchers combined HDMA data with three other large datasets ATOM, McDash, and Equifax which is the first time ever.
It also includes details on interest rates, loan terms and performances, property location, and borrower's credit, along with race and ethnicity.
Researchers at the Haas School of Business and Berkeley Law concentrated on 30-year, fixed-rate, single-family residential loans issued between 2008 and 2015 that were guaranteed by Fannie Mae and Freddie Mac.
This pretty much means that all loans in the pool were backed by the U.S government which followed the same rigorous pricing process based on only a grid of loan-to-value and credit scores that were put in place after the financial crisis.
Because private lenders are protected from default by the government guarantee, any other variations in loan pricing would be caused by lenders' competitive decision making.
The researchers could isolate the pricing differences that correlated with race and ethnicity apart from credit risk.
The analysis discovered significant discrimination from both face-to-face and algorithmic lenders were as follows:
Black and Latino borrowers paying up to 8.6% s higher interest rates on loans than Whites and Asians and more on refinancing loans.
For borrowers, these discrepancies cost them $250M to $500M annually.
For lenders, this comes to 11 to 17% higher profits on purchase loans to minorities based on the industry average 50-basis-point profit on loan issuance.
Professor Morse said the results are consistent with lenders using big data variables and machine learning to infer the extent of competition for customers and price loans.
The pricing could be based on geography including areas with fewer financial services or the characteristics of the applicants.
If artificial intelligence can determine which applicants might perform less comparison shopping and accept higher-costs, the lender has created, what Morse calls, algorithmic strategic pricing.
Morse also concurred that there are several reasons that ethnic minority groups may not be shopping around as much or it could be they live in financial deserts with less access to products and more monopoly pricing, or it could be that the financial system creates an unfriendly environment for some borrowers.
He further said that lenders may not be specifically targeting minorities in their pricing schemes but by profiling non-shopping applicants they are targeting them.
Discrimination is totally against the U.S. Fair Lending Laws which are designed to prohibit these kinds of practices.
There are several U.S courts that have ruled that loan pricing differences that vary by price or ethnicity can only be legally justified if they are based on the borrower's creditworthiness. The novelty, we can rule out the possibility that these pricing differences are due to differences in credit risk among the borrowers.
The data collected revealed some good news. Lending discrimination is on the decline.
This suggests that the rise of new FinTech, or Financial Technology, platforms and simpler online application processes for traditional lenders has increased competition and made it easier for people to comparison shop which is great news for underserved home buyers.
The researchers discovered that FinTech lenders did not discriminate on accepting minority applications. But, traditional face-to-face lenders were still 5% more likely to reject minorities.