Conference reveals how discrimination still harms communities of color

Consumer Action was proud to co-sponsor the Understanding Segregation: Tools to Advance Access and Opportunity conference hosted by Fair Housing Advocates of Northern California on April 1 in San Rafael. The conference provided an opportunity for local stakeholders to come together to address pressing fair housing issues affecting communities of color both in the San Francisco Bay Area and nationwide. Consumer Action’s Audrey Perrott was in attendance.

“Fair Housing Advocates of Northern California provides a vital opportunity for advocates to have open forums to discuss racism, segregation and the racial wealth gap,” Perrott said. “Racism still persists in every institution in America. We must continue to work to provide increased economic opportunities and promote equal access to housing.”

The conference featured keynote addresses from Stephen Menendian, assistant director and director of research at the Haas Institute for a Fair and Inclusive Society at the University of California, Berkeley, and Lisa Rice, president and chief executive officer of the National Fair Housing Alliance (NFHA).

Menendian presented findings from the Haas Institute’s report Racial Segregation in the San Francisco Bay Area. The report reveals how segregation is embedded in American culture, even though America is growing in diversity. According to Menendian, 59 percent of black and white Americans would need to move into new neighborhoods to create integration in the U.S.

According to the report, segregation between white and Latino populations has risen dramatically since the 1970s, particularly in the San Francisco Bay Area, while segregation between white and Asian populations has also increased significantly since the 1980s. Black and white segregation and black and Latino segregation have decreased, however.

Menendian also discussed how the New Deal, Fair Housing Administration (FHA) and Veterans Administration (VA) programs were all designed to create wealth and homeownership for whites while marginalizing blacks and other people of color. Tens of millions of home mortgages were insured with federal subsidies—mortgages that were often denied to people of color. This phenomenon allowed whites to accrue wealth while others did not, and partly explains the wealth gap today. On average, white people now have $14 to every $1 black people have. This wealth is distributed through the system over time in different ways: Parents give money and gifts to their children, allowing them to put downpayments on homes or pay for college.

After Menendian summarized the overarching demographic data, Lisa Rice introduced the concept of “weaponized data” and how it impacts the underrepresented communities that advocates serve daily. Financial services and the housing industry often rely on computers, automated systems and algorithms to make choices about whom to offer home loans to and at what price, arguing that technology provides a better level of efficiency and that this is passed on to the consumer in lower costs.

The challenge, Rice said, is that the public, and even industry, believes that computer data is innocuous and that “computers are colorblind.” This is a dangerous notion, and, unfortunately, industries are building artificial intelligence (AI) systems and not asking the right questions about the data being fed into the system to “teach” the machines how to make decisions. Data inputs have the very real potential to cause and perpetuate bias. According to Rice, industry must consider where data is coming from, if it has been “scrubbed and cleaned” of existing errors and how it has been scrubbed and cleaned, before it is put into the system.

Rice pointed out that AI decision-making introduces a profound opportunity for the perpetuation of discrimination and inequalities at a massive level, particularly in the housing industry, as computers are now able to offer automated verification to process if a person is employed, their rental and banking information, their credit report and more. The data the computer has access to during the underwriting process determines how much “home” a consumer can buy and at what interest rate, which also impacts where they can buy.

Even seemingly innocuous data inputs can harm consumers. Rice provided the example of an address, which is one piece of data that exists in a computer system and is analyzed by AI during the underwriting process. An address is not just a fixed point on a map: Humans and computers (operating in accordance with the rules or algorithms that humans have designed for them) can know and/or infer a lot about a person based on an address. This includes the person’s income and whether they rent or own a home currently, their chance of graduating high school, if they have graduated from college, their credit score, how much wealth they’ve accumulated, and even their life expectancy and chance of contracting a deadly disease—all due in part to the highly racialized neighborhoods in which we live. Machines have been programmed to make inferences based on factors like whether an address is located in a neighborhood that is a food desert, a credit desert, an education desert, an area of economic growth or opportunity, or an area where police have a high level of engagement.

“It is important for stakeholders to understand the ramifications of segregation,” Perrott said. “Segregation is not just about groups of people living together by ethnic composition. It is also about a systemic lack of access to opportunity and resources, racial inequality, poor health outcomes and more.”

The conference concluded with a panel discussion on the racial wealth gap, remedies and reparations. Panelists included: Anne Price, president of the Insight Center for Community Economic Development; E.J. Toppin, a staff researcher at the Haas Institute for a Fair and Inclusive Society at UC Berkeley; and Christopher Elmendorf, a professor at the UC Davis School of Law.