British and American banks put under spotlight for discriminatory mortgage lending practices – a cautionary tale for Australian banks

Criticism of British banks for discriminatory lending practices

There have long been complaints from ethnic minority communities about lack of access to mortgages and credit in the UK. Recently, there has been an increased focus on lenders’ use of AI in the mortgage market. While the use of AI reduces costs and shortens the assessment process, it can introduce inherent biases which prevent certain customers from gaining access. Even though discriminatory characteristics may be excluded, other proximate characteristics which are not illegal may correlate with these protected characteristics resulting in inherent biases despite there being no explicit intention to discriminate. For example, postcodes may be discriminatory if people from a certain nationality tend to reside there.

A government review of bias in algorithmic decision-making published in November 2020 found that AI-powered decisions “are being made in the context of a socio-economic environment where financial resources are not spread evenly between different groups”. As a result, it warned, if financial institutions rely on making accurate predictions about how likely people are to repay mortgage debt, and specific individuals or groups are historically underrepresented in the financial system, there is a risk that these historic biases could be entrenched further through algorithmic systems.

On the other hand, a number of Britain’s biggest banks have been criticised for refusing to give mortgages to self-employed people who received government grants during the COVID-19 pandemic. Banks have been said to categorise individuals who received such grants as high risk, with those working in sectors like entertainment, hospitality and travel being the worst affected. Some of these policies go against assurances from British banking watchdog the Financial Conduct Authority (FCA), and consumer groups such as Fairer Finance are now calling on the FCA to urgently investigate which lenders are penalising borrowers in this way and take action.

American banks held accountable for discriminatory lending practices based on the ‘Disparate Impact’ standard

Instances of mortgage discrimination in the US have been occurring since at least the 1930s, when banks practised ‘redlining’ or denial of financial services to residents of areas based upon the racial or ethnic composition of those areas, either directly or through selectively raising prices. Last year, US watchdog Consumer Financial Protection Bureau (CFPB) shifted its focus towards the use of the “disparate impact” standard to hold major financial institutions accountable for discriminatory lending practices. The “disparate impact” standard is a legal tool which detects policies and practices that may seem neutral on their face but have disproportionately negative impact on protected classes of people.

In January this year, acting CFPB director Dave Uejio released a statement confirming that fair lending enforcement will be a top priority for the CFPB, with a focus on racial equity, but that the CFPB will also “look more broadly, beyond fair lending, to identify and root out unlawful conduct that disproportionately impacts communities of colour and other vulnerable populations”. In its annual Fair Lending Report to Congress released in April this year, the CFPB reported that over the past year it had conducted prioritised assessments in order to gain an understanding of industry responses to pandemic-related challenges, and help ensure that financial institutions were attentive to practices that may result in consumer harm. The CFPB also reported that it had taken enforcement action against two financial institutions and referred four matters to the Department of Justice. CFPB has indicated an intention to place greater emphasis on fair lending and efforts to address racial equity for underserved communities in 2021 and beyond. In response to CFPB’s supervisory and enforcement priorities and rulemaking agenda, consulting firms such as Protiviti have urged financial institutions in the US to evaluate the impact of these heightened regulatory expectations and reconsider the robustness of their compliance management systems.

In June this year, the Biden Administration announced, as part of its strategy to address the racial wealth gap, “a first-of-its-kind interagency effort to address inequity in home appraisals, and conduc[t] rulemaking to aggressively combat housing discrimination.” In his first week in office, President Biden issued a memorandum directing Department of Housing and Urban Development (HUD; a department within the federal government) to address discrimination in the housing market. HUD’s proposed rule on countering discrimination in housing has now been published in the Federal Register. These proposed rules will provide the legal framework for HUD to require private and public entities alike to rethink established practices that contribute to perpetuate inequities.

Fair lending in Australia

In Australia, the Royal Commission into Misconduct in the Banking, Superannuation and Financial Services Industry in 2018 saw an increased focus on responsible lending, and Australia’s financial institutions have since tightened access to credit and imposed stricter conditions on its customers.

Financial institutions in Australia have started to explore the deployment of AI in loan systems to accelerate approval times. In response, in December 2019 ASIC’s updated version of Regulatory Guide 209 ‘Credit licensing: Responsible lending conduct’ reflected the increased uptake of technology use, such as data aggregation services, in verification of a customer’s financial situation. Organisations such as the Australian Human Rights Commission (AHRC) have suggested that the use of AI in home loan assessment could expose financial institutions to claims of bias and legal breaches. Whilst the hidden problems behind data-driven approaches in consumer lending have not yet been investigated by financial regulators such as ASIC, given its focus in other jurisdictions and the warning from the AHRC, Australian lending institutions are advised to tread carefully when utilising AI in home loan assessments.

Analogous to the COVID-related grants given out in the UK, the Australian government provided COVID-19 disaster payments to help keep Australians in jobs and to support businesses affected by the significant economic impact of the pandemic. At its peak, more than 3.8 million Australians were on JobKeeper — approximately three in every 10 Australian workers. On 14 July 2021, the government announced additional support for New South Wales, including an increased and expanded COVID-19 Disaster Payment, as the state’s strict lockdown affects jobs in key sectors, such as hospitality and much of retail. There is yet to be an analysis of the way in which Australian lenders have treated consumers who accessed these payments. If the Australian regulators follow in the footsteps of the UK, there may soon be an investigation into which lenders may be unfairly penalising borrowers.

The lesson for Australia

Discriminatory lending practices are a high priority for the UK and US government and regulators. The use of AI and treatment of COVID-19 disaster payment recipients in consumer lending are particular issues of concern. It is only a matter of time before these issues are investigated by Australian regulators.

Lenders in Australia should get on top of these issues early, and consider such questions as: How proactively are we identifying deficiencies in our lending practices? Are we holistically addressing all types of discrimination? Among other actions, they should review their fair lending programs and servicing practices to determine whether they are operating in a fair manner throughout all phases of the credit lifecycle.