Around 75% of white American families were homeowners in the first quarter of 2020, according to data from the United States Census Bureau. However, only 44% of Black American families owned their homes at the same time. This gap is larger than it was in 1960, when racial bias in mortgage lending was a matter of policy in many states. Similarly, insights from the Federal Housing Finance Agency demonstrate significant differences in mortgage application approval rates between white Americans and minority applicants. In 2020, for instance, white Americans had a more than 15% higher chance of being approved for a loan based on HMDA data.

Even as other deeply embedded social institutions have become more inclusive of people of color, equal access to housing is seemingly out of the country’s grasp. That is largely because the modern lending ecosystem is underpinned by the very technologies and processes that were originally designed to facilitate discrimination rather than prevent it. Fair lending laws alone have been unable to narrow the yawning gap in homeownership rates.

With this in mind, how do we prevent homeownership inequities from persisting well into this century?

For many, the answer to that question is simple: We would use artificial intelligence. After all, credit score algorithms and automated mortgage application reviewers do not see skin color. In practice, however, integrating unbiased AI-powered technologies into a biased lending system presents significant obstacles.

The Human Obstacle

Human lenders have both implicit and explicit biases that might affect what products they offer to prospective borrowers and how they view and evaluate loan applications. The promise of AI-driven advanced analytics is that it could remove those biases from application assessments and instead focus solely on the facts.

First, we must recognize that AI in general is more promising than reality in many fields, and it might take time before an AI system could reliably make credit determinations. Moreover, AI development often perpetuates current processes rather than creating a new way of thinking.

Traditional mortgage lending metrics (e.g., credit scores) are focused on identifying borrowers who are simply likely to default in the short term. With or without AI, new approaches to evaluating credit could deliver insights on underlying borrower characteristics that are more indicative of long-term creditworthiness. By evaluating a wider range of data (in addition to borrower credit history, income, property value, and other relevant information), we can begin to create a system that supports homeownership rather than discourages it. Nevertheless, that will not happen by accident.

The Power of Purpose

Although policy has struggled to create a fair and inclusive path toward homeownership, technology can still succeed — but only if its creators deliberately include the elimination of racial bias in mortgage lending as a key success metric. Industry innovation does not necessarily result in progress for all Americans; achieving greater equity will be difficult if we do not pursue it directly.

As such, the data scientists and developers building AI technologies today must use decidedly different approaches from those employed in the past. Specifically, they should be trained with data that ensures less consistency with “historical” approaches in pursuit of greater equity for all.

What does this mean? In short, we have centuries of data on loans and mortgages issued in the past, but given the inequities we are seeing, it is likely not representative of a system that enables equal access to housing. Rather than training algorithms using just historical credit data, we should envision a more just scenario and incorporate data that corresponds to a more equitable world. A reliance on flawed inputs will only exacerbate existing biases.

Of course, the surge in companies such as Rocket Mortgage and other online lending platforms has already reduced the need for in-person applications, which in turn has helped to reduce discrimination significantly. Nevertheless, as an industry, we cannot assume that technology alone will solve a problem that is centuries old.

The Human Solution

Today, there is a tremendous amount of data available to power algorithms that make the mortgage lending process more streamlined and efficient. The question, now, is this: Can that data also help create equal access to housing for Americans of color? 

Not without our help and our focus on addressing inequality.

The AI technologies and mortgage risk analytics that anchor tomorrow’s lending processes must be designed with the express goal of eliminating racial bias in mortgage lending. Misuse of data can promote additional discrimination, and failing to keep this in mind will result in new technologies that simply do not account for the struggles of marginalized communities. To avoid perpetuating systemic inequities that have spanned generations, actual intelligence is needed.

A magical solution does not exist — but by combining the power of human foresight with the power of emerging technology, we could certainly take many more steps forward.

According to a report by the Research Institute for Housing America, climate change risk is rapidly increasing in the housing industry and will continue to demand more attention and regulation in the near future.

Climate change will impact risk factors in the housing industry in nearly every corner of the globe. Wildfires are becoming more common and the area they ravage more extensive. Hurricanes and severe storms are happening with more frequency and severity. Potential damage from excess heat and droughts elevates risk to properties every day.

However, flooding is currently one of the highest risk factors posed to the housing industry. Many housing areas are used to the idea of flood risk and are adequately prepared and protected, but many properties that were never at risk before are now in the danger zone. The housing market is currently in a vulnerable position.

When Floods Outpace Insurance Policies

Depending on geography, more properties without previous flood risk are increasingly likely to experience flood damage. Homes and communities that were erected in floodplains are used to the protocols: Safety procedures such as evacuating or securing the area, working with insurance to cover damage, or receiving aid to rebuild when possible.

Under the National Flood Insurance Program, homes that are federally backed by programs such as Freddie Mac and Fannie Mae in certain areas require the owners to carry flood insurance. Those homes are located in floodplains that are defined based on a 100-year flood probability.

The problem is that floodplain boundaries are rapidly changing. One hundred years' worth of flooding data is not as relevant as it used to be when flooding zones are becoming more and more volatile. Even as the area of potential flooding damage overflows into neighboring regions, the floodplain boundaries have not been redrawn recently enough to impact flood policy uptake.

That means many homes that are at risk of future flooding are not likely or required to carry flood insurance. Experts are predicting that the National Flood Insurance Program will be stretched to its limits very soon, and that banking and insurance regulation will need to act quickly to spread and manage climate-related risk. It's possible that soon, the total cost of owning homes will outpace the value of the home.

This becomes very concerning when we consider the likelihood of mortgages going unpaid; a lack of flood insurance then quickly becomes not just a housing risk but a credit risk for the owners and an economic risk for the country if housing prices plummet and people’s debts begin to far outvalue their assets.

Updating Risk Calculations on Climate Change Analysis

Firms currently vary in their preparedness to face climate change insurance risk. As data becomes more advanced, some firms have begun to license property-level climate risk data, and specialist analytics firms are appearing with expertise in climate models.

The Fed and the SEC are also trying to adapt regulations to fit the new (and ever-changing) reality of climate change risk. There are new committees dedicated to assessing climate change analysis and determining systemic risk to the entire financial world, including the Supervision Climate Committee. These regulators will need updated methods to quantify risk and mandate disclosure, but for now, changes are nascent and firms will have to add their own experience to the bank of loss exposure research.

Financial firms are facing — or are about to face — considerable pressure from investors, governing and regulatory bodies, and insurance and banking regulators concerning the way they calculate risk. They will probably also feel some pressure from employees and workers in the financial sector, who are becoming increasingly alarmed about the impending disruption of climate change.

Firms will need to manage climate risk alongside their broader risk management strategy. For that to work, they’ll need to understand climate change data and the set of exposure scenarios that are relevant to them. For example, McKinsey predicts that about one-third of the planet's land will be affected by climate change. In addition, flooding exacerbated by climate change is expected to double the damage to capital stock by 2030.

Financial institutions urgently need to understand how to calculate and explain the risks posed by climate change, both for their own risk management strategies and for stakeholders. Quantifying climate change risk will be an evolving science. Property portfolios will require new risk scores based on the potential hazards that climate change will bring. Those scores will then need translating into commonly used financial measures, such as credit risk, market risk and prepayment risk.

As financial firms wait for regulatory approaches to become clear, they will need to continue to educate themselves and to remember that climate change models will shift rapidly — the best climate change risk strategy will be the one that is most able to change.

Mortgage market participants are keenly aware that the Federal Reserve has been scaling back its UST and MBS purchases and factoring the outcomes of its actions on stakeholders across markets. In this Policy Perspectives article, we take a retrospective look back at the March 2020 ease from a mortgage markets view point and highlight how the outcomes of intervention manifest through the interaction of related primary and secondary mortgage markets activities. We show that Fed activity can have unintended and disruptive impacts on the functioning of housing finance and result in wealth effects that benefit the more affluent segments of the housing economy.

Read Now

The growing prevalence of artificial intelligence in the mortgage industry is shining a new light on the human biases that have pervaded the industry since its inception. AI is meant to bring fairness and objectivity to mortgage decisions, but it can’t perform fairly if it was built on an unfair system.

In particular, racial bias in mortgage lending is a prevalent issue. The homeownership gap between the Black and white populations has remained relatively unchanged for more than a century, and today, it’s as wide as ever. Moreover, Black borrowers were 2.5 times more likely to be rejected for a home loan last year than their white counterparts — and that data does not account for applicants who ended up not making a home purchase.

Equipping lenders with more software and better algorithms will not reduce this gap. Before AI can be deployed effectively as a tool for positive change in the mortgage industry, a widespread shift in perspective must take place.

Importantly, lenders must change their definition of who qualifies as a “good” or successful homeowner in order for AI to operate with true objectivity. To reduce inequity in the mortgage industry, lenders need to change the question from “Who is delinquent?” to “If someone is delinquent, what can cure the delinquency to ensure long-term success?”

The Delinquency Dilemma

Historically, lenders have relied on delinquency as an influential metric when assessing borrower capacity and have (both consciously and unconsciously) equated it with the moral worth of mortgage applicants. In the midst of increasingly numerous and devastating natural disasters and the ongoing COVID-19 pandemic, however, the delinquency metric has come under scrutiny.

As an indicator of potential success in mortgage fulfillment, delinquency is not an accurate representation of a borrower. It is increasingly being understood as a result of circumstances, and not necessarily the result of a person’s ability to own a home.

A credit score, for example — which is based on measures of delinquency — is not a viable indicator of a person’s long-term ability to afford a car or home. Still, it will exert a disproportionate influence on the costs of borrowed capital, which are often prohibitive for BIPOC mortgage applicants.

If nothing else, the social, political, and economic uncertainty that has characterized the past several years has shown that delinquency alone cannot be a viable metric. As people around the world dealt with the pandemic, a halting economy, and disruption in nearly every aspect of life, it became clear that delinquency simply was not a relevant differentiating metric.

It’s also important to realize that circumstances resulting in delinquency have historically impacted people of color disproportionately. According to the Consumer Financial Protection Bureau’s May 2021 report on the characteristics of mortgage borrowers through COVID-19, BIPOC homeowners faced higher rates of delinquency and forbearance than their white counterparts. Specifically, Black and Hispanic borrowers account for only 18% of all mortgage borrowers, yet these groups represented 33% of mortgages in forbearance and 27% of the mortgages that were delinquent.

There are numerous social, economic, and political factors that impact why BIPOC communities are affected more heavily than others in extenuating circumstances. To begin with, BIPOC families have historically had less generational wealth. According to a September 2020 report from the U.S. Federal Reserve, white families have eight times more wealth on average than Black families, and five times more wealth on average than Hispanic families.

If the industry continues to use the same metrics that exacerbated this wealth disparity in the first place, then equity in lending will always be out of arm’s reach.

Progressing Toward Equality

Thankfully, the wider perspective has begun to shift over the past few years. Rather than punishing delinquent borrowers with additional fees or removing them from their homes, lenders are seeing the value of assisting homeowners so they can remain in their homes over the long term. After all, penalizing short-term financial hardship is not as profitable as helping a borrower successfully complete payments over the course of the mortgage.

As such, lenders are beginning to focus on different types of metrics, which will have important (and positive) implications for mortgage decisions and even AI-led mortgage analytics.

Increasingly, lenders are realizing that forbearance, loss mitigation, income disruption assistance, and other approaches are far more effective when it comes to extending homeownership. They’re considering attributes that might make borrowers more likely to re-perform if given some leeway as well as the systems that will be needed to ensure temporary setbacks are rectified.

This is a massive step in the right direction. As lenders continue to shift their focus toward metrics of sustainable homeownership instead of delinquency, the hurdles these borrowers face should become smaller.

That said, AI-powered lending tools must be deliberately and thoughtfully designed around those metrics, and with the intention to create a more equitable system. Otherwise, technology will reinforce old ways of thinking — and racial bias in mortgage lending will persist.

The LDM v3.0.2 library adds AutoLDM to the v3.0.1 library.

Key benefits include:

• AutoLDM is a loan-level model that produces monthly default, prepayment, severity, balance, and delinquency projections.
• The projections are sensitive to individual borrower attributes (e.g., credit score, contract rate, loan term, delinquency status) and vehicle characteristics (e.g., vehicle age, type, new/used).
• The model utilizes a delinquency state transition framework to model the migration of the borrower conditional on their attributes and the unemployment projections.
• AutoLDM covers the full credit spectrum of loans from subprime through prime borrowers.
• Extensive support of AutoLDM is available from experienced modelers.
• Model validation documentation is available.

AutoLDM is available via the LDM library and through the AutoKinetics application. For a full list of updates, read our LoanDynamics Model v3.0.2 Release Notes.

We are working closely with our third-party vendors on the integration of this release into their platform. For more information about the availability of this release through your vendor system, please contact michelle@ad-co.com. For all other requests, please contact support@ad-co.com.

Release notes for all our products are available at https://www.ad-co.com/support/release-notes. 

To access AutoLDM demo on-demand, please click here.