NCRC Mortgage Market Analysis
October 2024
Photo: © Annap – stock.adobe.com
In 2023, mortgage lending was shaped by rising home prices and a return to historically normal interest rates, creating an unprecedented perfect storm: high values attracted home investors looking for expensive assets and rental revenue, while higher borrowing costs limited the options for homebuyers.
This pinch, which has helped compound the nation’s long-running crisis in affordable housing, is due to the over-financialization of housing in America. Instead of cooling off home values, prolonged pandemic-era low interest rates spurred a gold rush of investors, both large and small, buying US homes as they borrowed at super low rates against their existing homes. The housing finance system failed to adapt to this power shift from traditional owner-occupant borrowers toward medium- and large-scale landlording operations, leaving vast unmet demand for affordable options like manufactured homes or smaller mortgages.
As a result, the gap between renters and owners is widening. Homeowners are trapped in gilded cages, unable to move due to rising rates, while renters face skyrocketing rents, making it impossible to save for homeownership.
The future hinges on local communities. They need to curb the influence of market-distorting investors and build more housing—denser and cheaper than today’s single-family mini-mansions.
The local changes the housing market requires to function correctly are only possible with corresponding change at wider and higher levels of market activity and oversight. The problem demands a reconfiguration of housing finance.
Once used to fund construction costs, mortgage debt now serves Wall Street’s needs. The financial tail is wagging the household-wealth dog.
This report gives an overview of how mortgage lending changed in 2023.
— Key Takeaways
Regulatory oversight needs to be strengthened to close loopholes that led to a nearly total absence of demographic data from secondary-market purchased loans.
Credit unions increased their market share, particularly in home equity and home improvement lending, as non-bank lenders retreated.
— Key Findings
- Decrease in Mortgage Applications and Refinancing: The number of loan application records reported in 2023 dropped to 11.5 million, down from 16 million in 2022 and 26 million in 2021. This marks the lowest number of applications since Home Mortgage Disclosure Act (HMDA) reporting rules were enhanced in 2018. Refinancing activities decreased by 60% in 2023, indicating a marked shift away from refinancing as interest rates normalized.
- Growth in Lending to Black and Hispanic Borrowers: The share of loans to Hispanic borrowers increased from 12% in 2021 to 13.9% in 2023, with home purchase loans for Hispanic borrowers rising from 14.9% to 16.5%. Black borrowers saw an increase in their share of home purchase loans from 9% in 2021 to 9.3% in 2023, indicating a positive trend towards greater inclusion in the housing market.
Rise in Credit Union Market Share: Their share of home improvement loans rose from 33.1% in 2018 to 46.2% in 2023, and their share of home equity lending increased from 29.2% to 39.9%.
- Impact of Rising Interest Rates and Housing Costs on LMI Borrowers: Rising interest rates and increased housing costs disproportionately affected low- and moderate-income (LMI) borrowers, making it more challenging for them to enter the housing market. These economic pressures have exacerbated existing disparities, with higher property values making homeownership less attainable for LMI individuals and families.
- Investor Activity and Its Impact on the Housing Market: There has been a noticeable increase in investor purchases of single-family homes, particularly in lower-cost units and Sunbelt markets. This trend has made it more difficult for families to transition from renting to homeownership, as investors often outbid potential homebuyers.
Methodology
This analysis examines mortgage lending trends from 2018 through 2023, captured by the Consumer Finance Protection Bureau (CFPB) as a part of the Home Mortgage Disclosure Act (HMDA) dataset. This data covers about 88% of the US mortgage market. We explore the shifts from refinancing to home purchasing and the evolving roles of diverse borrower groups in a transforming market.
In 1989, the Financial Institutions Reform, Recovery, and Enforcement Act (FIRREA) significantly expanded the scope of HMDA by requiring the reporting of data about applicant and borrower characteristics to enable regulatory bodies and the public to identify possible discriminatory lending patterns.
In 2015, the Consumer Financial Protection Bureau (CFPB) issued a rule that further expanded the data requirements under HMDA, including providing the public with information about an applicant’s age, property value, and loan terms. The rule was intended to improve the public’s understanding of how consumers access mortgage credit and to help regulators, researchers, and advocates identify potential discriminatory lending patterns and monitor compliance with fair lending laws.
In HMDA, several definitions are crucial for understanding the landscape of home financing.
- Applicant race refers to the primary race and ethnicity identified by loan applicants on their applications, providing insight into demographic trends in mortgage lending. Beginning with this report, NCRC is eliminating records without demographic data from our calculations of the percent of loans made to specific races. The reasons for this are laid out clearly in our section on purchased lending and missing data.
- Loan purpose categorizes the reasons for borrowing: home purchase for buying a home; refinance for renewing a mortgage without extracting equity; cash-out refinance for refinancing a mortgage while withdrawing equity; home improvement for securing funds to enhance a property; and home equity lending for obtaining a secondary loan for purposes such as debt consolidation or starting a business.
- Loan type is bifurcated into conventional loans and non-conventional loans, the latter including government-backed options like FHA, VA, or RHS loans, which reduce lender risk.
- Lender type distinguishes between the institutions offering these loans, including banks, mortgage companies, and credit unions, each playing a distinct role in the financial ecosystem. Understanding these terms helps illuminate the processes and players in the mortgage industry.
Unless otherwise mentioned, all analysis in this report is limited to applications on owner-occupied, 1-4 unit, site-built homes. Data is downloaded as a flat file from the CFPB and then queried from our SQL database using Tableau. Some variables are created in SQL, such as race and ethnicity, and closing fees. Interest rate and rate spread outliers were identified statistically and eliminated from the dataset where interest rates or rate spread was discussed in this report. Those outliers represented 1.37% of all loan records from 2018-2023.
NCRC employs AI tools to enhance data aggregation, summarization, and analysis. These tools are used under strict human oversight, adhering to NCRC’s Responsible Artificial Intelligence Usage & Author Responsibility Policy. All AI-generated outputs are reviewed to align with NCRC’s ethical standards, with authors retaining full accountability for accuracy.
Introduction
While COVID-19 remains a danger to public health, the acute shocks that the pandemic inflicted upon the US and global economies are now past. But those waning economic phenomena appear likely to have a long tail: The impacts of this global crisis will shape our economic and social landscapes for years to come.
Housing markets are a central forum in which those long-term shifts will play out. They both drive and reflect wider economic realities. This makes them among the most valuable venues for analysis and understanding – to the extent they can be rendered visible, in detailed and consistent fashion, across quarters and years and decades.
Few other areas of the economy are required by law to have their transactions and patterns logged and scrutinized in such detail as in housing finance and home lending. It is this rigorous set of rules that make analysis such as this report possible. We are able to identify and explore the radical swings of the American mortgage marketplace over the past half-decade and unearth key trends that hint at what’s to come.
Yet this high-resolution clarity required by statute and regulation – so valuable to lenders, borrowers and policymakers alike – is becoming worryingly blurry. The vital data-capturing activity the law has built into the housing finance system is being thwarted. The statistical transparency researchers, investors, borrowers and oversight leaders rely upon is endangered.
It is nonetheless possible to identify several signals within the noise.
What we have found is not encouraging. The data contain numerous indications that the US housing market is serving the wrong functions in our economy – capturing capital where it is neither needed nor constructive, rather than building sustainable opportunities to build wealth where it is absent.
Below we detail these trends, as well as a handful of brighter spots in the picture – while also sounding an alarm about threats to our ability to continue seeing into the housing market in these ways at all.
What follows is an assessment of what is visible, an explanation of what is not and an evaluation of what should be done about all of it.
Overview of 2018-2023 Lending
The mortgage market in 2023 tells a clear story: Rising interest rates prompted a shift back to home purchase lending, marking a significant change in consumer behavior and market dynamics. The role of Black and Hispanic borrowers in this story grew, yet they continued to face higher costs in securing loans. The national trends assessed in this report suggest answers to some key questions:
What will be the lasting impact of the pandemic on the housing market? Will it manifest as a significant wealth transfer to homeowners who capitalized on historically low rates? Or will it exacerbate the divide between those who owned homes at the start of 2020 and those who did not, potentially accelerating the wealth gap between Black and White Americans?
All Applications and Originations
Figure 1
The rising rates of 2023 effectively ended the refinance market, which had already taken a massive hit in 2022. Home purchase lending declined as well, as higher rates, prices and closing costs pushed many would-be buyers out of the market.
In 2023, lenders reported a total of 11,468,465 loan application records (LAR), marking a significant reduction from 16 million in 2022 and 26 million in 2021. This represents the lowest number of applications reported since enhancements were made to the Home Mortgage Disclosure Act (HMDA) in 2018. The surge in lending seen in 2020 and 2021 was driven by historically low interest rates, which encouraged a boom in refinancing activities. However, as interest rates normalized in recent years, this frenzy has substantially decreased.
The retreat of borrowers from the refinance market is stark, with refinance lending, including cash-out refinance loans on owner-occupied single-family homes, dropping by 60% to 927,484 in 2023. The burst of activity in home equity and home improvement lending observed during the period of low rates has also diminished. This report will discuss how the era of low interest rates has influenced homeowner wealth and its implications for decades to come. The data for 2023 suggests that the refinancing boom has ended, prompting a need to evaluate the impact of pandemic-era rates and post-pandemic home values on future homeownership and the racial wealth gap.
Figure 2
The dynamics of the mortgage lending market have seen significant changes recently. Independent, non-bank mortgage companies held 47.7% of the overall mortgage market (including all loan types) in 2023, a drastic decrease from 60% in 2021. Meanwhile, banks have maintained a consistent output, issuing about 35% of loans over the last two years, an increase from earlier years but still below their 42.6% market share in 2018. Credit unions have shown remarkable growth, with their market share rising from under 10% in 2020 to 18% in 2023. This growth is especially notable in the home equity lending market, where credit unions have increased their share dramatically; from 33.1% in 2018 for home improvement loans to 46.2% in 2023, and from 29.2% to 39.9% in home equity lending over the same period.
These shifts in the market share for all categories of home lending taken together do not, however, correspond to any meaningful shift in the market share of loans to purchase a home. Despite reducing their share of the overall market, mortgage companies still issued 64% of all home purchase loans in 2023, slightly down from over 65% in 2021, showing their continued strength in this area.
Mortgage companies withdrew swiftly from the refinancing market—dropping from 61% of refinance loans in 2021 to just 20% in 2023. This rapid shift underscores the vital role that depository institutions like banks and credit unions play in mortgage lending, highlighting their stability and enduring presence in the market, especially in periods of changing demand such as the decrease in refinancing.
Given this context, NCRC advocates for including independent mortgage lenders under the Community Reinvestment Act (CRA) regulations. The quick exit of mortgage companies from certain lending areas, particularly refinancing when interest rates normalize, reveals a gap in meeting the credit needs of communities — a gap that could be addressed by extending CRA oversight to ensure these lenders also contribute positively to community development and financial inclusion.
Across Different Borrowers and Neighborhoods
Figure 3
Low- and moderate-income (LMI) communities and families, as well as Black and Hispanic borrowers, continue to grow their share of the mortgage market.
Across all loan types, there was a considerable reduction in the total number of loans issued from over 13.3 million in 2021 to about 4.9 million by 2023 as interest rates have risen back to a more historically normal level. Within this overall contraction in lending, the proportions of loans to LMI borrowers and within LMI census tracts increased slightly. Specifically, loans to LMI census tracts rose from 15.3% of loans in 2021 to 18.9% in 2023, and loans to LMI borrowers increased from 25.4% in 2021 to 27.5% in the same period.
These rises appear to be due to higher-income families strategically withdrawing from the market, rather than to an influx of LMI borrowers. Many homeowners that benefited from low rates in 2020 and 2021 now find that they are essentially “trapped” in those homes. A family that refinanced their $300,000 mortgage at 2% would have a principal and interest payment of $1,100. Selling that home and taking on a new mortgage of the same amount at 7% would saddle them with a monthly payment of almost $2,000. Even if they want to sell they are not able to afford homes of similar value at current interest rates.
Excluding the 12% of loans that lack demographic data, we do find evidence that the share of all mortgage loans to Hispanic borrowers increased substantially, rising from 12% in 2021 to 13.9% in 2023. And in what might be a sign of a historic point, Hispanic lending for home purchase loans rose from 14.9% in 2021 to 16.5% of all home purchases in 2023, closely aligning with their overall share of the US adult population (17%).
Black borrowers also saw their share of mortgage lending rise, though much more modestly: 9.3% of home purchase loans went to a Black borrower in 2023, up from 9.0% in 2021. This means that the 12% of the adult US population who are Black remain underrepresented in the mortgage market.
This progress signifies a positive development where Black and Hispanic families are gaining a larger share of the housing market, particularly as the market transitions from predominantly refinances to more home purchases due to shifts in interest rates.
Meanwhile, the percentage of loans granted to Asian borrowers remained relatively consistent, averaging around 6%, which is slightly above their representation in the population.
Figure 4
Lending in Less-White Places
The lending landscape in majority-minority census tracts has seen notable increases over recent years. In 2018, the proportion of all mortgages originating in tracts where non-Hispanic Whites comprised less than 50% of the population was just above 19%. By 2023, this figure had escalated to over 24%. A change occurred in 2022 when the HMDA dataset transitioned to the new 2020 Census tract boundaries. This update, which occurs every decade, involves the creation of new tracts to reflect changes revealed by the latest Census data.
During this redrawing of census tract boundaries, areas with growing populations are frequently divided. This division can result in the formation of new census tracts that qualify as majority-minority, meaning non-Hispanic Whites make up less than half the population. Consequently, if a newly defined tract falls into this category, the mortgage data from that year may suddenly show an increase in loans to these areas, even if the actual lending behavior hasn’t changed dramatically.
This shift can introduce what appears as a discontinuity in the data, as seen in 2022. The 2023 data are therefore a new opportunity to observe changes free of that one-year tract-shifting noise.
Figure 5
Since the new boundaries took effect in 2022, there has been a notable increase in the share of loans directed toward majority minority census tracts and a corresponding decrease in loans to lower minority tracts. This trend may be attributed to rising housing prices and interest rates, which have made lower-priced homes in higher minority neighborhoods more appealing to homebuyers. The shift raises concerns about gentrification, including the possibility of intra-racial gentrification, prompting an analysis of the share of lending in these communities by the race and income of borrowers.
The overall share of home purchase loans made in high minority tracts (defined as areas with over 75% non-White residents) increased by 0.5 percentage points, from 11.1% in 2022 to 11.6% in 2023. This change represents a percent increase of approximately 4.74%.
White and Asian borrowers experienced a decline in their share of loans in high minority tracts across all income levels. Conversely, Black and Hispanic borrowers, particularly those in middle- and upper-income brackets, increased their presence in these areas. Middle- and upper-income Black borrowers, as well as moderate-, middle- and upper-income Hispanic borrowers, saw an increased presence across all census tracts, including high minority areas.
Middle-income Hispanic borrowers exhibited the highest increase in loans in tracts with 25% to less than 50% minority residents, rising by 16.13%. Middle-income Black borrowers also had a notable increase of 6.36% in high minority tracts.
On the other hand, low-income White borrowers experienced the largest decrease in lending across all minority tracts, with the most pronounced decline in tracts with 50% to less than 75% minority residents, where the low-income White borrower share dropped by more than 20%. Low-income Asian borrowers also saw decreases in lending, particularly in tracts with 25% to less than 50% minority residents, with a decline of 14.32%.
These changes indicate a shift in the demographics of borrowers in majority minority tracts. The increase in the presence of Black and Hispanic borrowers, especially those in the middle- and upper-income brackets, suggests these groups are becoming more active in the home purchase market in high minority areas.
This trend underscores the importance of monitoring potential gentrification and ensuring equitable access to housing for all demographic groups. Unlike racialized gentrification, where upper-income White borrowers move into lower income high minority communities, intra-racial gentrification is when higher income home buyers displace lower income residents of the same race. It will remain difficult to probe for intra-racial gentrification without increased data disaggregation in HMDA and additional study to understand these shifts – and use that understanding to better address the challenges and opportunities in promoting fair lending practices and supporting diverse communities.
Figure 6
Lending in Less-Wealthy Places
We can conduct a similar analysis of geographic shifts in lending behavior focusing on the economic status of a given place, separate from the racial makeup discussed above. When we talk about LMI tract lending, we’re focusing on mortgage loans made in areas categorized as low- and moderate- income (LMI) tracts. These tracts are defined based on median household incomes relative to the surrounding metropolitan area. From 2018 to 2023, there’s been a noticeable increase in the share of mortgage loans in these LMI tracts across different loan categories. For instance, in the “All Loans” category, loans in LMI tracts rose from 16.5% in 2018 to 18.9% in 2023. Home purchase loans specifically saw an increase from 17.5% to 20.5% in the same period. This trend also extends to loan purposes like refinance and home improvement loans, all of which have seen varying degrees of growth in directing funds into LMI tracts.
Lending to LMI borrowers shows a bit more variation than lending in LMI tracts over this same period. The overall percentage of all loans going to LMI borrowers ticked up slightly from 27.0% to 27.5%. However, the specifics can differ significantly by type. For example, while home purchase loans to LMI borrowers actually decreased from 28.1% to 26.4%, refinance loans and cash-out refinance loans increased considerably. Refinance loans for LMI borrowers rose from 25.1% to 29.1%, and cash-out refinance loans jumped from 29.5% to 35.1%. This indicates a growing engagement of these borrowers in accessing their home equity.
Loan Costs & Home Values by Race and Class
Figure 7
Across all mortgage types, there’s a notable disparity in loan costs and wealth-building opportunities among different racial groups. Asian borrowers typically have higher incomes and consequently obtain larger loan amounts, which allows them to purchase properties of significantly higher values. For example, the average property value for Asians is around $723,144, which starkly contrasts with $381,886 and $433,064 for Black and Hispanic borrowers, respectively. This discrepancy suggests that Asian borrowers are accessing more affluent housing markets compared to their Black and Hispanic counterparts.
Interest rates and rate spreads further complicate the financial landscape. Black and Hispanic borrowers face higher average interest rates of 6.86% and 6.89%, respectively, compared to 6.67% for Asian borrowers. This indicates that even though Black and Hispanic individuals are buying less expensive homes, they are doing so at relatively higher costs.
When isolating the data to only home purchase loans, these trends continue. Black and Hispanic borrowers still encounter higher costs relative to the value of the properties they are buying. This is evident not just in higher interest rates but also potentially in higher average total charges to the borrower, suggesting that the cost of acquiring a mortgage is more burdensome for these groups.
Figure 7b
The current trends indicate a troubling forecast for the racial wealth gap, particularly among Black and Hispanic communities. As interest rates climb, these families often find themselves confined to the lower end of the housing market due to financial constraints. This limitation on purchasing power not only restricts the immediate value of their investment but also the potential long-term wealth that homeownership typically offers.
Since homeownership remains a primary vehicle for wealth accumulation in the US, the ongoing high home prices, coupled with rising interest rates, likely mean that the racial wealth gap will continue to widen. The reality that Black and Hispanic families face more significant barriers to becoming homeowners only exacerbates this issue, hindering their ability to build wealth at a pace comparable to other racial groups. This systematic disparity underscores the need for targeted policies to address these imbalances and foster a more equitable housing finance system.
Investor Lending
Investor purchases of housing continue to distort many local markets, keeping home prices from falling as rates have risen.
The primary focus of this report has been on owner-occupied lending. However, since the onset of the pandemic, there has been a noticeable surge in investor purchases of single-family homes. This increase is particularly pronounced in lower-cost units and in Sunbelt markets, where the market share of these properties is increasingly dominated by medium and large-scale rental operators. Initially spurred by low interest rates and high rents, investor activity has moderated but remains elevated compared to pre-pandemic levels. Investors often target lower-cost properties and frequently outbid families attempting to transition from renting to homeownership, especially in competitive markets.
HMDA captures data on the occupancy of homes that serve as collateral for mortgages, categorizing them as either a principal residence (owner-occupied), a second residence or an investment property.
There has been a steady rise in the proportion of loans made to investors, reaching a peak in 2022. For example, the percentage of all loans made to investors climbed significantly from 2018, saw a slight dip in 2021, but soared to its highest in 2022. This trend aligns with national data that indicate a surge in investor activity during the same period, fueled by low interest rates and a competitive housing market.
Figure 8
In the realm of home purchase loans, the data shows a consistent increase in loans granted to investors. This trend underlines the role investors play in setting prices and rents across many communities. This can also adversely affect the availability and affordability of homes for first-time and lower-income homebuyers. In 2023, over 8% of home purchase loans were made to investors, underscoring the substantial role they play in this market segment.
Refinance loans to investors also witnessed growth, particularly in 2022 and 2023. Over 12% of refinance loans in 2023 were secured on investment properties, indicating that investors are leveraging lower interest rates to improve their financial positions on existing properties.
Although large corporate buyers control a relatively small portion of the total single-family rental (SFR) market nationally—just over 3%—their impact is considerably more significant in specific local markets. In areas with high growth or high rental demand, such as Atlanta, Phoenix and parts of Florida, their market share can be substantially higher. Since the pandemic especially, large scale investors in the single-family rental space have aggressively increased their share of the market. These corporate landlords also pursue evictions at a higher rate than smaller investors. This concentration can significantly influence local housing prices and rental rates. HMDA does not contain information about large institutional lenders such as this. The investor lending captured here is to smaller investors, who are often using equity from a primary home to fund the downpayment on an investment home and financing the balance.
These insights and the underlying data highlight the evolving dynamics of the housing market, where investor activity continues to play a pivotal role in shaping market outcomes. This trend necessitates ongoing monitoring and potentially new policies to ensure a balanced and equitable housing market. While investors continue to wag the dog in real estate, turning homes into commodities rather than places to live, it’s clear that some tail-pulling by policymakers is overdue. Measures could include enhancing the transparency of property ownership and imposing more stringent regulations on large-scale investment activities to protect the interests of first-time and lower-income homebuyers. Local policy solutions might also involve setting limits on short-term rentals, implementing progressive tax rates on non-owner occupied single-family homes, and strengthening protections for renters to prevent predatory evictions. Let’s remember: homes are for people to live in, not just investment portfolios to grow. If investors are playing Monopoly with real estate, maybe it’s time to change the rules of the game.
Disaggregated Data
More detail on the specific race of mortgage borrowers has opened new avenues of understanding mortgage discrimination and how it plays out.
In 2018 when the latest version of HMDA data began, it included a more detailed look at race than ever before. Data disaggregation allows the loan applicant to identify themselves with more precision. Limited to Hispanic, Asian and Hawaiian or Pacific Islanders, borrowers could choose a number of more specific groups.
At the time, comments submitted to the CFPB complained about the burden of adding the additional questions to the mortgage application. Some even suggested that collecting such data would be a potential barrier for some borrowers.
Figure 9
In fact, the collection of this data has not been shown to be a burden, lenders simply include the additional question in place of the previous race and ethnic group question. Applicants have responded enthusiastically to the new options, with over 80% of Asian, 55% of Hispanic and 60% of Hawaiian or Pacific Islander applicants choosing to use a more precise definition of their background in 2023. In 2020 NCRC found that Asian borrowers were often upper-income Chinese and Asian Indian borrowers. NCRC and the CFPB also found that other Asian groups, particularly Filipino and Vietnamese, often display similar patterns of denials and loan costs as Black and Hispanic borrowers. This observation was impossible prior to the collection of disaggregated data.
NCRC has recently called for the expansion of the collection of disaggregated data for other racial groups that are thought to be homogenous in their experience with the mortgage market, such as Black people. As with Asian borrowing, the current schema of data collection may obscure different outcomes for African immigrants, Caribbean immigrants, their descendants who were born in the United States, descendants of American Chattel Slavery and other borrowers that currently can only identify as Black. There are documented differences in outcomes between African immigrants, Caribbean immigrants and descendants of American Chattel Slavery in income, college degree attainment and poverty; descendants of American Chattel Slavery have the worst outcomes in all three categories. As NCRC and other groups increase the pressure to collect more disaggregated data, the inherent nature of lenders or other organizations will be to again claim that the collection of such data is a burden. The experience of the 2018 HMDA rule, and the value of the findings derived from this data, is that not only is such collection not a burden but that to avoid protracted and costly litigation the lending industry should meet with community organizations and regulators to agree on a systematic method that allows all borrowers to identify themselves more precisely. As regulators begin the shift away from the two-tiered ethnicity/race classification for Hispanic and non-Hispanic, and add Middle Eastern or North African (MENA) as a race, the CFPB should consider other improvements to HMDA data collection that could be implemented as well.
This data is critical. As regulators initiate the data collection process for small business data under Section 1071 of the Dodd-Frank Act, lenders and their funded trade organizations will predictably resort to tired claims that data collection is an undue burden. Such objections are patently absurd and should be dismissed outright. The only “burden” here falls on those lenders who neglect their responsibility to serve underserved markets.
Same-Sex Lending
A growing field of study has found that discrimination occurs along many different vectors other than race and income.
In addition to disaggregated racial data, the growing need to collect Sexual Orientation and Gender Identity (SOGI) data must be addressed. As with race and ethnicity, the addition of such data results in nearly no burden for lenders and borrowers appear to be eager to share more about themselves than they are normally asked to provide. In a report on same-sex mortgage lending in 2020, NCRC identified patterns that suggested different outcomes for same-sex borrowers. Further disaggregation of this data would help better understand if there is a discriminatory impact on a family’s ability to secure a mortgage based on their sexual orientation and gender identity.
Figure 10
Although HMDA does not collect data on the sexual orientation or gender identity of borrowers, the identification of applicant pairs as either both male or both female provides an indirect indicator of loans potentially issued to LGBTQ+ couples. Over the period from 2018 to 2023, there has been a gradual increase in the share of these loans. For example, loans to male/male applicant pairs have risen from 3.5% in 2018 to 5.1% in 2023, and female/female pairs have seen a rise from 10.6% in 2018 to 10.8% in 2023. This trend suggests a growing representation of same-gender pairs among mortgage loan recipients.
Figure 11
The total share of loans to applicants of the same gender increased from 5.3% in 2018 to 7.0% in 2023. This growth may reflect a broader societal acceptance and recognition of LGBTQ+ individuals, potentially encouraging more couples to apply jointly for mortgages. Furthermore, the increment in these percentages aligns with broader trends of legal and social recognition for same-sex marriages and partnerships over these years. While we must be cautious in directly associating these trends with an increase in LGBTQ+ borrowers—given the absence of explicit SOGI data—the correspondence of rising percentages with significant civil rights advancements offers a compelling context. This data assists in understanding the landscape of mortgage lending to a possibly underrepresented group, underscoring the importance of considering SOGI in future housing finance research and policy development to ensure equitable treatment and opportunities in homeownership.
Multi-Racial Lending
Figure 12
As the US becomes more multi-racial, disaggregated data and a more nuanced way of understanding mortgage data is critical.
The trends in mortgage lending to multi-racial borrowers and interracial couples reflect broader shifts in the demographic composition and identity in America. As the US has experienced significant diversification, the proportion of individuals identifying as multi-racial has seen substantial growth. The share of Americans that reported a multiracial background saw an increase of 276% from 2010 to 2020 as the Census developed new methods of encouraging respondents to identify themselves with more precision.
This demographic shift is mirrored in the mortgage sector. Home purchase borrowers in 2023 reported their racial identity. Among those that offered this data, specifically for home purchase loans, this figure rose to 15% in 2022 . This rise is part of a broader national trend where racial and ethnic minorities, particularly Latino or Hispanic groups, have been pivotal in population growth across many states and metropolitan areas, often offsetting declines in the White population and injecting youth into the demographics.
Figure 13
Moreover, the increasing incidence of interracial couples and the resulting rise in multi-racial identities are influenced by greater social acceptance and the changing social fabric of the US. This is evident in the growing percentage of interracial marriages and children from interracial backgrounds, highlighting a societal shift towards embracing multiracial identities. The increase in loans to multi-racial applicants and couples not only reflects these demographic trends but also underscores the need for continued vigilance against potential discriminatory lending practices that could undermine the gains made in housing equity for this increasingly significant segment of the population.
How the rising share of multiracial lending impacts the future wealth of multiracial people remains to be seen. As interest rates have risen, multiracial buyers, like Black and Hispanic buyers in general, are getting a larger share of a smaller number of mortgage loans. They are also paying more for it than buyers did during the pandemic. The long-term impact of this on issues such as the racial wealth gap is a topic in need of further study.
The increase in the number of multi-racial applicants and interracial couples raises new concerns of structural or overt racism on the part of lenders that may harm consumers. Expanded data collection and matched-pair testing by communities, consumer protection groups like NCRC and regulators should look for evidence of disparate impact based on the more nuanced experience of multi-racial loan applicants.
Purchased Lending and Missing Data
Purchased loans, as reported under HMDA, are those that were bought by a reporting lender from another lender that made the original credit decision at the time the loan was originated. These contrast with originations, where the reporting lender is directly involved in making the credit decision. Despite making up a significant portion of the mortgage market, purchased loans frequently bypass the same level of scrutiny as originations.
The reporting of these loans reveals a notable gap in data transparency, particularly concerning demographic and income information. For example, in 2023, while the overall quality of data collection showed improvement — loans lacking demographic data decreased from 14.5% in 2021 to 12.2% – the absence of demographic data with purchased loans remained stark. Over 90% of purchased loans did not include demographic data, and a significant number lacked income details as well. Combined with the similar issues in originations, this represents a substantial portion of the dataset where crucial information is missing.
This lack of data is exacerbated by a loophole in HMDA regulations that allows lenders to omit demographic and income data from reports on purchased loans. In 2023, this resulted in a large discrepancy: only 12% of originations lacked demographic data, yet 90% of purchased loans were reported without this essential information. The impact is profound, as these gaps hinder our ability to understand who is receiving loans and under what terms, which is vital for assessing fairness and inclusivity.
Figure 15
Additionally, the public dataset for HMDA does not include the Universal Loan Indicator (ULI), which makes it difficult to track whether a loan reported as purchased had also been reported as an origination by the initial lender. This limitation complicates efforts to fully grasp the mortgage market’s dynamics, particularly the flow of credit to low- and moderate- income (LMI) and BIPOC communities.
Further complicating the landscape is what we find when we look at who lenders report that they sell their loans to. This includes entities like the Government Sponsored Enterprises (GSEs) Fannie Mae and Freddie Mac, as well as commercial banks and other financial institutions. The data for 2023 showed that out of the 1.4 million purchased loans, 863,743 were reported as sold to other lenders who might conceivably report them as purchased loans in HMDA. This leaves over 500,000 loans that only appear in the data as “purchased” loans, and we know very little about them. This discrepancy raises concerns about the transparency of loan reporting and the actual flow of mortgage credit to low- and moderate- income (LMI) and BIPOC communities.
One significant issue with the reporting of purchased loans is a loophole in HMDA regulations that allows purchasers to omit demographic and income data from their reports. This loophole means that while only 12% of mortgage originations lacked demographic data in 2023, a staggering 90% of purchased loans reported missing this crucial information. For home purchase loans 83% also lacked borrower income data.
The major role of purchased loans in the market and the significant gaps in reporting call for enhanced regulatory oversight to close these loopholes. Ensuring comprehensive and transparent data reporting is essential for enabling a clearer understanding of how mortgage credit is distributed, especially to historically underserved communities.
Figure 16
The challenge of missing demographic data in loan applications is substantial and merits careful consideration in any analysis of lending practices. In the 2023 data set for site-built, 1-4 unit owner-occupied homes, a comparative review reveals statistically significant differences (Figure 17). The “No Data” race/ethnicity category differs from the combination of the others in key values like borrower income, amount of the loan, loan costs, loan to value ratio, rate spread and lender credits but not in the interest rate of the loan. This indicates that borrowers who fail to identify their race and ethnicity differ from a mix of other borrowers, tending to be more affluent and to borrow higher amounts, all suggesting different economic status. Additionally, the lower loan to value ratios and rate spreads indicate that their risk profiles are different, and that they use a broader range of loan products. Since owner-occupancy is used as a selection criteria, the “No Data” category cannot consist of institutional investors. Because of these factors, it is recommended that the data be divided into “Known” and “Unknown” race/ethnicity categories prior to demographic analysis. This is because borrowers who choose not to identify have significantly different parameters in the types of loans which they are obtaining, which are not comparable to the other demographic categories.
For this analysis, we have opted to exclude these “No Data” records from our examination of lending patterns by race. This decision marks a shift from our previous reporting methods and from the standard practices observed in HMDA reporting. Our rationale is grounded in the aim to preserve the integrity of the data analysis. Including loans without demographic data could potentially dilute the analytical value of the remaining 88% of the dataset, which contains more complete information. Including such incomplete records would distort our understanding of the racial and ethnic compositions of loan applicants, especially when making comparisons to Census data, which do not account for a “No Data” category.
Figure 18
To maintain clarity and reliability in our analysis, and to avoid introducing a non-existent racial category within HMDA contexts, it is essential to focus on records with complete data. While it remains critical for lenders to address and minimize the occurrence of missing data, our analytical approach excludes these records to ensure a more accurate reflection of demographic lending trends. This methodological adjustment allows us to provide a clearer, more precise evaluation of racial and ethnic disparities in mortgage lending.
Tying It All Together
The stakes of mortgage lending, redlining and inequality extend far beyond academic discourse; they profoundly impact public health. NCRC’s extensive research illustrates the grave consequences of denying communities access to mortgage credit. In redlined areas, residents suffer increased mortality rates from preventable diseases like obesity, high blood pressure, kidney disease and stroke—all conditions exacerbated by the environmental and economic strains these communities endure. Future research must address the specific impacts of redlining on COVID-19 mortality rates and explore how the era of ultra-low interest rates has created a divide in America between those lucky enough to own a home before the pandemic and everyone else.
This divide traps existing homeowners in properties they cannot sell, while simultaneously barring potential new homeowners from entering the market due to escalating rates and inflated housing costs driven by investor activity. HMDA data is a vital tool in tracking the investment flows that shape these neighborhoods. We urgently need more comprehensive data on small business and community investment to effectively craft policies that mitigate the harsh realities of redlining. This can lead not only to enhanced economic prosperity but also to significant improvements in public health. The extensive benefits of detailed data collection, encompassing income, race, sexual orientation and gender identity, decisively outweigh any concerns over burden or privacy. It’s imperative that efforts to curtail this essential data collection be recognized as not just misguided but as detrimental to the health and well-being of our communities. It is crucial that data collection efforts be intensified, not hindered, to break the vicious cycle of poverty and environmental degradation that afflicts these areas.
Top 100 Metros
Figure 19
Patterns of discrimination are intricately mirrored in the landscape of mortgage lending across the United States. Each metropolitan area exhibits unique manifestations of these patterns, influenced by various factors. Historically, the practice of redlining, initially institutionalized in the 1930s by the federal Home Owner’s Loan Corporation (HOLC), delineated and reinforced existing segregation and trapped the most economically disadvantaged and marginalized populations in certain urban sectors. These practices left a lasting mark and ingrained a legacy of racial and socioeconomic segregation into the financial fabric of American cities for decades. Even today, mortgage lending continues to trace these historic lines of division based on race and class, established over a century ago. From cities in the rustbelt to those in the sunbelt, the enduring impact of past prejudices is evident in the spatial distribution of mortgage lending.
As home purchase lending returned to the fore in 2023, it is important to see which racial groups saw their share of the home-buying market improve. Unlike refinance lending, which declined by about 60% from 2022 to 2023, home purchase lending declined by just 20%.
Nationally, a major trend influenced by rising interest rates was the marked increase in the use of discount points. Discount points allow a homebuyer to pay an upfront fee at closing in exchange for a reduced interest rate over the life of the loan. In 2019, homebuyers paid an average of just under $1,800 at closing in discount points. By 2023, this average had risen to approximately 2.5 times that amount, reaching around $4,500. This shift reflects an adaptation to the dual pressures of rising home values and higher borrowing costs. As property values increased, borrowers found themselves needing to secure larger loans, despite these loans being offered at higher interest rates.
During this period, many closing costs, including origination fees — which are payments made to the loan originator for processing the loan — are regulated to not exceed 3% of the total loan amount. Consequently, as the average loan amount grew from $268,873 in 2019 to $372,364 in 2023, the cap on these fees increased correspondingly. Origination fees climbed from $1,630 to nearly $3,700, tracking closely with the increase in average loan amounts.
The escalation of these costs functions much like the formidable barriers of a medieval castle, hindering many potential homeowners’ transition from renting to owning. This financial barrier effectively locks out numerous individuals from the primary avenue for wealth accumulation in America, trapping them in a cycle where they contribute to the wealth of landlords, institutional investors and Wall Street banks instead of building their own.
The pandemic, the historically low interest rates that followed, and the rise in both rates and housing prices as COVID-19 became endemic, played havoc with the mortgage market nationally. Many metro areas experienced changes that will have repercussions after the pandemic has faded from our collective memory. In the 100 metros where the most loan originations were reported from 2019 through 2023, this section of the report looks at specific loan, borrower and neighborhood characteristics to see where the changes were most pronounced over that timeframe.
The period from 2019 to 2023 marked major changes in the landscape of home purchase lending across various metros. The data reveals shifts in loan characteristics, borrower demographics and neighborhood profiles, painting a comprehensive picture of how the mortgage market evolved during these years.
For instance, Myrtle Beach, SC-NC, experienced a steady climb in home purchase loans, starting at 8,844 in 2019, peaking at 12,330 in 2021, and leveling off at 9,766 in 2023. This 10.4% increase underscores the area’s growing appeal among both local buyers and those relocating.
In a similar fashion, Lakeland, FL, showed a positive trend, with loans rising from 11,311 in 2019 to 14,604 in 2021. Despite a minor drop in 2022, the number stayed high at 11,810 in 2023, marking a 4.4% increase over the period.
On the other hand, Port St. Lucie, FL, saw some fluctuations. Home purchase loans fell slightly from 7,830 in 2019 to 7,481 in 2023, despite peaks in 2020 and 2021. Rural Texas mirrored this trend, peaking in 2021 before a significant decline by 2023, influenced by broader economic factors. Deltona, FL, had substantial growth during the pandemic years but later retracted, resulting in an 8.4% overall decrease.
In stark contrast, metros like Fresno, CA, and San Diego, CA, faced sharp declines in home purchase loans, with Fresno dropping by 40.2% and San Diego by 41.7%. Ogden, UT, and Salt Lake City, UT, also experienced notable reductions, highlighting volatility and market adjustments. The largest decrease was seen in Oxnard, CA, pointing to significant market challenges.
Examining average borrower income from 2019 to 2023 reveals robust growth in several metros. Miami, FL, saw the highest percentage increase, from $129,034 in 2019 to $211,302 in 2023. North Port, FL, and Provo, UT, also witnessed substantial increases, reflecting economic growth and demographic shifts. In contrast, Baton Rouge, LA, and Pittsburgh, PA, had more modest gains, indicating stable but less dramatic improvements. Interestingly, rural Wisconsin saw a decrease in average borrower income, suggesting economic challenges or demographic shifts.
Average loan amounts grew significantly, especially in Florida and California. Cape Coral, FL, had the most notable rise, with loan amounts increasing from $232,427 in 2019 to $363,791 in 2023. Stockton, CA, and Miami, FL, also saw substantial increases, highlighting robust housing markets. San Francisco, CA, had a smaller percentage increase, though high absolute loan amounts indicate substantial financial barriers to entry.
The average property value from 2019 to 2023 also saw significant growth in several metros. North Port, FL, had the highest percentage increase, driven by high demand. Cape Coral, FL, and Portland, ME, recorded substantial growth as well. Meanwhile, Honolulu, HI, and New Orleans, LA, had more modest increases, indicating stable markets.
Interest rates saw notable increases across various metros, reflecting broader economic trends. Omaha, NE-IA, had the most substantial rise, with rates going from 4.02% in 2019 to 6.60% in 2023. San Jose, CA, and Portland, ME, also saw significant increases. Conversely, Jacksonville, FL, and Las Vegas, NV, experienced relatively smaller increases, suggesting regional differences in the impact of interest rate changes.
Borrower charges increased significantly across several metros, reflecting rising housing costs. Reno, NV, saw the largest increase, with total charges rising from $6,649 in 2019 to $15,832 in 2023. Tucson, AZ, and Phoenix, AZ, also recorded large increases. Pittsburgh, PA, and Chicago, IL, had more modest gains, indicating stable but less dramatic growth in borrowing costs.
From 2019 to 2023, the share of LMI borrowers showed interesting trends. San Francisco, CA, saw a significant increase, suggesting growing inclusion in a high-cost market. Akron, OH, and Greensboro, NC, also recorded notable increases. Conversely, Orlando, FL, and Salt Lake City, UT, experienced substantial declines, highlighting challenges in maintaining affordability amid rising housing costs.
Lending data in minority majority census tracts revealed varying trends. San Francisco, CA, and San Jose, CA, saw notable increases in lending, highlighting efforts to support predominantly minority areas. In contrast, Orlando, FL, and Salt Lake City, UT, experienced declines, indicating challenges in supporting minority-majority areas amid rising costs.
Trends in minority borrowers showed significant changes. San Francisco, CA, and San Jose, CA, saw notable increases in the share of minority borrowers, highlighting efforts to promote homeownership among minority populations. In contrast, metros like Orlando, FL, and Salt Lake City, UT, experienced declines, indicating challenges in supporting minority borrowers amid rising housing costs.
San Francisco, CA, and San Jose, CA, saw significant increases in the share of Asian borrowers from 2019 to 2023, reflecting efforts to promote homeownership among Asian populations. In contrast, metros like Orlando, FL, and Salt Lake City, UT, experienced declines, indicating challenges in supporting Asian borrowers amid rising housing costs.
Atlanta, GA, and Washington, DC, saw significant increases in the share of Black borrowers, highlighting efforts to promote homeownership among Black populations. In contrast, metros like Orlando, FL, and Salt Lake City, UT, experienced declines.
Finally, data revealed notable increases in the share of Hispanic borrowers in Houston, TX, and Los Angeles, CA. Conversely, Orlando, FL, and Salt Lake City, UT, experienced declines, highlighting challenges in maintaining support for Hispanic borrowers amid rising housing costs.
REPORT MATERIALS
Report PDF
TABLE OF CONTENTS
Subscribe
Get NCRC news and
alerts by email.
