Reading HMDA Data for Mortgage Lending Patterns

Why Fair Lending Analysis Matters

The mortgage market is the largest consumer credit market in the United States, over $12 trillion in outstanding mortgage debt and millions of new applications each year. Access to mortgage credit determines whether families can build wealth through homeownership, the single largest source of household net worth for most Americans. Disparities in lending access have compounding effects across generations.

Congress created the Home Mortgage Disclosure Act specifically to provide the public and regulators with data to monitor lending patterns. Before HMDA, there was no systematic way to see whether lenders were serving all communities fairly. Today, HMDA covers 12.2 million applications annually from nearly 5,000 lenders, the largest mortgage dataset in the world.

PlainLender makes this data accessible by lender, metro area, and state. You can examine how lending patterns vary geographically and between institutions.

Key Metric: Denial Rate Disparities

What it tells you: Denial rate disparities measure how much denial rates differ between borrower groups applying to the same lender or in the same market. For example, if Lender A denies 8% of applications from one group and 18% from another group, the denial rate disparity is 10 percentage points (or a 2.25x ratio). Large disparities across multiple lenders in the same market suggest structural patterns that may warrant investigation.

What it doesn't tell you: Denial rate disparities in HMDA data are "unadjusted" - they do not control for credit scores, assets, employment stability, property characteristics, or other legitimate underwriting factors. Academic research consistently finds that a significant portion (though not all) of observed denial rate disparities can be explained by these omitted variables. A 10-percentage-point gap in raw denial rates might shrink to 2-3 points after controlling for creditworthiness factors.

Reading tip: Compare denial rates across lenders in the same market to see which institutions show the largest and smallest demographic disparities in the published tables. Gaps versus peers in the same geography are a research signal for further file-level review by agencies; they are not a PlainLender finding of illegal discrimination. Use lender profiles to view the as-reported fields side by side.

Geographic Lending Patterns

What it tells you: HMDA data includes the census tract where each property is located, enabling analysis of whether lenders are serving all neighborhoods within their market area. Geographic analysis can reveal whether certain neighborhoods receive disproportionately few loan originations relative to their population and housing stock. It can also show whether certain areas have higher denial rates or receive different loan products (e.g., more government-insured FHA loans vs. conventional loans).

What it doesn't tell you: Low lending volume in a neighborhood might reflect lower demand (fewer home sales, lower homeownership rates) rather than restricted supply. It might also reflect property conditions (homes too deteriorated to qualify for mortgages) or local economic factors (population decline, job losses). Not every geographic gap is evidence of redlining, but persistent gaps across multiple lenders in economically viable neighborhoods are a red flag.

Reading tip: On PlainLender's metro pages, examine the geographic distribution of reported lending activity. Neighborhoods with substantially lower origination volume relative to population are geographic patterns in the public data; income controls from census sources are needed before attributing cause.

Interest Rate Differentials

Since 2018, HMDA has included the interest rate on originated loans. This allows analysis of whether different borrowers pay different rates for similar loans, a key dimension of fair lending that was previously invisible in the data.

Interest rate analysis in HMDA must be done carefully. Rates legitimately vary based on credit score, loan-to-value ratio, loan type, and market conditions at the time of rate lock. Two borrowers who closed a month apart might have different rates solely because market rates moved. To draw meaningful conclusions, analysts typically compare borrowers with similar income levels, loan amounts, and loan types who applied at the same lender around the same time.

Large, unexplained rate differentials at a specific lender, rates that cannot be accounted for by loan characteristics and market timing, may indicate pricing practices that regulators would want to examine. This is exactly the type of analysis that led to the Ally Financial settlement in 2013, where dealer markup practices resulted in borrowers in certain groups paying higher rates.

Practical Framework: Analyzing a Lender's HMDA Profile

1. Compare denial rates to market peers. Is this lender's denial rate significantly higher or lower than other lenders in the same metro? A lender with a 25% overall denial rate in a market where peers average 15% may have stricter underwriting, or it may attract more marginal applicants through marketing. Context matters.
2. Examine denial rate disparities. Do denial rates differ significantly between borrower groups at this lender? Compare to the same disparities at peer lenders. If Lender A shows a 12-point gap while peers show 4-6 points, the difference may warrant closer examination.
3. Check denial reasons. What are the primary reasons this lender cites for denials? Are they consistent with the lender's stated underwriting criteria? A lender that disproportionately cites "collateral" (appraisal) issues in certain neighborhoods may be applying appraisal standards unevenly.
4. Look at geographic coverage. Does this lender originate loans throughout its market area, or are there neighborhoods it avoids? Compare the lender's geographic footprint to its branch locations and marketing reach.

Frequently Asked Questions