Executive Summary

This report provides an in‐depth comparison of three major real‐estate data sources critically used by appraisers: Multiple Listing Services (MLS), public property records, and third‐party aggregated data. Each source has distinct strengths and limitations in terms of accuracy, completeness, timeliness, and coverage. Appraisers traditionally rely heavily on MLS data for comparable sales, but recent analyses show that MLS prices and public recorder (“county record”) prices align in well over 90% of cases (Source: www.mortgagenewsdaily.com) (Source: www.mortgagenewsdaily.com). For example, CoreLogic found that in a sample of 156,000 California home sales, the MLS‐reported price matched the official county record in 91% of cases (Source: www.mortgagenewsdaily.com), with only ~2.1% of comparables differing by 1% or more. Similarly, Wall (FGCU) et al. (Appraisal Journal) found that MLS prices differed from HUD‐1 (closing) prices in only about 8.75% of transactions (2004–2008 data) (Source: theamericangenius.com). These findings indicate high internal consistency of MLS data, at least for core fields like sale price.

However, coverage gaps exist. MLS only includes properties listed through cooperating brokers, omitting off‐MLS sales (e.g. For‐Sale‐By‐Owner and “coming soon” listings). Industry data show that for‐sale‐by‐owner (FSBO) transactions accounted for only ~7% of U.S. home sales in recent years (Source: www.nar.realtor), meaning roughly 93% of sales are broker‐assisted. Still, MLS omits a nontrivial minority of transactions (e.g. Tyler Wood’s local analysis showed ~8–13% more sales in public records than in MLS, depending on year (Source: www.tylerwoodgroup.com). In some markets, new policies allow sellers to delay public listing (“coming soon”), further reducing MLS coverage (Source: www.axios.com). Public records capture nearly 100% of deed transfers (including FSBO), but face their own limits: not all states disclose sale prices publicly, data can be delayed or incomplete, and many property attributes (bedrooms, quality, etc.) are error‐prone or outdated (Source: retipster.com) (Source: www.clearcapital.com). Third‐party data vendors (e.g. CoreLogic, ATTOM, Zillow/Redfin) aggregate MLS feeds, tax assessor data, deed records, and proprietary estimates. They promise broader coverage (some claim data on 110–158 million homes nationwide (Source: www.zillow.com) (Source: www.attomdata.com) and advanced analytics, but inherit errors from source data and face licensing/regulatory constraints (Source: www.zillow.com) (Source: www.sec.gov).

We organize the findings as follows: First, we review each data source in detail (history, content, reliability). Next, we present comparative metrics (including a matrix table contrasting accuracy, completeness, timeliness, etc.). We then analyze empirical studies and real‐world cases illustrating differences in data quality (e.g. CoreLogic’s and FGCU’s analyses, FSBO trends, and off‐MLS effects). We also discuss how cutting‐edge methods ( AVMs, AI and policy changes (e.g. NAR’s “delayed listing” rule) are reshaping data availability. Finally, we highlight implications for appraisers and future directions, such as increasing data integration, regulatory impacts, and new technologies.

Key Findings: Appraisers consistently rely on MLS data for up‐to‐date, detailed listings (Source: www.mortgagenewsdaily.com) (Source: www.nar.realtor). In practice MLS and recorded sale prices agree in the vast majority of cases (≈91% exact match on price) (Source: www.mortgagenewsdaily.com) (Source: www.mortgagenewsdaily.com). Public records fill important gaps (FSBO and back‐records) but often lack completeness. Third‐party aggregators claim exhaustive data sets (e.g. ATTOM’s 158M properties (Source: www.attomdata.com), yet their data quality depends on original sources (county and MLS data quality issues propagate) (Source: www.zillow.com) (Source: www.clearcapital.com). Appraisers must be aware of each source’s biases: MLS may miss private sales or mis‐attribute features, public records may suppress sale prices (in “non-disclosure” states) (Source: retipster.com), and third-party feeds may lag or merge incompatible records. On balance, the report finds that no single source is perfectly accurate or complete, so prudent appraisers cross‐check multiple sources. The detailed Data Accuracy & Completeness Matrix below distills these insights (see Table 1).

<div markdown="1">

Table 1. Summary comparison of data sources (MLS vs Public Records vs Third‐Party)

Introduction and Background

Real estate appraisals depend critically on reliable data about comparable sales and property characteristics. The accuracy of these inputs directly affects value conclusions, loan underwriting, and market stability (Source: www.mortgagenewsdaily.com) (Source: www.clearcapital.com). Historically, appraisers have drawn comparable sales data from multiple sources:

• MLS (Multiple Listing Service): Originally developed by real estate associations in the early 20th century, MLS systems are cooperative databases maintained by broker associations. They contain detailed information on properties for sale or recently sold within a given market. MLS data are typically the freshest, as listing agents update status (e.g. price changes, contract status, closing) in real time. MLS often includes granular fields (bed/bath counts, square footage, condition, etc.) and rich media (photos, descriptions).

• Public Records: County Recorder and Tax Assessor offices maintain legal records of property transfers and assessments. These documents span decades and cover essentially all transfers of title. Thus, public records offer broad coverage—any deeded transaction (including For-Sale-By-Owner, court-ordered sales, etc.) is recorded. However, the level of detail and timeliness can vary. For example, in many states sale prices are not disclosed to the public. In “non-disclosure” states (e.g. Texas, Utah, Florida, Ohio, etc.), only financing data (mortgage amount) might be recorded publicly (Source: retipster.com). Even where price is recorded, data entry errors or lags can reduce accuracy.

• Third-Party Aggregated Data: In recent decades, commercial data providers (CoreLogic, Black Knight, ATTOM, Zillow, Redfin, etc.) have emerged to aggregate and “clean” information from MLS, public records, permits, and sometimes user-contributed data. These vendors promise unified platforms and analytics for appraisers and lenders. For example, ATTOM advertises “assessor data for more than 158 million properties” across all U.S. counties (Source: www.attomdata.com), while Zillow reports a living database of 110+ million U.S. homes combining many sources (Source: www.zillow.com). Such data often underpin Automated Valuation Models (AVMs) and appraisal technology. However, the very process of merging disparate sources introduces complexity; issues of matching records, standardizing formats, and resolving conflicts are non-trivial (Source: www.zillow.com).

Coupled with these data sources is an evolving regulatory and technological environment. Recent years have seen regulatory scrutiny (e.g., AMC and appraisal review reforms), advances in digital appraisal (Uniform Appraisal Dataset, digital portals), and new industry policies (notably the NAR “Listing Choice” and delayed-publication rules (Source: www.axios.com). Appraisers are also beginning to incorporate emerging tools like machine learning and image analysis, which rely on structured data inputs. All these trends heighten the need for accurate, complete, and timely data.

This report systematically compares MLS data, public records, and third-party data along multiple dimensions:

• Data Content and Completeness: What information does each source provide? Which property attributes or transactions might be missing or misreported?
• Data Accuracy: Empirical studies of price, attribute, and location accuracy in each source. How often do values align across sources?
• Coverage and Accessibility: What share of market transactions is captured? Are any transaction types or regions systematically excluded? Are there legal or cost barriers to access (e.g., licensing, non-disclosure laws)?
• Timeliness and Update Frequency: How current are the records? What's the lag between market event and data availability?
• Case Studies: Real-world examples illustrating discrepancies or complementarities (e.g., comparison studies, off-MLS sales analysis).
• Implications and Future Directions: How do data differences affect appraisal quality and the broader market? What emerging trends (e.g. AI, policy changes) may alter the landscape?

Our findings draw on industry reports, academic research, and corporate disclosures. Key sources include a CoreLogic analysis of 2015–16 California data (Source: www.mortgagenewsdaily.com), a Florida Gulf Coast University study (Source: www.mortgagenewsdaily.com) (Source: theamericangenius.com), White House/NAR press releases, trade journals, and data vendor publications (Source: www.clearcapital.com) (Source: www.attomdata.com) (Source: www.zillow.com). We also leverage recent news articles (Reuters, Axios), and professional insights (NAR Magazine, ATTOM blogs).

By cross-examining multiple perspectives, this report aims to equip appraisers, regulators, and researchers with a thorough understanding of the Data Accuracy & Completeness Matrix in 2025. In the sections that follow, each data source is examined in depth, supported by quantitative evidence and expert commentary.

MLS Data for Appraisals

The Multiple Listing Service (MLS) is widely regarded as the primary data source for comparable sales in residential appraisals. MLS systems, operated by local Realtor associations, require members (licensed brokers/agents) to enter listings and sales, sharing them with cooperating brokers. Key features of MLS data:

• High granularity: MLS records include detailed property attributes (bedrooms, bathrooms, square footage, lot size, year built, amenities, etc.), often more so than public records. Photo galleries and narrative descriptions add context.
• Up-to-date status: Agents update listings in real time (e.g. price changes, under-contract, pending, closed) as soon as events occur. This means MLS reflects the latest market conditions.
• Accuracy by design: Because MLS data drives commissions and marketing, agents have strong incentives to enter data correctly. MLS organizations enforce rules and audits to minimize errors. An NAR article emphasizes that “everyone who uses MLS data counts on it being complete and accurate”, urging agents to “touch every field” when entering a listing (Source: www.nar.realtor).
• Widespread usage: A CoreLogic analysis (2017) found that MLS was the source for about 90% of comps used in appraisals: “nearly nine out of ten of the comparable sales used by appraisers contained a direct reference to the local MLS as the source of data.” (Source: www.mortgagenewsdaily.com). Likewise, MLS is central for broker price opinions and CMAs (Source: www.nar.realtor).
• Policy compliance needed: Because MLS is cooperative, all brokers in an area theoretically should list active properties in the local MLS. Participation is voluntary but near-universal among professionals. Redfin’s SEC filings note that access to listings depends on membership in local MLS boards: “We get listings data primarily from MLSs… We also source listings from public records, other third-party providers, and individuals.” Importantly, “MLS participation is voluntary.” Broker choices (e.g. pocket listings) can thus limit MLS coverage (Source: www.sec.gov).

Accuracy and Consistency of MLS Price Data: MLS‐reported sale prices are often treated as de facto truths by appraisers, but studies caution that MLS prices may sometimes deviate from official recorded prices. A detailed study by Allen, Lusht, and Weeks (published in The Appraisal Journal) compared 400 sales in a Southeastern market (2004-08) by matching MLS‐reported sale prices against HUD‐1 closing statements. They found MLS and HUD prices differed about 8.75% of the time: specifically, 6.25% of transactions had MLS prices higher than HUD (“overstated”) and 2.50% had MLS lower than HUD (Source: theamericangenius.com). The average overstatement was 6.69% of HUD price (~$14,038 on a $210k home), with the largest single discrepancy 21.44% (Source: theamericangenius.com). This indicates that while the vast majority (≈91.25%) of MLS prices matched closing prices exactly, a small share had non-trivial errors. The errors tended to cluster when market conditions were shifting, suggesting possible bias around peak price turning points (Source: theamericangenius.com). In practice, modern appraisal review tools flag any discrepancy >1% (Source: www.mortgagenewsdaily.com) for further check. Notably, the Center for FHFA/U.S. Treasury reports that over 91% of comps matched exactly in a large 2015–16 California sample (Source: www.mortgagenewsdaily.com) (mirroring the 91.25% match found by Allen et al. for 2004–08). These studies collectively suggest that MLS data is highly accurate on average, but appraisers should verify any outliers.

Data Fields and Completeness: MLS listings generally contain all key attributes relevant to a sale. Essential fields such as final sale price, address, sale/list dates, and geolocation are mandatory in most MLS systems. Many MLS also capture additional details: financing type, buyer’s agent, occupancy code, etc. Agents often update status fields (e.g. “pending”, “active under contract”, “closed”) which aid in tracking time on market. By contrast, public records lack most of these granular details. Therefore, MLS often exceeds public data in completeness of fields for comparable analysis.

One caveat: MLS is chaptered by locale. Large metro areas may have multiple overlapping MLS feeds, and national aggregators must license each. Redfin notes membership in “more than 130 MLSs” across the U.S. (Source: www.sec.gov). MLSs vary in naming conventions and data fields, so merging them requires careful normalization. However, from the appraiser’s perspective, any MLS in the subject’s market usually provides the richest data set of recent comps.

Coverage Limitations: MLS does NOT capture every property sale. Off-market transactions (FSBO, short sales not listed publicly, sales by attorneys, etc.) generally do not appear in MLS. NAR data shows FSBO sales are now a small share (~7% in 2024) (Source: www.nar.realtor), but these constitute nearly all of non-MLS sales. In some states or markets, small numbers of brokerages practice pocket listings (keeping a home signed without MLS entry). Recent policy changes may expand this: for instance, NAR’s new “delayed listing” rules allow sellers to hold properties off MLS initially (Source: www.axios.com). One study in the Phoenix market found that off-MLS recorded sales typically closed at about $4,200 lower than those listed on MLS (Source: www.axios.com). This implies that missing off-market sales can bias comp sets upward. Over a year, even a 7–13% gap (as observed in Tyler Wood’s Big Bear analysis (Source: www.tylerwoodgroup.com) in total sales means MLS indices could overstate average market values.

MLS Data Quality Controls: MLS organizations invest in quality control. For example, the ARAB MLS blog outlines metrics for accuracy, completeness, consistency, timeliness, etc. to score data quality (Source: arabmls.org) (Source: arabmls.org). NAR also educates agents on avoiding listing errors (Source: www.nar.realtor). Nonetheless, deliberate manipulation (“MLS spraying”, false inflation of prices, stale data) has been noted as an industry issue. Poor data entry can mislead users, as what MLS calls “the truth” essentially depends on user inputs (Source: theamericangenius.com) (Source: www.nar.realtor). Appraisers must remain vigilant: if a comparable’s MLS price seems anomalous, one should verify via other sources (closing statement, county record). Automated audits are now common – CoreLogic notes that appraisal review systems automatically flag material price discrepancies between the reported price and robust data sources (Source: www.mortgagenewsdaily.com).

In summary, MLS is the preeminent source for up-to-date, detailed listing data. It provides near-real-time insight into recent market activity. Its data is generally trustworthy (over 90% price-match with official records in multiple studies (Source: www.mortgagenewsdaily.com) (Source: theamericangenius.com). But completeness gaps (off-market sales) and occasional input errors mean MLS alone cannot tell the whole story. An informed appraiser uses MLS as the starting point but corroborates key comparables with public records or other data where possible.

Public Records Data for Appraisals

Public property records encompass data from government sources, primarily the County Recorder/Clerk and Tax Assessor offices. These records have distinct characteristics compared to MLS:

• Legal Source of Sales: By law, when real estate changes hands via a deed, that transaction is recorded in the county where the property lies. Thus, public records theoretically capture all legal conveyances—even those not listed on MLS** (Source: www.tylerwoodgroup.com). Properties sold as FSBO, by litigation, or via out-of-area agents must record the deed. Tyler Wood notes that “In a perfect world, public records will catch all…sales that occur, regardless of which MLS it is in, or if the property is sold by owner.” (Source: www.tylerwoodgroup.com). This makes the records a gold standard for total market counts: in the Big Bear analysis, public records showed 13–12% more sales than MLS over three years (Source: www.tylerwoodgroup.com).

• Mandatory Data Elements: County deeds typically report grantor/grantee names, sale date, and often sale price (depending on state law). Tax assessor databases add attributes like assessed value, land area, building square footage, and owner mailing addresses. In disclosure states, the sale price on the deed is reliably available. In non-disclosure states, sale prices are not recorded in public files. For example, Texas law excludes sale price from recorded documents: “Sales prices are not listed in any publicly recorded documents nor is the sales price shared with a governmental agency…” (Source: retipster.com). In those states (roughly 12–15 states as of 2025), public records can’t furnish prices at all. Instead, lenders or assessment offices rely on appraisals or AVMs to estimate those values. Even in disclosure states, the recorded price may sometimes be obscured by seller credit, or mis‐recorded.

• Updatedness: County recording often occurs a few days to weeks after closing, depending on workload. Some jurisdictions have modern e‐recording systems, but others still rely on manual data entry or scanning. Thus, public record sales data is typically lagging behind real time. For example, a sale in March might not appear in the official record until April or later. Appraisers must account for this delay. Many counties now provide online search or bulk data feeds, but availability varies widely.

• Field Quality and Completeness: Public records generally lack many of the granular fields needed by appraisers. Key property characteristics (bedroom count, improvements, property condition, exact square footage) are often missing or outdated. Assessor rolls may list an old square footage or an outdated building sketch. Zillow notes that data from county sources is often incomplete: “Data might be missing because counties did not collect all the required data or errors were made during the key-in process.” (Source: www.zillow.com). The ClearCapital blog emphasizes that “assessor data” (square footage, age, etc.) is crucial for valuation accuracy (Source: www.clearcapital.com), implying that gaps in assessor records can lead to incorrect appraisals. In practice, appraisers seldom trust a county square footage to be perfectly accurate; they often verify measurements or use MLS/builder records for that. On the positive side, public records do reliably include some essential elements (legal description, lot dimensions, tax law classifications) and, importantly, they are the official record of ownership and transaction.

• Errors and Limitations: Even core fields can be wrong. Clerical errors, revamped parcel boundaries, or owner changes can cause inconsistencies. Moreover, liens and mortgages are recorded with deeds, which means mortgage amounts (if publicized) are visible. Several studies point out that these financing figures can at best be approximate stand-ins for sale price. In the Texas example, only the mortgage amount is visible if one existed (Source: retipster.com), which may not equal the purchase price. Data aggregators often have to infer sale price from deeds of trust or from nearby comparables.

• Accessibility: Most public record data is nominally free or low-cost, but varies by county. Many counties publish deeds and tax rolls openly online; others require fee‐based requests or physical visits. In contrast to MLS (which requires a brokerage membership), public data is generally open to anyone (except some privacy restrictions). Some states sell standardized datasets. Aggregators frequently license the entire set of records from hundreds of counties (as ATTOM claims (Source: www.attomdata.com), transforming scattered public data into appraiser-friendly products.

Despite these limitations, public records yield critical checks and balances for appraisers. A sale price reported on public record is the legally binding value attached to a transaction (except when withheld). A mismatch between MLS price and public deed price (where available) can prompt a review. According to CoreLogic’s findings, when MLS and public prices differ, MLS tends to be slightly lower on average. In their California sample, among the 9% of cases with disagreement, MLS price averaged about $605 (0.085%) below the public-record price (Source: www.mortgagenewsdaily.com). This suggests that MLS listing agents may underreport the ultimate closing price slightly (perhaps reporting contract price or including concessions). However, only ~2.7% of comps had MLS > public prices (Source: www.mortgagenewsdaily.com). In many cases, the difference was negligible (<0.25%, i.e. a few hundred dollars on a typical property) (Source: www.mortgagenewsdaily.com).

Non-Disclosure States and Pricing Gaps: Appraisers must be especially cautious in non-disclosure states. Here, the only authoritative sale price may be what the MLS or parties report. Zillow’s analysis stresses that non-disclosure laws create “barriers to finding recent sales comps.” Communities in these states often rely on professionals with MLS access to share sale prices. In fact, one workaround is to work with a cooperating agent: “The simplest way to access the MLS without becoming licensed is to establish a relationship with someone who is.” (Source: retipster.com). Public records in those states will show the deed but no price. Overall, the absence of price data means appraisers in such markets rely even more on MLS and third-party sources for sale information, underscoring the importance of the other data sources.

Implications of Public Record Gaps: Public records do capture private sales (e.g. “pocket listings” that closed via deed). Thus, they often reveal more transactions than MLS. In the Big Bear example, 572 sales were recorded vs 496 in MLS in the first 9 months of 2008—a 13% jump (Source: www.tylerwoodgroup.com). Over multiple years, MLS data undercounted sales by 8–13% (Source: www.tylerwoodgroup.com). This suggests that relying solely on MLS could understate market activity. However, not all missing sales would have been good comparables (some FSBO sales may not be arm’s-length, or trust assumption, etc.), so the practical impact on valuation depends on context.

In summary, public records provide comprehensive market totals and the legal sales price (where disclosed), but often lack the rich descriptive data of listings. They serve as a crucial reference: appraisers check public records to ensure no transactions were overlooked by MLS searches, and to confirm sale dates and prices. Yet, data completeness flags abound – particularly attribute data (square footage, etc.) and sale price in certain jurisdictions. Appraisers must hedge by combining public information with MLS and other sources to assemble a full picture of each comparable.

Third‐Party Aggregated Data for Appraisals

Third-party real estate data providers arose to “fill gaps” by combining multiple sources and offering value-added analytics. These include major tech companies (e.g. Zillow, Redfin), data vendors (CoreLogic, ATTOM, Black Knight/Veros), and niche data firms. Their propositions to appraisers typically emphasize:

• Unified access: Instead of visiting dozens of county websites or multiple MLS systems, appraisers can query a single platform. For example, ATTOM advertises a REST API with “property, deed, mortgage, foreclosure, and neighborhood information updated daily” (Source: appraisersforum.com). Zillow/Redfin provide extensive online property details (often free to consumers).
• Expanded fields: Beyond raw records, many aggregators infer or append data. This includes key metrics (AVM values, Zestimates, Redfin Estimates), improvement data (permits, renovations), neighborhood trends (price indices like Zillow’s HVI), and risk factors (flood zones, crime, etc.). CoreLogic’s “Total Home Value” index and Black Knight’s str تکنologies are examples.
• Data cleaning: Vendors argue that they “clean up” public/private data to reduce errors. For instance, ATTOM states it “brings together and cleans up public and private real estate data to eradicate inconsistencies” (Source: www.attomdata.com). This suggests they correct obvious typos, unify naming conventions, de-duplicate entries, and fill gaps by inference.
• Coverage breadth: Aggregators often claim near-100% national coverage. For example, ATTOM boasts “assessor data for more than 158 million properties” in 3,000+ counties (Source: www.attomdata.com). Zillow similarly compiles data on ~110M U.S. homes (Source: www.zillow.com). This far exceeds the ~2 million sample size discussed in CoreLogic’s analysis, implying these services index historical records as well as new ones.

Accuracy and Data Quality: While third-party platforms offer convenience, their accuracy is only as good as the underlying sources. Zillow’s Tech Hub blog openly discusses the challenges: public records are heterogeneous, with “data missing” from many counties and key-in errors during data entry (Source: www.zillow.com). When public records datang incomplete, aggregators may impute or extrapolate missing values (e.g. using neighborhood medians for missing square footage). This can introduce non-random errors. Zillow notes that “errors in public records data” must be addressed through normalization and validation (Source: www.zillow.com) (Source: www.zillow.com). Inaccuracies in public data thus propagate into AVMs and home price indices unless corrected.

MLS data licensing also constrains aggregator accuracy. Many platforms license MLS feeds (e.g. Zillow’s data feed agreements with boards). However, if an MLS restricts access or limits fields, or if membership lapses, aggregator feeds may suffer omissions. Redfin’s filing acknowledges that MLS rules “typically do not contemplate multi-jurisdictional online brokerages”, leading to uneven rules and possible compliance issues . In some cases, broker demand for disconnection has led to pulled MLS data (e.g. some brokerages in California in 2021). Such restrictions could cause third-party databases to lose or delay updates for certain areas.

Empirical Performance: Third-party AVMs can be evaluated by comparing their estimates to actual sales. While full analysis of AVM accuracy is beyond this report’s scope, general findings are instructive. Nationwide, Zillow reports its median Zestimate error (the typical deviation from actual sale price) around 1–2% for homes on the market, but higher for off-market and unique properties (Source: www.zillow.com). Redfin States (and others) publish similar metrics. These small median errors imply that on average, third-party models align closely with true values, but the tails can be wide (as the Allen et al. study showed for MLS vs HUD: a few homes had 20%+ error (Source: theamericangenius.com).

Example – Zillow’s Data Strategy: Zillow’s “Solving the Challenges of Public Records Data” (2016) sums up many key points. It describes Zillow’s “living database of over 110 million U.S. homes”, built from county records, MLS listings, rental listings, and mortgages (Source: www.zillow.com). It highlights two quality issues: completeness (county data often incomplete/missing fields) and consistency (different formats and labeling). This necessitates data engineering: normalization of addresses, deduplication of records across counties, and cross-checking with real estate feeds. The article also notes that Zillow updates Zestimates daily to incorporate new data, reflecting a constant integration pipeline (Source: www.zillow.com).

Attom and Other Vendors: ATTOM’s marketing places emphasis on the breadth and cleanliness of its data. In its 2024 blog, an ATTOM executive writes: “As an appraiser, the quality of your information impacts the accuracy of your valuations.” The post touts that its “unique, reliable assessor data” improves appraisal speed and accuracy (Source: www.attomdata.com). ATTOM’s property reports include not only assessor and deed data, but also tax liens, mortgages, and even foreclosure status (Source: www.attomdata.com) (Source: www.attomdata.com). Realist/CoreLogic similarly provide combined MLS+public interfaces. These services can reduce manual search time. For instance, ATTOM claims appraisers can query millions of local sales in seconds through their cloud platform (Source: www.attomdata.com), whereas traditionally an appraiser might have to manually request records or scour multiple MLS modules.

However, one must note ATTOM and peers are commercial entities with self-interest. Their claims should be weighed against independent findings. While ATTOM says it has 158M property assessor records (Source: www.attomdata.com), it does not guarantee every record is error-free. Likewise Zillow’s assurances of data “accuracy” come with caveats about sources.

Pros and Cons Compared to MLS/Public: Third-party data can substantially improve completeness by supplementing MLS with non-MLS sales and supplementing public records with AVM estimates. For example, if a subjective field is missing (like property condition), some AVMs use statistical proxies (like age + tax assessment data) to estimate value. Also, third-party platforms могут integrate non-traditional data (e.g. school ratings, local economic indicators) which could inform an “investment” appraisal perspective.

But aggregators also introduce new biases. If an AVM relies heavily on prior MLS comps, it will reflect any systematic tracking bias therein. If county tax data is stale, the AVM might undervalue appreciating assets. Tools that automatically suggest comparables (like Zillow’s feature “sale price nearby”) sometimes produce questionable matches due to data mismatches. Without local market context, an algorithm may pick dissimilar comps (e.g., older homes in another neighborhood with a similar recent sale price).

Third-Party Data Accessibility: For appraisers, one advantage is ease of pull: many suppliers provide subscription access to nationwide data. Some public record aggregators (e.g. DataTrace, LandVision) focus on title work, but the trend is more appraisal‐focused products. For example, CloudCMA and RPR (Realtors Property Resource) are broker tools that mix MLS and public data for agent comps; appraisal tools like DataMaster interface with MLS and can pull prior sales.

However, quality often varies by region. In a less-populated area, a third-party may have sparser data (if MLS coverage is spotty or county isn’t in their pipeline yet). In contrast, major metros are well-covered. Appraisers should independently verify that a vendor’s data is up-to-date for their locale.

In summary, third-party aggregated data offers breadth and analytical power, but comes with its own quality considerations. These tools can dramatically accelerate data gathering and highlight trends, yet prudent appraisers use them as supplements—not replacements—for primary sources. Cross-referencing a proposed comp’s details across MLS, county, and a paid aggregator can confirm data integrity. As one appraisal educator put it, if an MLS data source “goes down,” one should have contingency plans (like backup databases) (Source: www.mckissock.com). In practice, appraisers often use a hybrid approach: start with MLS, validate with public, and leverage third-party for large-scale analysis or missing fields.

Data Accuracy and Completeness: Comparative Analysis

To synthesize the above discussion, we compare MLS, public records, and third-party data along key dimensions. Below we tabulate these differences explicitly:

The table highlights that accuracy (especially of sale price) is highest when multiple sources agree. Studies show MLS and county prices align ~91% of time (Source: www.mortgagenewsdaily.com) (Source: www.mortgagenewsdaily.com), but neither source is perfect. Completeness is split: public records win on coverage, MLS wins on detail, third-parties vary by provider.

Appraisers should leverage this matrix when selecting data. For example, an appraiser might rely on MLS for fresh comps and on public records for exhaustive sales count. A third-party tool can fill in missing fields (e.g. a tax lot area) or suggest AVM-based comps, but its output should be cross-verified. The modern expectation (per FHFA) is that appraisal data be “verifiable by multiple sources”. Using at least two of these sources per comp is prudent.

Case Studies and Real‐World Examples

To illustrate these concepts, we examine several empirical analyses and real-world anecdotes:

• MLS vs. Public Records Price Study (CoreLogic, 2017): CoreLogic Insights reviewed ~156,000 comparable sales from 2015–2016 in California. It found 91% identical sale prices between MLS sources and county recorder data (Source: www.mortgagenewsdaily.com). Differences, when present, were tiny: in 6% of comps the price differed by under 0.25% (<$1,000 on a $400k home), and only 2.1% differed by ≥1% (Source: www.mortgagenewsdaily.com). Notably, in disagreements, MLS prices averaged slightly lower (mean $605 lower, or 0.085%) than recorded prices (Source: www.mortgagenewsdaily.com). CoreLogic concluded that MLS data are largely reliable, and that “automated appraisal review tools will…alert…when detecting a material discrepancy” (Source: www.mortgagenewsdaily.com). This study underpins the conventional view that appraisers using MLS comps can be confident of price accuracy, though some vetting is still required on the rare outliers.

• MLS vs. Public Price Study (Florida Gulf Coast University, 2015): Wallace, Lusht, and Allen (2015, The Appraisal Journal) compared 400 sales (Florida) from 2004–2008. They matched MLS‐reported prices to HUD‐1 closing prices. Results: MLS prices differed 8.75% of the time, with 6.25% overstating the HUD price and 2.5% understating (Source: theamericangenius.com). The average IRS-overstatement was $14,038 (6.69%) on the reported price. These discrepancies clustered around a market peak, suggesting strategic listing inflation. Importantly, the overall match rate ~91.25%, very close to the CoreLogic figure. However, this study found larger errors (some over 20%). Appraisers should note that in a hot market, MLS may occasionally mislead, so corroboration is key. This study’s headline in The Real Daily (“How serious are errors in MLS…?” (Source: theamericangenius.com) underscored that even MLS data can warrant verification when suspicion arises.

• FSBO and Hidden Sales (NAR, 2024): According to the NAR’s 2024 Profile of Buyers & Sellers, only 7% of home sales were FSBO (Source: www.nar.realtor). This is an all-time low, meaning agents handled the other 93%. For appraisers, this implies MLS should cover most sales, but also that FSBO (7%) and off-MLS deals are not negligible. FSBO transactions tend to close at lower prices (median $380k vs $435k with agents) (Source: www.nar.realtor), which can skew market medians. When those FSBO homes are not visible on MLS, relying solely on MLS can overestimate local values. The NAR analysis also notes that nearly 40% of FSBOs sold to acquaintances, often for even less (nearly $100k less) (Source: www.nar.realtor). As a case in point, an appraiser unaware of nearby FSBOs at a discount might overvalue a subject. Thus, while FSBOs are shrinking, appraisers should still consider local knowledge and public records to catch them.

• Off‐MLS Listings (Phoenix 2025): A recent Axios report highlighted NAR’s new “delayed listing” policy, allowing sellers to hold their homes off public view initially (Source: www.axios.com). Zillow research cited in the article found that in Phoenix, homes sold off the MLS typically fetched $4,200 less than comparable MLS-listed homes (Source: www.axios.com). If this broader phenomenon holds, future appraisal data may become less complete. For example, a high-end seller might delay listing to keep price expectations down or avoid scrutiny. Appraisers will need to track local MLS policies and may have to use techniques like querying “coming soon” records or MLS “office exclusive” fields.

• Regional MLS Gaps (Big Bear, CA): In 2008, Realtor Tyler Wood analyzed local sales in Big Bear Lake, CA. He found MLS reporting 13% fewer sales than public records for Jan–Sept 2008 (496 MLS vs 572 public) (Source: www.tylerwoodgroup.com). Over several years, MLS underreported sales by 8–13%. The gap was attributed to out-of-area agents not entering listings and owner‐sales not on MLS (Source: www.tylerwoodgroup.com). Despite being dated, this case underscores a perennial issue: a region’s MLS may not capture all market activity, especially in resort or retirement areas where FSBOs and absentee owners are common. It also validates the idea that public records will catch those missing sales (Source: www.tylerwoodgroup.com), which is crucial when analyzing a market’s full breadth rather than just MLS view.

• MLS Data Quality Initiatives (NAR 2025): A April 2025 NAR Magazine article by Donna Halfpenny emphasized that “MLS data helps…appraisers obtain market data and comparable sales” and urged careful data entry (Source: www.nar.realtor). It warns that “everyone who uses MLS data counts on it being complete and accurate” (Source: www.nar.realtor). This reflects a growing awareness: legislatures and regulators (e.g. the Appraiser Qualifications Board) expect high data standards. The article’s advice implies that errors (wrong bed count, stale status, etc.) are not mere nuisances but affect critical decisions. In practice, appraisers benefit from these industry efforts: better agent compliance means fewer blatant errors. However, it also serves as a reminder that if an appraiser does find inconsistent MLS info, it’s likely due to human error and should be corrected from another source.

• Third‐Party Data Enrichment (ATTOM Case): ATTOM frequently showcases case studies of appraisers using its property reports. For example, an ATTOM blog writes about an appraisal firm reducing data collection time by integrating ATTOM’s APIs (Source: www.attomdata.com). In one case, an appraiser used ATTOM to retrieve assessor and recorder data for a rural county where the local MLS had limited info. The result was a more robust comp set than relying on one MLS feed alone. While vendor “success stories” should be taken with skepticism, they illustrate how aggregated data can unlock time savings. Appraisers have reported that being able to download sales history across counties (something public records support but MLS cannot) helped them find comps in adjacent markets. Of course, one must still assess that data critically—if ATTOM’s county data is old, the appraiser may overlay MLS notes to update it.

• International Digital Records (Blockchain Pilot): Though not yet mainstream in the U.S., innovative examples exist globally. South Africa launched a pilot blockchain-based property registry (2023) (Source: www.onlinemarketplaces.com), aiming for transparent, tamper-proof records. If such technology spreads, it could redefine “public records” in the future: instantly verifiable transfer history with built-in price data. An American regulatory context isn’t there yet, but it hints at a future where deed data might become real-time and authenticated, solving many county-record lags and errors. Appraisers should watch such developments; in a blockchain scenario, “public records” could become far more timely and complete than today.

These cases highlight that data accuracy and completeness are not matters of opinion but of evidence. Appraisers using any source must cross-check and be aware of systematic quirks (e.g. MLS tends to slightly underreport prices (Source: www.mortgagenewsdaily.com); public records can’t show price in Texas (Source: retipster.com). The overall message is consistent: use multiple sources to confirm facts. As the CoreLogic analysis put it, when MLS vs public diverge, automated tools flag it for “appraisal reviewers and underwriters” (Source: www.mortgagenewsdaily.com). In the field, it falls to the appraiser to resolve the discrepancy.

Implications and Future Directions

The interplay of MLS, public, and third-party data has direct consequences for appraisal practice, regulatory policy, and market transparency. We discuss key implications and look ahead to evolving trends:

1. Reliance on Data Quality Over Source: All stakeholders increasingly acknowledge “data is king” in appraisal. Experts note that “access to accurate, current, and historical data can fill gaps left by other tools and ensure accuracy” (Source: www.clearcapital.com). Indeed, the Uniform Collateral Data Portal (UCDP) and UAD initiatives by FHFA owe their existence to the need for standardized, reliable appraisal inputs. Lenders are investing in PropTech to reduce appraisal delays and revision risk. This focus means that discrepancies among sources cannot be ignored. Errors that once might have been glossed over are now caught by algorithmic checks.

2. Consolidation of Data Platforms: We see a trend toward bigger data platforms. Zillow Group’s acquisition moves (Redfin, etc.) and mergers (Black Knight acquiring Optimal Blue) aim to create one-stop shops. For appraisers, this could simplify workflows (one login for multiple counties). But caution: consolidated datasets still need vetting. The FTC’s April 2025 scrutiny of Zillow/Redfin partnerships (FTC v. Zillow lawsuit) reveals tensions in how data is licensed and shared (though about rentals, it signals regulatory attention to data power). Appraisers should remain independent; if a single platform claims it has “all the answers,” one must compare with source data.

3. Regulatory Changes: Policies now directly affect data availability. The NAR “coming soon” rule (2025) and similar local policies introduce new classes of transactions (private and delayed listings). If many sellers use them, MLS data may capture fewer comps initially. Appraisers might need to rely more on historical records or even buyer-seller disclosures. Conversely, more online transparency is mandated for appraisal reviews (e.g. Delivery of appraisals electronically to GSEs). The Biden administration’s push for digital government suggests future counties may need to modernize records (OMB’s “deliver a digital-first experience” memo, 2023 (Source: bidenwhitehouse.archives.gov). If counties standardize and expedite data sharing, public records could become far more appraiser-friendly.

4. Technology Integration – AI and Big Data: Machine learning models (AVMs, hybrid valuation) are increasingly part of appraisal review and even initial valuation. A recent survey of AI in appraisal highlights multimodal approaches (images + structured data) (Source: arxiv.org). These rely on massive training data drawn from aggregated sources. In effect, MLS and public data feed neural networks. The more accurate/comprehensive that training data, the better the AI. We can expect “AI‐augmented appraisals” to not only cross-check numeric fields but also flag anomalous property images or descriptions. However, overreliance on AI without data ground-truth could perpetuate biases.

5. Focus on Completeness – The UAD 3.6 Example: Regulatory guidelines (e.g. the Uniform Appraisal Dataset) now specify exactly which fields must be reported for each comparable. Appraisers thus need all those fields filled. If MLS lacks one (say, view quality), they must note it. The new UAD training resources (released 2025) emphasize completeness. Future practice may even demand citing source for each field (MLS/report). This could create incentives for better data fidelity in MLS and public records. Conversely, if gaps remain, UT’s new UAD might allow certified appraisers to use next-best info (e.g. an ATTOM-provided field).

6. Potential “Convergence”: An ideal future might see the three sources converge. Some jurisdictions have considered requiring (or already require) recording certain listing details with deeds. Real estate commissions could mandate that any sale (even by owner) be reported in a public database with a few key fields. Meanwhile, MLSs might expand their coverage by allowing more “office exclusive” practices while still capturing sales headcounts (as in Redfin’s regulatory disclosures (Source: www.sec.gov). Third-party vendors may incorporate blockchain or unique identifiers to reduce duplication. If a national property ID system were adopted (as some countries use), it would tremendously help join disparate data.

7. Risk Management Implications: From a lender perspective, data gaps translate into risk. An appraisal based on incomplete data could lead to mispriced collateral. For example, not knowing a neighboring home sold big (off-market) could bias comps down. Uber-accurate public data (enabled by open data initiatives) may alleviate this. FHFA’s move to publish aggregate appraisal stats (UAD dashboards (Source: www.fhfa.gov) shows an appetite for transparency. Perhaps next steps include anonymized public reporting of appraisal "errors" vs actuals, akin to AVM error stats – which would indirectly pressure data accuracy.

8. Broader Market Impacts: At the macro level, if MLS undercounts sales (as observed by Tyler Wood (Source: www.tylerwoodgroup.com), market indices might appear softer. Public analysts (Case-Shiller, FHFA index) rely on broad data. If appraisals lean heavily on MLS, micro markets could show artificial strength. Recognizing these distortions is crucial, especially in rural or developing markets. Tools like Redfin (with their own housing market data) could bring more public records into publicly visible indices.

In conclusion, the gap between MLS, public, and aggregated data will likely narrow as technologies evolve and policies adapt. However, fundamental differences will remain: MLS will always be real-time and agent-curated; public records will always be the legal source of transactions; third parties will always aggregate. Appraisers’ best practice will continue to be a judicious blend of all three: use MLS for the latest and richest comps, public records for completeness, and third-party tools for analytics and efficiency. The future holds more data, not necessarily error-free, making critical appraisal skepticism and verification skills more important than ever.

Conclusion

This report has surveyed the data accuracy and completeness landscape facing appraisers in 2025. Multiple Listing Services (MLS) remain the cornerstone for contemporary comparable sales data: appraisers cite MLS as the source for nearly 90% of comps (Source: www.mortgagenewsdaily.com). MLS listings provide rich, timely details that are hard to match. At the same time, Appraisal Journal and industry analyses consistently find that MLS prices agree with official public records in roughly 91–92% of cases (Source: www.mortgagenewsdaily.com) (Source: theamericangenius.com), validating MLS reliability in most transactions.

Nevertheless, no source is perfect. Public records give the full universe of sales, uncovering FSBO and off-market deals that MLS misses (Source: www.tylerwoodgroup.com) (Source: www.tylerwoodgroup.com), but often without sale prices (in non-disclosure states) (Source: retipster.com) and with lagging or incomplete property details. Third-party aggregators promise to fill these gaps by combining MLS and public data (and more) in one place (Source: www.attomdata.com) (Source: www.zillow.com). They can dramatically speed data gathering and offer advanced analytics, yet they inherit the imperfections of their inputs (Source: www.zillow.com) (Source: www.clearcapital.com).

Our comparative matrix (Table 1) highlights how accuracy tends to be highest when corroborated across sources, while completeness often requires multiple sources. Case studies illustrate this vividly: in one California analysis, 91% of MLS comps exactly matched county prices (Source: www.mortgagenewsdaily.com), yet errors did occur and needed flags. FSBO sales (NAR: ~7%) (Source: www.nar.realtor) and new “coming soon” practices (Source: www.axios.com) show that appraisers still must go beyond MLS. Leading data vendors (ATTOM, CoreLogic) emphasize “data is king” (Source: www.clearcapital.com), urging appraisers to leverage all available feeds.

Going forward, appraisers will increasingly rely on a multi-pronged data validation approach. Automated appraisal review systems are already common, and they typically derive confidence scores by cross-checking MLS-derived figures against normalized public records (as CoreLogic’s tools do (Source: www.mortgagenewsdaily.com). The industry trend is clear: integrate, cross-verify, and automate. For example, a cause for alarm would be a 10% discrepancy between an MLS-reported price and tax assessor data; today’s technologies would flag and prompt an explanation.

It is critical that users understand each source’s provenance. Redfin’s disclosures remind us that “MLS participation is voluntary” (Source: www.sec.gov), meaning data streams can change with policy, litigation, or corporate decisions. Appraisers should stay informed about local MLS rules (e.g. on coming-soon listings) and maintain contacts in local associations to fill any information gaps. Meanwhile, government openness initiatives may soon improve public availability of deed data; some states may even consider requiring limited price disclosure in nondisclosure states to enhance transparency (a proposal occasionally discussed in policy circles).

In final summary: MLS, public records, and third-party data each play essential, complementary roles. Relying on any single source risks blind spots. This report’s analysis and the accompanying tables aim to help appraisers weigh the trade-offs. The evidence shows that accuracy is high but not absolute; completeness is broad but not detailed. By using MLS for detail, public records for coverage, and third-party for integration, appraisers can form the most robust valuations. As the market evolves in 2025 and beyond, appraisers who master this data matrix will be best equipped to deliver credible, defensible appraisals.

References: (Data and quotations in this report are drawn from the sources cited inline. Key references include CoreLogic and FGCU studies (Source: www.mortgagenewsdaily.com) (Source: www.mortgagenewsdaily.com) (Source: theamericangenius.com), NAR research (Source: www.nar.realtor) (Source: www.nar.realtor), third-party data providers (Source: www.attomdata.com) (Source: www.attomdata.com) (Source: www.zillow.com), and recent industry journalism (Source: www.axios.com) (Source: www.tylerwoodgroup.com).)