Investigative Review of TransUnion

Financial Breakdown of the Settlement

The company did not admit wrongdoing as part of the settlement. This is a standard clause in such agreements. Yet the magnitude of the payment and the extent of the required changes speak for themselves. The company must operate under a consent order that dictates its behavior for years to come. This order provides the regulators with a method to monitor the company’s compliance. Any future violations of this order could result in even stiffer penalties. The $23 million settlement is a landmark moment in the fight for fair housing. It exposes the of the tenant screening industry. It reveals how sloppy data practices can derail lives. It also shows that regulators are catching up to the technological realities of the background check economy. The days of “black box” screening with zero accountability appear to be numbered. TransUnion must prove that it can operate a tenant screening business that respects the law and the rights of consumers. The rigorous monitoring required by the settlement determine if the company is truly capable of reform. This case also highlights the vulnerability of renters in the digital age. A single database error can render a person homeless. The speed at which these reports are generated frequently the ability of consumers to correct them. The settlement attempts to slow this process down. It forces the company to verify before it reports. It prioritizes accuracy over speed. This shift is essential for a fair housing market. The regulators have drawn a line in the sand. Inaccurate data is not just a customer service problem. It is a violation of federal law. The enforcement action against TransUnion serves as a case study in corporate negligence regarding consumer data. The company profited from selling information it knew or should have known was flawed. It built blocks to correction. It misled consumers about the sources of that information. The $23 million penalty is a receipt for these practices. It is a tangible acknowledgment of the harm inflicted on thousands of renters. The true measure of this settlement be in the improved accuracy of future reports. Until then, the record stands as a testament to the need for vigilant oversight of the credit reporting industry.

The Static Snapshot: Data Ingestion vs. Real-Time Reality

The fundamental technical failure driving inaccurate tenant screening reports lies in the architectural difference between a live court docket and a static commercial database. When a landlord files an eviction petition, the court clerk enters this action into the public record. Almost immediately, automated data scrapers, software bots designed to harvest public records, copy this entry. These scrapers, operated by third-party data aggregators or directly by screening bureaus, capture the “filing” event. This initial data packet contains the tenant’s name, the filing date, and the case type. At this precise moment, the commercial database matches the court’s reality.

The begins seconds later. Court cases are; they evolve. A tenant might pay the rent the day, leading to a dismissal. A judge might rule in favor of the tenant, finding the eviction baseless. In jurisdictions, if a tenant wins or settles, the court orders the record sealed or expunged, legally removing it from the public eye to protect the tenant’s future housing prospects. In the court’s live system, that record or becomes inaccessible to the public. Yet, in the TransUnion Rental Screening Solutions (TURSS) ecosystem, the original “filing” record frequently remains frozen in time. The system ingested the initial accusation failed to return to the source to verify the outcome.

This “snapshot” method creates a zombie record: a piece of data that is legally dead digitally alive. The screening report presented to a prospective landlord shows an active eviction filing. It does not show that the case was dismissed, nor does it honor the court’s seal. The database treats the initial scrape as a permanent fact rather than a temporary status. To the algorithm, the absence of a “judgment” code does not automatically trigger a re-verification; instead, the system defaults to reporting the existence of the case itself, which landlords almost universally interpret as a red flag.

The Vendor Cascade and the Broken Chain of Custody

TransUnion does not send human runners to every county courthouse in America. Instead, the company relies on a complex supply chain of third-party data vendors. The Federal Trade Commission (FTC) complaint from October 2023 specifically identified LexisNexis Risk and Information Analytics Group as one such vendor from whom TURSS obtained eviction proceeding records. This reliance on intermediaries creates a game of “telephone” where data integrity degrades at each handoff. A local court updates its docket; a regional aggregator scrapes that docket; a national vendor buys the regional data; TransUnion buys the national data.

In this cascade, a “seal” order issued by a judge faces multiple blocks. The court might process the seal immediately, if the regional aggregator only refreshes its bulk data purchase once a month, or once a quarter, the sealed record remains visible in the vendor’s feed. TransUnion, situated at the end of this chain, ingests the data provided by the vendor. If the vendor fails to send a “delete” command or a “status update” code, TransUnion’s database retains the original entry. The FTC found that TURSS failed to take reasonable steps to ensure the accuracy of this vendor-supplied data, trusting the supply chain blindly while the actual court records shifted.

This structural blindness is compounded by the financial incentives of the industry. Real-time verification, pinging the court’s live database the moment a landlord requests a screening report, is expensive. It requires sophisticated API integrations with thousands of local court systems, of which charge access fees. Bulk data purchasing is cheap. It allows the screening company to maintain a massive internal library of records that costs fractions of a penny to query. The result is a business model that prioritizes the low cost of stale data over the high cost of accurate, real-time data. The tenant pays the price for this efficiency in the form of rejected applications.

The “Wildcard” Matching Logic and Mixed Files

Beyond the problem of stale data lies the problem of loose matching logic. Eviction records in civil courts rarely contain unique biometric identifiers like fingerprints or full Social Security numbers. They contain a name and an address, and sometimes a partial date of birth. To maximize the number of “hits” found, and thus appear more valuable to landlord customers, screening algorithms use “wildcard” or partial matching. This logic accepts records that are “close enough” rather than exact matches.

If a tenant named “Michael A. Brown” applies for an apartment, the system searches its database for that name. yet, if the algorithm is tuned loosely, it might return an eviction record for “Michael Brown” (no middle initial) or “Mike Brown” living in the same city or county. This creates a “mixed file,” where the innocent applicant inherits the eviction history of a stranger. When combined with the failure to update sealed records, the error multiplies. A “Michael Brown” might have had an eviction filed, then sealed. The applicant “Michael A. Brown” is then flagged not just for a stranger’s case, for a stranger’s sealed case that should not be visible to anyone.

The 2023 enforcement action highlighted that TURSS procedures failed to prevent the inclusion of multiple entries for the same eviction case. a chaotic data environment where a single eviction event, perhaps scraped multiple times by different vendors or at different stages (filing, hearing, judgment), appears as a list of repeat offenses. To an untrained landlord reviewing the report, it looks like the applicant is a serial non-payer, when in reality, they are the victim of a single, resolved, or misattributed legal dispute that the database duplicated rather than reconciled.

The Disposition Disconnect: Filing vs. Judgment

A specific technical failure identified in regulatory investigations is the inability of the screening system to distinguish between a “filing” and a “judgment.” A filing is an allegation; a judgment is a legal fact. In the United States justice system, anyone can file a lawsuit against anyone else. The mere existence of a filing proves nothing regarding the tenant’s conduct. Yet, TransUnion’s reports have frequently displayed filings in a manner that suggests guilt. The system captures the “Eviction Action” tag fails to capture the “Dismissed” or “Judgment for Defendant” tag that follows weeks later.

This error is not passive; it is a failure of database maintenance. When a case is dismissed, the court docket reflects that disposition. A screening system committed to “maximum possible accuracy”, the standard required by the Fair Credit Reporting Act (FCRA), would require a logic gate: If a record is older than 60 days and absence a final disposition, query the source again before reporting. TURSS did not employ such a safeguard. Instead, it allowed open-ended filings to indefinitely. A tenant who defeated an eviction attempt three years ago finds the initial filing still staring back at them on a SmartMove report, with no indication that they won the case.

The persistence of these records is particularly damaging because of how landlords use the data. property management platforms set automated filters to reject any applicant with a “hit” in the eviction database, regardless of the outcome. By feeding unresolved or sealed filings into these automated decision engines, TransUnion usurps the judicial process, enforcing a penalty (homelessness) for a crime (non-payment) that a court of law determined did not happen. The database’s inertia overrides the judge’s gavel.

The load of Correction

The architecture of this error places the entire load of quality control on the victim. Because the system absence an internal self-correcting method for sealed records, the error is only discovered after the damage is done, when a rental application is denied. The tenant must then obtain the report, identify the sealed record, and file a dispute. This process forces the tenant to prove the non-existence of a record that the court has already ordered to be destroyed or hidden.

Even when a tenant successfully disputes a record, the fix is frequently temporary or localized. TransUnion might suppress the record in its own database, if the underlying vendor feed continues to supply the bad data in the monthly bulk update, the zombie record can re-emerge. also, correcting the record at TransUnion does not automatically correct it at the source vendor or at other screening agencies that buy from the same vendor. The data propagates through the ecosystem like a virus, and the “cure”, a dispute resolution, treats only one symptom on one report, leaving the root cause in the vendor’s database untouched.

This pattern of scrape, stagnate, and report creates a widespread barrier to housing. The technical reality is that TransUnion’s systems were designed to ingest data easily purge it with difficulty. The “delete” function is reactive, triggered only by consumer complaints or specific lawsuits, while the “add” function is proactive and automated. Until the ingestion logic is fundamentally rewritten to prioritize real-time verification over bulk accumulation, sealed and resolved eviction records continue to haunt tenant screening reports, both the law and the truth.

A specific, widespread failure within TransUnion Rental Screening Solutions (TURSS) created a digital mirage that destroyed the housing prospects of American renters. This phenomenon, identified by federal regulators as the “multiple entry” error, transformed single, frequently resolved legal disputes into what appeared to be patterns of serial delinquency. For years, the automated systems powering SmartMove and other TransUnion screening products treated sequential updates in a single court case not as modifications to an existing file, as entirely new, distinct adverse records. A tenant facing one eviction proceeding, regardless of the outcome, would frequently see their screening report populate with two, three, or four separate “eviction” entries, branding them as a high-risk repeat offender to any prospective landlord.

The Mechanics of File Stuffing

The error originated in the crude data ingestion methods used to scrape public court dockets. An eviction proceeding is a process, not a singular event. It involves a sequence of legal filings: the initial complaint, a summons, a hearing record, a judgment, and chance a writ of possession or a dismissal. In a properly managed database, these distinct documents share a unique case identifier, allowing the system to merge them into a single, coherent narrative of the event. TURSS, yet, failed to implement basic deduplication logic for these records until at least April 2021.

When a court clerk updated a docket, for instance, recording that a judgment was satisfied or a case was dismissed, TransUnion’s algorithms frequently captured this new piece of data as a fresh eviction record. The system did not overwrite the previous status; it stacked the new entry on top of the old one. Consequently, a tenant who successfully fought an eviction and had the case dismissed might see a screening report listing both the initial “Eviction Filed” and the subsequent “Eviction Dismissed” as two separate negative items. To a landlord reviewing the report, the nuance of the case numbers (frequently identical or differing only by a suffix) was lost in the visual impact of seeing multiple red flags. The tenant did not look like someone who had one legal dispute; they looked like someone who had been evicted multiple times in rapid succession.

The Serial Offender Mirage

The psychological impact of this data presentation on landlords cannot be overstated. Property managers spend less than five minutes reviewing a screening report. Their primary goal is risk mitigation. When a report summarizes a candidate’s history with a list of three “Eviction” records, the immediate assumption is that the applicant is a serial non-payer. The landlord does not investigate the case numbers to see if they match. They do not cross-reference the dates to realize that three “evictions” occurring within a two-month span is legally impossible. They simply move to the applicant.

This “file stuffing” rendered the accuracy of the underlying data irrelevant. Even if the final disposition of the case was recorded correctly as a dismissal (which it frequently was not), the presence of the initial filing as a separate, active record contradicted the resolution. The report created a paradox: the tenant was simultaneously evicted and not evicted, sued and dismissed, all in the same document. For the consumer, correcting this was a bureaucratic nightmare. Disputing one entry might remove the duplicate, yet leave the original filing, or vice versa. Because the system treated them as separate events, the consumer had to prove the non-existence of multiple phantom evictions rather than correcting the status of a single real one.

Federal Enforcement and the 2023 Settlement

The of this negligence triggered a joint enforcement action by the Federal Trade Commission (FTC) and the Consumer Financial Protection Bureau (CFPB). In October 2023, these agencies announced a $23 million settlement with TransUnion, explicitly citing the multiple entry failure as a primary violation of the Fair Credit Reporting Act (FCRA). The complaint filed by the regulators detailed how TURSS “failed to follow reasonable procedures to prevent the inclusion of multiple entries for the same eviction case.”

The timeline revealed in the federal complaint is damning. The agencies alleged that this practice was standard operation for TURSS until April 2021. For years, while TransUnion marketed its SmartMove product as a sophisticated, solution for independent landlords, its backend systems were churning out reports with duplications that a simple database query could have prevented. The settlement required TransUnion to pay $15 million to the FTC and $8 million to the CFPB, the largest amount ever recovered by the FTC in a tenant screening matter. even with the payout, TransUnion settled without admitting or denying the findings of fact, a standard legal maneuver that allows the company to avoid formal liability while paying to close the investigation.

The Human Cost of Algorithmic Laziness

The victims of this glitch were frequently those already in precarious housing situations. Consider a tenant who withheld rent due to uninhabitable conditions, a legal right in jurisdictions. The landlord files for eviction. The court rules in favor of the tenant, and the case is dismissed. In the real world, the tenant won. In the TransUnion database, the tenant had two records: the filing and the dismissal. A future landlord sees the report, notes “Eviction” listed twice, and rejects the application. The tenant, who exercised their legal rights and prevailed, is punished more severely than a tenant who simply left without a fight.

This duplication also compounded the damage of “zombie records”, sealed or expunged cases that should not appear at all. If a court ordered a case sealed, TransUnion’s system might remove the specific record associated with the sealing order fail to delete the duplicate entries generated from earlier stages of the case. The “shadow” of the sealed eviction remained visible, listed as a separate filing that the automated purge logic missed. This failure violated the core FCRA requirement that consumer reporting agencies maintain “reasonable procedures to assure maximum possible accuracy.”

Regulatory condemnation

Samuel Levine, Director of the FTC’s Bureau of Consumer Protection, described the reports produced by TransUnion as being “ridden with errors and based on data from secret sources.” The CFPB’s Director, Rohit Chopra, echoed this sentiment, stating that Americans were put at risk of wrongful housing denials because TransUnion failed to follow the law. The regulators identified that TransUnion did not make random mistakes; they failed to implement the necessary software architecture to handle the complexity of court data. The “glitch” was not a bug; it was a feature of a system designed to prioritize data volume and acquisition speed over data hygiene.

The settlement order imposed strict injunctive relief, forcing TransUnion to overhaul its matching logic. The company is legally required to implement procedures that prevent the display of multiple entries for the same eviction proceeding. They must also ensure that the disposition of the case (e. g., dismissed, judgment for defendant) is and accurately reported. The fact that a federal court order was required to compel a multi-billion dollar data company to perform basic record deduplication speaks volumes about the industry’s voluntary commitment to accuracy.

The Persistence of the Problem

While the 2023 settlement addressed the specific practices of TURSS, the legacy of these duplicate records in the rental market. Data sold by TransUnion prior to the cleanup may still reside in the cached databases of smaller, downstream tenant screening companies that purchase bulk data do not update it regularly. A “multiple entry” error generated in 2020 could theoretically still appear on a report generated by a third-party reseller in 2026 if that reseller has not refreshed their source files. The digital stain of a duplicated eviction is difficult to scrub once it has leaked out of the primary reservoir and into the wider ecosystem of background check vendors.

The “multiple entry” glitch serves as a clear example of how technical incompetence in the data brokerage industry into real-world harm. It transforms the procedural steps of the justice system into a weapon against the accused. By failing to distinguish between a case update and a new case, TransUnion rewrote the history of thousands of renters, presenting them as chaotic, litigious, and risky. The $23 million penalty, while record-breaking, represents a fraction of the economic damage inflicted on families who were denied housing, forced into sub-standard apartments, or compelled to pay higher security deposits due to a phantom record that never should have existed.

The “One-and-Done” Data Scraping Model

The core mechanical failure driving the persistence of dismissed eviction cases in TransUnion reports lies in a data collection method best described as “one-and-Done.” TransUnion Rental Screening Solutions (TURSS) and its third-party vendors frequently use automated scraping algorithms to harvest data from county court dockets. These scripts identify the initial filing of an eviction action, an Unlawful Detainer, and immediately log it into the consumer’s file. This initial capture is highly, ensuring that a landlord’s legal action appears on a credit or tenant screening report almost instantly.

The defect arises in the maintenance of that record. Eviction litigation is; a filing is the opening salvo, not the final judgment. In jurisdictions across the United States, a significant percentage of eviction filings do not result in a judgment against the tenant. They are dismissed, withdrawn by the landlord, or resolved through settlement agreements where the tenant remains in the property. Yet, the automated systems used by TURSS have historically failed to return to the court docket to verify the final disposition of the case. The scraper grabs the “Eviction Filed” status and moves on, leaving the record frozen in time. Consequently, a tenant who successfully defended themselves in court and had the case dismissed three weeks later continues to carry a “Eviction Filed” flag on their TransUnion report for up to seven years.

This failure to update creates a “zombie record”, a piece of data that is technically accurate regarding the past (a filing occurred) materially misleading regarding the present (the case was dropped). Under the Fair Credit Reporting Act (FCRA), consumer reporting agencies must follow reasonable procedures to assure “maximum possible accuracy.” Reporting a filing without its subsequent dismissal violates this standard because it omits the exonerating context. A landlord reviewing the report sees only that the applicant was sued for eviction, leading to an automatic rejection, even though the applicant prevailed in the legal system.

The 2023 Federal Enforcement Action

The widespread nature of this failure was laid bare in October 2023, when the Federal Trade Commission (FTC) and the Consumer Financial Protection Bureau (CFPB) jointly announced a $23 million settlement with TransUnion. The complaint filed by the agencies specifically targeted the company’s failure to update eviction records. The regulators found that TURSS did not employ reasonable procedures to ensure that the eviction data it sold to landlords reflected the current status of public records. The investigation revealed that until April 2021, TURSS frequently reported multiple developments in the same eviction proceeding as separate events or failed to report that a case had been dismissed entirely.

The agencies noted that TURSS obtained eviction records from third-party vendors, such as LexisNexis Risk and Information Analytics Group, failed to validate the accuracy of this data. The chain of custody for this information exacerbated the error rate. A court clerk updates a physical or digital docket; a third-party vendor scrapes that docket (chance missing the update); TransUnion buys the data from the vendor. If the vendor does not push a “dismissal” update, or if TransUnion’s ingestion system does not overwrite the previous “filing” entry, the error. The 2023 settlement required TransUnion to overhaul these procedures, yet for millions of renters screened prior to this enforcement, the damage was already done. The settlement highlighted that TransUnion continued to report “Judgment Amounts” for cases that were actually dismissed, falsely inflating the financial risk associated with the tenant.

Statistical Reality: The High Rate of Non-Eviction Outcomes

The magnitude of this reporting failure is amplified by the statistical reality of eviction court outcomes. Research from the Princeton Eviction Lab and various legal aid organizations indicates that a substantial portion of eviction filings do not end in a removal order. In Cook County, Illinois, data from 2014 to 2017 showed that approximately 39% of completed eviction cases did not result in a judgment against the tenant. In other jurisdictions, dismissal rates can be even higher depending on local tenant protections and the prevalence of “pay and stay” agreements.

When TransUnion reports all filings as negative markers without updating the disposition, the error rate of their product mirrors the dismissal rate of the court system. If 40% of cases in a specific county are dismissed, and TransUnion fails to update those records, then 40% of the eviction data for that county is materially misleading. This is not a marginal error; it is a structural deficiency that disproportionately affects low-income renters who are more likely to face eviction filings also more likely to resolve them through payment plans or legal aid intervention. The “filing-only” reporting model punishes tenants for exercising their legal right to a defense, as the mere existence of the case on the docket, regardless of the outcome, becomes a barrier to future housing.

The “Satisfied” Judgment Loophole

Beyond dismissed cases, a parallel failure exists regarding “satisfied” judgments. In situations where a court does rule against a tenant, the tenant may subsequently pay the judgment amount in full. Once paid, the court record is updated to reflect that the judgment is “Satisfied.” This distinction is important for future landlords, as it demonstrates that the tenant honors their financial obligations even after a dispute. Yet, TransUnion’s reporting systems have frequently failed to capture this status change. A report may continue to list the judgment as “Unpaid” or simply display the original judgment amount without the “Satisfied” indicator.

This specific failure to update transforms a resolved debt into an active financial liability in the eyes of a screening algorithm. Automated rental scoring models penalize outstanding judgments far more severely than satisfied ones. A tenant who paid off their landlord five years ago may still be scored as if they currently owe thousands of dollars. The load then shifts entirely to the consumer to prove they paid, a process that requires obtaining certified court records and engaging in a dispute process with TransUnion that can take 30 days or longer, a timeframe that guarantees the loss of the apartment they applied for.

Case Study: Belluccia v. TransUnion Rental Screening Solutions

The legal battleground has provided clear documentation of these practices. In the class action lawsuit Belluccia v. TransUnion Rental Screening Solutions, Inc., filed in 2021, the plaintiffs alleged that TURSS “regularly and illegally reports eviction information pertaining to cases and judgments that have been dismissed, withdrawn, satisfied, or have resulted in a judgment for the tenant.” The complaint detailed how the plaintiff, a prospective tenant, was denied housing because TURSS reported a prior eviction filing omitted the serious fact that the case had been dismissed.

The Belluccia case exposed the reliance on “distilled” data. The lawsuit argued that TURSS and its vendors “infrequently re-check dockets to find updated dispositions.” This creates a widespread asymmetry: the negative information (the filing) is captured immediately, while the positive information (the dismissal) is ignored. The plaintiffs argued that this practice was not negligent willful, as TransUnion was aware of the nature of court records yet chose a cheaper, static data collection method. The settlement of such cases frequently involves monetary relief also mandates changes to business practices, forcing the credit bureau to implement “reasonable procedures” to verify dispositions before reporting them. Even with these legal mandates, the lag time between a court ruling and a database update remains a serious vulnerability for renters.

Comparative Data Status Table

The following table illustrates the disconnect between the actual legal status of an eviction case and the status frequently reported by TransUnion prior to consumer disputes.

The load of Dispute

TransUnion’s defense in these matters frequently relies on the FCRA’s dispute method, arguing that consumers have the ability to correct errors. This argument ignores the temporal reality of the rental market. In a competitive housing environment, a rental unit is frequently on the market for less than 48 hours. The FCRA allows consumer reporting agencies 30 days to investigate a dispute. By the time TransUnion verifies the dismissal and updates the report, the apartment is long gone. The failure to update records proactively shifts the load of accuracy from the multi-billion dollar data broker to the individual consumer, who must constantly police their own file to ensure that resolved legal matters do not destroy their future prospects.

The “Wildcard” Matching Algorithm

The core of the quality control failure lies in the algorithmic matching criteria used to generate reports. To deliver “instant” results, TURSS employs what regulators describe as “wildcard” or “blind” matching. Instead of requiring an exact match across multiple identifiers, such as full name, full date of birth, and social security number, the system frequently links records based on partial data. A common name combined with a partial date of birth or a matching zip code can trigger a “hit.” This loose matching logic creates false positives at an industrial. An applicant named “James Smith” may be flagged for an eviction judgment belonging to “James A. Smith” or “Jamie Smith” simply because they resided in the same jurisdiction. The Consumer Financial Protection Bureau (CFPB) has explicitly warned that name-only matching violates the Fair Credit Reporting Act (FCRA), yet TURSS continued to use procedures that failed to distinguish between distinct individuals. The system operates on a presumption of guilt: if the data looks similar, it is reported as a match, shifting the load of disproving the record onto the applicant.

Reliance on Unverified Third-Party Data

TURSS does not send human runners to courthouses to verify records. Instead, it purchases bulk data from third-party vendors, such as LexisNexis Risk & Information Analytics Group. This supply chain introduces a “broken telephone” effect where data degradation occurs at every handoff. When a court seals a record or dismisses an eviction case, that update must travel from the court clerk to the aggregator, and then from the aggregator to TransUnion. The breakdown occurs because TURSS failed to implement reasonable procedures to audit these vendors. The October 2023 joint enforcement action by the FTC and CFPB revealed that TURSS did not validate the accuracy of the data stream. If a vendor provided a file containing sealed records, TURSS reported them. If a vendor failed to update a “filing” to a “dismissal,” TURSS reported the filing as an active eviction. also, for years, TURSS failed to disclose the names of these third-party vendors to consumers, preventing victims from correcting the source of the error. A tenant would dispute the error with TransUnion, the bad data would remain at the vendor level, ready to reappear on the background check.

Automated Adjudication: The “Decline” Recommendation

The absence of quality control is most visible in the “SmartMove” recommendation engine. This feature allows landlords to set criteria, such as “no prior evictions”, and receive an automated “Accept” or “Decline” recommendation. This removes human judgment from the equation entirely. When the algorithm incorrectly tags an applicant with a resolved or sealed eviction, the system automatically generates a “Decline” recommendation. Small- landlords, who are the primary target for SmartMove, frequently absence the legal expertise to interpret the raw data or the time to verify it. They rely on the red “Decline” banner. Consequently, a data error becomes an immediate housing denial without the landlord ever examining the underlying file. This automated adjudication amplifies the harm of inaccurate data, as the decision is made instantaneously by a machine programmed to minimize risk for the landlord, regardless of the truth.

Violation of the “Maximum Possible Accuracy” Standard

The Fair Credit Reporting Act, specifically 15 U. S. C. § 1681e(b), mandates that consumer reporting agencies follow “reasonable procedures to assure maximum possible accuracy” of the information concerning the individual. TURSS’s procedures fell short of this statutory requirement. The company argued that its practices were consistent with industry standards. Yet, federal courts and regulators have clarified that an industry standard of mediocrity does not excuse a violation of federal law.

The “Multiple Entry” Failure

Another manifestation of poor quality control was the “multiple entry” glitch. TURSS systems frequently interpreted updates to a single case as separate eviction events. A single eviction proceeding involves multiple steps: the initial filing, a hearing, a judgment, and chance a writ of possession. TURSS frequently reported each of these administrative steps as a distinct eviction record. A tenant who fought one eviction case and won might see their screening report list three or four “eviction records” corresponding to the various docket updates. To a landlord reviewing the report, this looked like a pattern of serial delinquency rather than a single legal dispute. This was not a data error a software logic failure, a refusal to group related records into a single case file. The company knew or should have known that court dockets produce sequential entries, yet their parsing software failed to synthesize this information accurately.

Offloading Quality Control to the Victim

Perhaps the most cynical aspect of TURSS’s operation is its reliance on the dispute process as its primary quality control method. Rather than investing in front-end verification to prevent errors, the company use a “publish, correct later” method. This places the entire load of accuracy on the tenant. For a rental applicant, this retroactive correction is useless. By the time a tenant receives an adverse action notice, identifies the error, files a dispute, and waits the statutory 30 days for an investigation, the apartment is gone. The housing market moves faster than the FCRA dispute timeline. TURSS capitalizes on this temporal disconnect. They sell the report instantly, collect the fee, and incur costs only if the consumer has the stamina to fight back. The $15 million settlement in 2023 was a penalty for this business model, acknowledging that “reasonable procedures” implies preventing errors before they are sold, not just fixing them after the damage is irreversible. The persistence of these errors shows a corporate culture that views accuracy as a luxury product. In the high-volume world of tenant screening, TURSS built a of automation designed to protect landlords from risk, in doing so, it exposed millions of renters to the hazard of unverified, inaccurate, and damaging data.

The Vendor Behind the Curtain: LexisNexis Risk Solutions

The operational core of TransUnion Rental Screening Solutions (TURSS) is not a direct pipeline to county courthouses, a commercial contract with third-party data brokers. For years, the primary engine driving TransUnion’s eviction reporting was not the judicial system, LexisNexis Risk and Information Analytics Group, Inc. (LexisNexis). This relationship, obscured from consumers until federal regulators intervened, reveals a fundamental flaw in the tenant screening industry: the reliance on second-hand data that is frequently stale, incomplete, or categorically prohibited from release. Federal investigations exposed that TURSS purchased bulk eviction proceeding records from LexisNexis to populate its SmartMove and resident screening reports. Rather than retrieving real-time data from the jurisdiction where a case was filed, TransUnion queried a static database maintained by LexisNexis. This distinction is important. When a court updates a file, dismissing a case, sealing a record, or ruling in favor of the tenant, that change is reflected immediately in the court’s live system. It does not, yet, automatically propagate to the third-party aggregator’s database unless a specific update pattern occurs.

The Chain of Custody Failure

The data supply chain operates on a “garbage in, garbage out” principle. The Federal Trade Commission (FTC) complaint filed in October 2023 detailed how this specific vendor relationship compromised accuracy. TURSS procedures allowed the reporting of eviction records without verifying when LexisNexis last updated them. If LexisNexis scraped a court docket on Monday, and the judge sealed the case on Tuesday, a TransUnion report generated on Wednesday would still display the eviction as active and public. This lag creates a “zombie record” phenomenon. The record is legally dead at the source remains alive in the commercial database. The FTC found that TURSS failed to impose specific accuracy requirements on LexisNexis regarding these updates. TransUnion accepted the data stream as truth, monetized it, and delivered it to landlords without independent verification. The reliance on this single vendor point of failure meant that any error in the LexisNexis repository became a widespread error in TransUnion’s product.

Concealing the Source

Perhaps the most egregious violation identified by the Consumer Financial Protection Bureau (CFPB) was TransUnion’s systematic concealment of this third-party source. When consumers requested a copy of their file to dispute an error, TURSS frequently listed the source of the eviction record as the local court (e. g., “Circuit Court of Cook County”) rather than LexisNexis. This misrepresentation crippled the consumer’s ability to fix the problem. A tenant seeing “Circuit Court” as the source would contact the court clerk. The clerk would correctly inform them that no record existed or that the file was sealed. The tenant, armed with this information, would return to TransUnion, only to be told the record was “verified.” Because TransUnion was actually verifying the data with LexisNexis, not the court, the error. The tenant was trapped in a loop, fighting a ghost record held by a vendor they did not know existed.

The “Verification” Loophole

The Fair Credit Reporting Act (FCRA) demands that Consumer Reporting Agencies (CRAs) follow reasonable procedures to assure maximum possible accuracy. TransUnion’s arrangement with LexisNexis outsourced this legal obligation. When a dispute arose, the standard procedure was frequently to ping the vendor’s database again. If LexisNexis still held the erroneous data, TransUnion treated the dispute as resolved in favor of the existing report. This circular verification process bypasses the authoritative source: the public record itself. By treating the vendor’s database as a proxy for the judicial system, TransUnion insulated itself from the cost of manual court checks while exposing renters to the high risk of outdated information. The FTC investigation revealed that TURSS only began disclosing LexisNexis as a source in June 2021, and only after receiving a Civil Investigative Demand from federal regulators.

Regulatory

The $15 million portion of the 2023 settlement specifically addressed these tenant screening failures. The consent order requires TransUnion to implement procedures that prevent the inclusion of sealed records and to accurately disclose the third-party vendors used. This enforcement action confirms that the industry practice of blindly trusting bulk data aggregators like LexisNexis is not a defense against FCRA violations. It is, instead, a documented liability.

Table: The Data Pipeline

The Fiction of ‘Judgment Amounts’

In the lexicon of tenant screening, few data points carry as much weight as the “Judgment Amount.” To a property manager, this figure represents a confirmed financial liability, money a court has legally determined a tenant owes to a previous landlord. It suggests a proven history of non-payment and a direct risk to future revenue. Yet, for years, TransUnion Rental Screening Solutions (TURSS) has systematically corrupted this data point by populating it not with the final amount awarded by a judge, with the initial amount demanded by the landlord. This practice converts unproven allegations into reported financial facts, branding tenants with debts they do not owe and may have successfully defeated in court.

The distinction between a claim and a judgment is the foundation of civil law. When a landlord files an eviction suit, they include an ad damnum clause, a specific request for damages. This figure frequently includes back rent, future rent, late fees, legal costs, and sometimes arbitrary punitive damages. It is a wish list, not a verdict. In jurisdictions, landlords this number to ensure they capture all chance costs, knowing the court likely reduce it. A tenant might be sued for $5, 000, after presenting evidence of uninhabitable conditions or accounting errors, be ordered to pay only $500, or nothing at all. In cases where the tenant wins or the case is dismissed, the legal debt is zero.

TransUnion’s automated scraping systems have frequently ignored this judicial outcome. Instead of parsing the final disposition to find the actual money judgment, the algorithms grab the dollar figure available in the court docket: the initial claim. This raw data is then fed into the “Judgment Amount” field of the SmartMove or ResidentScreening report. The result is a report that tells a future landlord the applicant has a “Civil Judgment” for $5, 000, even if the case was dismissed three years ago. This is not a clerical error; it is the fabrication of a debtor profile that exists only in the database of the screening company, not in the records of the court.

The FTC and CFPB Findings

This specific deception was a central pillar of the October 2023 joint enforcement action by the Federal Trade Commission (FTC) and the Consumer Financial Protection Bureau (CFPB). The regulators alleged that TURSS failed to follow reasonable procedures to assure maximum possible accuracy, a direct violation of the Fair Credit Reporting Act (FCRA). The complaint detailed how TransUnion’s systems labeled money a landlord claimed a consumer owed as a “Judgment Amount.” This labeling created the false impression that a judge had reviewed the evidence and ordered the tenant to pay that specific sum.

The regulators found that this practice was not an glitch a structural defect in how TransUnion acquired and processed public record data. By relying on third-party vendors and automated scrapers that prioritized speed and volume over granular accuracy, TransUnion allowed the landlord’s opening argument to become the final word in the tenant’s permanent record. The settlement required TransUnion to pay $15 million specifically for these tenant screening violations, sending a clear signal that conflating claims with judgments is an illegal practice. Yet, for the tenants affected before this intervention, the damage was already done.

The method of the Error

To understand why this error, one must examine the architecture of court record scraping. Court dockets are frequently messy, unstructured text files. A human clerk reading a docket can easily distinguish between “Plaintiff seeks $2, 500” and “Judgment entered for Defendant.” Automated scrapers, yet, look for patterns. They scan for the dollar sign ($) and keywords like “rent,” “damages,” or “suit amount.” The initial filing almost always contains these clear, scrape-friendly figures. The final disposition, by contrast, might be a handwritten note, a checkbox on a form, or a complex order that requires context to interpret.

TransUnion’s vendors, incentivized to provide data that looks complete, frequently default to the most easily accessible number. If the final judgment amount is missing, ambiguous, or difficult to parse, the system reverts to the filing amount. This creates a “better safe than sorry” method for the landlord-customer, warning them of the worst-case scenario, it is catastrophic for the tenant. A tenant who settles a case by agreeing to move out in exchange for a waiver of all back rent, a common “cash for keys” or “mutual release” agreement, technically owes zero dollars. Yet, because the initial suit demanded thousands, the screening report continues to display that demand as an active debt.

The ‘Civil Judgment’ Misnomer

The presentation of this data exacerbates the harm. TransUnion reports categorize these entries under headings like “Civil Judgments” or “Monetary Judgments.” These terms have specific legal meanings. A civil judgment is a court order that permits a creditor to use state power, such as wage garnishment or bank levies, to collect a debt. By placing the landlord’s initial claim under this heading, TransUnion implies that the debt has reached this stage of enforceability.

Property managers, who frequently review these reports in seconds, rely on these headings to make quick decisions. They do not have the time or legal training to pull the original court file and verify if the “Judgment Amount” matches the actual ruling. They see “$3, 200 Judgment” and move to the applicant. The tenant is rejected not because they are a risk, because the screening report lied about the legal status of the dispute. The tenant is punished for being sued, regardless of the outcome of that suit.

The Verification Trap

When a tenant discovers this error and attempts to dispute it, they frequently fall into a verification trap. The tenant files a dispute stating, “I do not owe $4, 000; the case was dismissed.” TransUnion’s dispute resolution team then contacts the vendor or checks the court record. They see that an eviction case was filed and that the complaint did demand $4, 000.

Because the dispute investigators frequently verify the existence of the document rather than the nuance of the legal outcome, they may confirm the entry as “accurate.” The logic is circular: the report says the landlord demanded $4, 000, the court file says the landlord demanded $4, 000, therefore the report is accurate. This ignores the core complaint, that the report labels this demand as a “Judgment.” This bureaucratic sleight of hand allows the error to survive multiple rounds of disputes, leaving the tenant trapped by a debt that exists only on paper.

Financial Risk Profiling

The inclusion of these false judgment amounts distorts the financial risk profile of the tenant. landlords use debt-to-income ratios or specific thresholds for outstanding debt to filter applicants. A phantom $5, 000 judgment can push a tenant above these thresholds, triggering an automatic denial from the leasing software before a human even reviews the file.

Also, this practice disproportionately harms tenants who fight back against unjust evictions. A tenant who withholds rent due to a absence of heat or water is frequently sued for that rent. If the court rules in the tenant’s favor and abates the rent, the tenant has been vindicated. if the screening report still lists the withheld rent as a “Judgment Amount,” the tenant is penalized for exercising their legal right to safe housing. The report enforces the landlord’s initial retaliation, extending the punishment long after the court has declared the tenant innocent.

Regulatory Failure and Industry Standards

The persistence of this problem reveals a failure in industry standards regarding data dictionary definitions. There is no universal standard for what constitutes a “judgment” in tenant screening data feeds. While the credit reporting side of the business has strict codes (Metro 2 format) for reporting debts, the eviction screening side operates in a “wild west” of unstructured data.

TransUnion, as a major player, had the resources to impose strict data standards on its vendors. It could have required a separate field for “Claim Amount” and “Judgment Amount,” forcing vendors to distinguish between the two. It could have programmed its logic to display “$0” or “N/A” in the judgment field unless a specific monetary ruling was found. Instead, it chose a data architecture that conflated the two, prioritizing the presence of any dollar figure over the accuracy of the dollar figure.

The Human Cost

The human cost of this data negligence is measured in homelessness and instability. A tenant with a false judgment on their record is frequently locked out of institutional housing. They are forced into the private market, renting from slumlords who do not check credit charge predatory rents for substandard conditions. The false debt follows them, appearing on every subsequent application.

In one documented instance, a tenant who had successfully defended against an eviction claim involving a disputed security deposit found herself unable to rent an apartment for two years. Every application came back denied due to a “monetary judgment” that the court had explicitly denied. She carried the actual court order with her to viewings, trying to explain the error to leasing agents, the “Judgment Amount” on the TransUnion screen carried more authority than the judge’s signature in her hand.

This practice converts the civil court system, designed to resolve disputes, into a permanent blacklisting engine. By reporting claims as debts, TransUnion allows landlords to inflict financial damage on tenants simply by filing a lawsuit, regardless of the merit of that suit. The court’s dismissal becomes irrelevant; the filing itself, and the dollar sign attached to it, becomes the verdict.

The ‘Black Box’ Risk Score: Algorithmic Bias Based on Flawed Eviction Data

The modern tenant screening industry has evolved beyond simple background checks into a complex ecosystem of predictive analytics. At the forefront of this shift is TransUnion’s proprietary scoring model, known as **ResidentScore**. Marketed to landlords as a sophisticated tool to predict eviction risk, this algorithm reduces a human life to a three-digit number between 350 and 850. While TransUnion claims this score predicts eviction outcomes 15% more than traditional credit scores, the method behind it remains a “black box”, unclear, proprietary, and largely immune to consumer scrutiny. The fundamental danger of ResidentScore and similar algorithmic tools lies in the quality of the data they consume. In data science, the principle is absolute: “garbage in, garbage out.” When TransUnion’s automated systems scrape court databases for eviction filings, they frequently ingest records that are sealed, dismissed, or resolved in favor of the tenant. As established in previous sections, these records frequently absence serious context or updates. When this flawed data is fed into a risk-scoring algorithm, the result is a mathematically generated bias that permanently brands innocent renters as “high risk.”

The method of Algorithmic Unfairness

Unlike a credit score, which is regulated by strict federal standards and based on defined financial behaviors, a tenant risk score is a murky amalgamation of credit history, criminal records, and eviction filings. TransUnion’s marketing materials boast that ResidentScore analyzes “complex combinations of data” to identify tenants who might “skip” or face eviction. yet, the exact weighting of these factors is a trade secret. For a prospective tenant, this opacity is disastrous. A renter might review their credit report and see a decent score, only to be rejected by a landlord who relies solely on the ResidentScore. Because the algorithm penalizes applicants for the mere presence of an eviction filing, regardless of the outcome, a tenant who successfully defeated an unjust eviction in court may still receive a failing grade. The algorithm does not distinguish between a tenant who destroyed property and refused to pay rent, and a tenant who withheld rent legally due to uninhabitable conditions and won their case. To the code, both are simply “eviction records,” and both lower the score. This automated adjudication removes human judgment from the leasing process. Landlords, frequently encouraged by TransUnion’s marketing to “minimize risk,” treat the ResidentScore as a pass/fail threshold. If an applicant scores a 550, they are denied automatically. The landlord never looks at the underlying documents that might show the eviction was dismissed or that the “criminal record” was a case of mistaken identity. The computer says no, and the discussion ends.

Impact and Digital Redlining

The reliance on eviction filings as a primary variable in risk scoring creates a severe impact on minority communities, a phenomenon housing advocates describe as “digital redlining.” Statistical data confirms that Black and Hispanic renters face eviction filings at significantly higher rates than white renters, even when controlling for income. In jurisdictions, landlords use eviction filings as a routine rent collection tactic, filing cases that are subsequently dropped once payment is made. When TransUnion’s algorithm treats these filings as negative risk factors, it criminalizes poverty and race. A Black woman in a low-income neighborhood is statistically more likely to have a “serial filing” history, where a landlord files multiple times in a year without ever executing an eviction, than a white renter in a different zip code. ResidentScore interprets this pattern not as evidence of an aggressive landlord, as evidence of a “high-risk” tenant. The Consumer Financial Protection Bureau (CFPB) and the Federal Trade Commission (FTC) have explicitly warned against this practice. In their joint enforcement actions, regulators have highlighted that automated scoring systems can reproduce and amplify existing inequalities in the housing market. By baking historical discrimination into a proprietary score, companies like TransUnion sanitize bias, giving it a veneer of mathematical objectivity. A landlord can claim they aren’t discriminating based on race; they are simply following the “impartial” recommendation of a third-party algorithm.

The Persistence of “Zombie” Data in Risk Scores

One of the most insidious aspects of the ResidentScore is its memory. Even if a consumer successfully disputes an inaccurate eviction record on their screening report, there is no guarantee that the risk score immediately recalibrate. The “black box” nature of the algorithm means that the data points used to calculate the score at a specific moment are frozen in time. also, because the score is a derivative product, disputing it is nearly impossible. A consumer can dispute a specific line item, “This eviction judgment from 2019 is false”, they cannot dispute the score itself. If TransUnion removes the false record the algorithm has already tagged the consumer as “high risk” based on other correlated data points (such as the drop in credit score that frequently accompanies the financial of an eviction battle), the rental denial stands. The 2023 settlement between TransUnion, the FTC, and the CFPB underscored this failure. The regulators charged that TransUnion failed to ensure the maximum possible accuracy of the information it provided. This inaccuracy is not just a clerical error; it is the fuel for the risk score engine. When the engine runs on data that includes sealed records, files that are legally required to be invisible, it violates the core intent of consumer protection laws. The algorithm “unseals” these records by using them to calculate a lower score, punishing the tenant for a past they were legally absolved of.

Regulatory Failure and the absence of Accountability

even with the $15 million settlement and the $8 million in restitution ordered in October 2023, the structural problems with tenant risk scores. The settlement required TransUnion to improve its procedures regarding eviction data accuracy, it did not ban the use of algorithmic scoring itself. The industry continues to operate with a “catch me if ” mentality, prioritizing speed and volume over precision. The load of proof remains entirely on the renter. To correct a suppressed score, a tenant must realize they were denied because of it, a fact frequently obscured by generic adverse action notices. Then, they must obtain their report, identify the inaccurate data, file a dispute, wait for the investigation, and hope the score updates in time to secure the apartment. In tight rental markets, this delay is fatal. By the time the “black box” is forced to correct its math, the apartment is gone. This system creates a pattern of housing instability. A low ResidentScore leads to denials, forcing tenants into substandard housing with predatory landlords who are less likely to screen more likely to file evictions. These new filings further depress the score, trapping the tenant in a permanent underclass. TransUnion’s algorithm does not just predict risk;, it actively manufactures it by locking stable tenants out of the housing market based on flawed, outdated, and biased data.

The “Black Box” of ResidentScore represents the industrialization of housing bias. By automating the rejection process and insulating landlords from the human reality of their applicants, TransUnion has built a product that is for property owners devastating for renters. The algorithm does not care if an eviction was illegal, if a record is sealed, or if a tenant is a victim of identity theft. It sees only data, and in the world of tenant screening, that data is frequently wrong.

The Statutory Mandate: Maximum Possible Accuracy

Under the Fair Credit Reporting Act (FCRA), specifically 15 U. S. C. § 1681e(b), consumer reporting agencies are bound by a rigorous standard: they must follow reasonable procedures to assure “maximum possible accuracy” of the information concerning the individual about whom the report relates. This legal requirement is not a suggestion for “good enough” data nor a permission slip for approximate matching. It demands that companies like TransUnion Rental Screening Solutions (TURSS) use methods that actively prevent errors, rather than correcting them after a consumer’s life has been upended. Yet, the operational reality of TransUnion’s tenant screening business reveals a widespread disregard for this mandate, particularly regarding the handling of sealed, dismissed, or resolved eviction records.

The core violation lies in the between the nature of court records and the static nature of TransUnion’s databases. Court dockets are living documents; a case filed on Monday can be dismissed on Wednesday, sealed on Friday, or resolved through a settlement the following week. TransUnion, yet, frequently relies on a “snapshot” method of data collection. Third-party vendors scrape court filings at the moment of initiation, capturing the raw allegation of an eviction. Once this data enters TransUnion’s ecosystem, it frequently calcifies into a permanent mark. The company’s automated systems absence a consistent, real-time feedback loop to verify the current status of the case before selling the report to a landlord. Consequently, a tenant who successfully defeated an eviction attempt in court frequently finds the initial filing, legally irrelevant or sealed, staring back at them from a screening report, blocking their access to housing.

The “Snapshot” Failure: How Sealed Records Remain Visible

The persistence of sealed records in TransUnion reports constitutes a direct failure of the “reasonable procedures” standard. When a court orders a record sealed, it is a legal determination that the information should no longer be public or used against the individual. In jurisdictions, this sealing happens automatically if a tenant wins or if the case is withdrawn. yet, because TransUnion’s data acquisition model prioritizes speed and volume over currency, their databases frequently retain the record as it existed *before* the sealing order.

Federal investigations have exposed that TransUnion did not have adequate to check for these updates. In the October 2023 joint enforcement action by the Federal Trade Commission (FTC) and the Consumer Financial Protection Bureau (CFPB), regulators charged that TURSS failed to prevent the inclusion of sealed eviction records. The agencies found that TransUnion continued to report these prohibited items because they did not verify the public record status at the time the report was generated. Instead, they regurgitated old data harvested months or years prior. This practice shifts the load of accuracy entirely onto the consumer, who must discover the violation through a rejected application and then navigate a labyrinthine dispute process to force TransUnion to recognize a court order that already exists.

The Persistence of Resolved and Dismissed Cases

Beyond sealed records, the failure to update the *disposition* of eviction cases represents a massive structural flaw. A “filing” is an accusation; a “judgment” is a legal finding. In the world of tenant screening, yet, TransUnion has frequently blurred this distinction. The FTC complaint detailed how TURSS reported eviction proceedings without including the final outcome, such as a dismissal. A landlord reviewing such a report sees only that an eviction was filed, leading to the assumption that the tenant was a problem, even if the case was thrown out of court for absence of merit or if the landlord admitted to an error.

The method of this failure is rooted in cost-saving automation. verifying the final disposition of a case requires a second, frequently manual, check of the court docket, or a more sophisticated data pipeline that continuously queries for updates. TransUnion’s systems, yet, were designed to ingest data struggled to digest corrections. The 2023 settlement highlighted that until April 2021, TURSS frequently reported multiple developments in the same eviction proceeding as separate events. This “multiple entry” error not only inflated the perceived risk of the tenant also demonstrated a fundamental inability to track the lifecycle of a single legal case. By treating a dismissal or a procedural update as just another data point rather than a modification of the original record, TransUnion’s reports painted a picture of chronic instability where there was only a single, resolved legal dispute.

The 2023 Federal Enforcement Action

The magnitude of these failures was quantified in the landmark October 2023 settlement, where TransUnion and its subsidiary agreed to pay $15 million to settle charges related specifically to tenant screening inaccuracies. This was the largest amount ever recovered by the FTC in a tenant screening matter. The complaint filed by the FTC and CFPB was explicit: TransUnion violated the FCRA by failing to follow reasonable procedures to assure maximum possible accuracy. The regulators noted that the company did not accurately report the disposition of eviction cases, labeled monetary amounts incorrectly, and failed to prevent the inclusion of sealed records.

The settlement order required TransUnion to overhaul its procedures. It mandated that the company essentially stop reporting eviction records unless it could verify the current status and accuracy of the information. This legal intervention show that TransUnion’s prior methods were not accidental glitches a business choice to underinvest in accuracy. The company had the resources to build a system that checked for dismissals and sealing orders; they simply chose a cheaper model that relied on stale data until federal regulators forced their hand.

Misleading “Judgment” Labels

Another dimension of this accuracy failure involves the labeling of financial data. The FTC investigation revealed that TURSS frequently labeled money that a landlord *claimed* a consumer owed as a “Judgment Amount.” In legal terms, a claim is an unproven demand, while a judgment is a court-ordered debt. By conflating the two, TransUnion gave landlords the false impression that a court had already reviewed the evidence and ruled against the tenant.

This mislabeling is a textbook violation of Section 1681e(b). A reasonable procedure would distinguish between a plaintiff’s prayer for relief and a judge’s final order. TransUnion’s system, yet, frequently scraped the dollar figure from the initial complaint and presented it in a column that implied a finalized debt. For a prospective tenant, this error is catastrophic. A landlord sees a “Judgment Amount” of $5, 000 and assumes the applicant is a financial risk who stiffed a previous owner. In reality, the case might have been settled for zero dollars, or the tenant might have won. By presenting allegations as adjudicated facts, TransUnion’s reports actively deceived their customers and defamed the subjects of the reports.

The Human Cost of “Reasonable” Procedures

TransUnion has historically defended its practices by arguing that its procedures are “reasonable” given the volume of data it handles. They contend that no system can be 100% accurate. yet, the courts and regulators are increasingly rejecting this defense when the errors are widespread and preventable. The “maximum possible accuracy” standard requires more than just a high batting average; it requires specific to address known risks. TransUnion knew that eviction cases are frequently dismissed or sealed. They knew that their data sources were frequently static snapshots. By failing to implement a resynchronization process to catch these updates, they knowingly allowed inaccuracies to fester.

The result of this negligence is a “conditional” accuracy that works only for the data broker, not the consumer. A report that is 90% accurate regarding a consumer’s name 100% wrong regarding their eviction history is, for the purpose of housing, a total failure. The FCRA was enacted to prevent exactly this scenario, where a corporate entity controls the fate of an individual using flawed data that the individual cannot easily see or correct in time. TransUnion’s failure to update sealed and resolved records is not a technical oversight; it is a violation of the federal right to a fair and accurate credit profile.

The Phantom Source: How TransUnion Obscured Data Origins

For years, renters attempting to dispute inaccurate eviction or criminal records on their TransUnion screening reports encountered a deliberate procedural dead end. When a consumer requested the source of a derogatory mark, such as a sealed eviction or a dismissed criminal charge, TransUnion Rental Screening Solutions (TURSS) frequently provided a misleading answer. Instead of identifying the actual third-party data broker that supplied the erroneous file, TransUnion’s disclosures implied the information came directly from the “jurisdiction” or the court system itself. This obfuscation created a bureaucratic labyrinth. A renter might contact the specific county court listed on their report, only to be told by court clerks that the record did not exist, was sealed, or had been expunged. Armed with this proof from the court, the renter would return to TransUnion, only to find the dispute process stalled because the *actual* supplier of the data, a private vendor, remained hidden. The consumer could not correct the error at the source because TransUnion refused to name the source.

The Vendor Shell Game

Behind the veil of “public record” data, TransUnion relied on a complex network of third-party vendors to scrape, aggregate, and resell court data. These vendors, including companies like **LexisNexis Risk & Information Analytics Group**, **SJV & Associates**, **Tessera Data**, and **One Source Technology (dba Asurint)**, acted as the true furnishers of the information. By withholding these identities, TransUnion insulated its data supply chain from consumer scrutiny. If a background check vendor like Tessera Data scraped a court website in 2018 and captured an eviction filing that was dismissed two weeks later, that “ghost record” would in their database. When TransUnion purchased that data point years later, they reported it as fact. When the consumer disputed it, TransUnion pointed the finger at the court, not Tessera. The consumer was left chasing a ghost in the court system while the actual error resided on a private server they had no access to. This practice did not delay justice; it denied it. The Federal Trade Commission (FTC) and Consumer Financial Protection Bureau (CFPB) investigations revealed that this was a widespread feature of TransUnion’s business model, not an accidental oversight. The agencies found that TransUnion “failed in numerous instances” to identify these vendors in file disclosures, a direct violation of the Fair Credit Reporting Act (FCRA).

The “SmartMove” Blind Spot

The roadblock was particularly damaging for users of TransUnion’s **SmartMove** platform, a service marketed to smaller, independent landlords. SmartMove promotes a “push” model where renters initiate the screening request. yet, when these reports contained errors, the dispute method frequently routed consumers into a loop of automated responses. Because the data sources were masked, renters could not exercise their right to demand a reinvestigation from the specific furnisher of the data. Under the FCRA, a consumer reporting agency must provide the “sources of the information” upon request. TransUnion’s interpretation of “source” as the *originating court* rather than the *data vendor* deprived consumers of the ability to sue or demand corrections from the entity actually responsible for the libelous data.

Regulatory Crackdown on “Secret Sources”

The illegality of this practice was a central pillar of the October 2023 joint enforcement action by the FTC and CFPB. Samuel Levine, Director of the FTC’s Bureau of Consumer Protection, explicitly condemned the use of “secret sources,” stating that consumers “shouldn’t be shut out by tenant screening reports that are ridden with errors and based on data from secret sources.” As part of the $23 million settlement, federal regulators ordered TransUnion to overhaul its disclosure practices. The company is legally mandated to provide the full names and contact information of any third-party vendor that supplies criminal or eviction records. This requirement forces the “black box” of tenant screening to open, allowing consumers to see exactly whose algorithm or scraping bot tagged them as a risk.

The Cost of Obscurity

The consequences of this years-long practice were severe. Renters who could have easily fixed a database error with a direct phone call to a vendor like SJV & Associates were instead forced into weeks of homelessness or substandard housing while they fought a phantom battle with court clerks who had no power to change TransUnion’s internal files. By shielding its vendors, TransUnion prioritized the stability of its data partnerships over the accuracy of its reports. It treated the identity of its suppliers as a trade secret worth protecting, even at the cost of accurate reporting for the families whose housing prospects depended on that data. The “roadblock” was not a bug; it was a barrier designed to keep the of automated screening running without the friction of individual accountability.

Beard v. TURSS: Challenging Overly Broad Criteria and Mismatched Tenant Data

The class action lawsuit *Beard v. TransUnion Rental Screening Solutions, Inc.* (TURSS) stands as a definitive legal challenge to the automated, high-volume data matching practices that define the modern tenant screening industry. Filed in the United States District Court for the Western District of Virginia, the case exposed the catastrophic failure of TransUnion’s proprietary algorithms to distinguish between individuals with superficially similar names vastly different identities. The central allegation focused on TURSS’s use of “overly broad criteria”, a euphemism for loose matching logic that prioritizes the generation of “hits” over the verification of facts. This practice resulted in innocent renters being branded as criminals or sex offenders, barring them from housing opportunities based on data that did not belong to them.

The Case of Richard Adam Beard

Richard Adam Beard, the lead plaintiff, applied for a rental property and was subjected to a standard background check processed by TURSS. The resulting report delivered a shock to the chance landlord: it flagged Beard as a registered sex offender. The entry a conviction for “attempted use of communications systems to contact a minor child.” The between the plaintiff and the criminal record was absolute. The actual offender listed in the public registry was named “Gary,” not Richard. The offender was born in 1953, whereas Beard was born in 1977, a twenty-four-year age difference. also, the offender’s middle name did not match Beard’s. even with these disqualifiers, TransUnion’s screening system linked the sex offender registry entry to Beard’s consumer file. The landlord, receiving a report stamped with the authority of a major credit bureau, denied Beard’s application. This error was not a random glitch the predictable result of a specific data processing choice. The complaint alleged that TURSS used matching criteria so loose that it allowed criminal records to be attributed to consumers solely based on partial name matches, ignoring serious differentiators like middle names or dates of birth. In the industry, this is frequently referred to as “wildcard” matching, a method designed to catch individuals who might use variations of their name to hide a criminal past. When applied without rigorous secondary filters, this method ensnares innocent people who share a surname or a common name with a convict.

The Mechanics of “Overly Broad Criteria”

The core legal problem in *Beard* was the violation of the Fair Credit Reporting Act (FCRA), specifically Section 1681e(b), which mandates that consumer reporting agencies follow “reasonable procedures to assure maximum possible accuracy” of the information concerning the individual about whom the report relates. TransUnion’s defense of its matching logic frequently relies on the difficulty of obtaining complete identifiers from public record sources. Court records in jurisdictions may redact dates of birth or Social Security numbers to protect privacy. Faced with incomplete data, a screening company has two choices: exclude the record because it cannot be definitively linked to the applicant, or include the record as a “possible match” and shift the load of verification to the landlord or the tenant. The *Beard* complaint argued that TURSS systematically chose the latter. By setting its algorithms to accept partial matches, such as matching a last name and a partial name, or disregarding a mismatched date of birth, TURSS maximized the number of “hits” it could sell to landlords. A report that returns “no record found” is accurate less valuable to a risk-averse property manager than a report that flags a chance threat. Consequently, the financial incentive structures of the screening industry encourage over-reporting. In Beard’s case, the algorithm likely identified a match on the surname “Beard” and perhaps a loose geographic indicator, then disregarded the non-matching name “Gary” and the non-matching birth year. A human reviewer looking at the two files side-by-side would instantly reject the match. The automated system, operating without human intervention, processed the data as a valid hit. This automation allows TransUnion to generate millions of reports annually at low cost, it introduces a high error rate that disproportionately affects consumers with common names.

The Class Action Allegations

The lawsuit sought to represent not just Richard Beard, a class of similarly situated individuals. The “Policy Settlement Class” included consumers for whom TURSS reported a criminal or landlord-tenant record that did not belong to them, resulting from the company’s failure to use strict matching criteria. Attorneys for the plaintiff argued that TransUnion’s failure was willful. The company had been sued numerous times for similar FCRA violations and was well aware of the risks associated with loose matching logic. even with this knowledge, TURSS continued to use algorithms that did not require a full match of Personally Identifiable Information (PII). The complaint detailed that a reasonable procedure would require, at a minimum, a match of the consumer’s name, middle name, last name, and date of birth before attributing a serious criminal record to their file. The of this practice are severe. When a tenant screening report falsely labels an applicant as a sex offender, the damage is immediate and frequently irreversible. The applicant is denied housing, frequently without being told the specific reason, or is told the denial is due to “criminal history.” Even if the consumer disputes the report and corrects the error, a process that can take weeks, the apartment is rented to someone else by the time the record is cleared. The “maximum possible accuracy” standard exists precisely to prevent this type of irreparable harm.

The $11. 5 Million Settlement and Policy Reforms

In 2023, TransUnion agreed to settle the *Beard* class action for $11. 5 million. While the monetary relief provided compensation to consumers who had been harmed, the most significant outcome of the litigation was the “Policy Settlement”, a binding agreement requiring TURSS to overhaul its matching procedures. Under the terms of the settlement, TransUnion Rental Screening Solutions agreed to implement stricter matching logic for criminal and landlord-tenant records. The new prohibit the company from reporting a criminal record unless it can match the record to the consumer using more specific criteria. This involves requiring a match on the consumer’s full name and date of birth, or a match on the name plus a Social Security number or address history. also, the settlement addressed the problem of “stale” data. The plaintiffs alleged that TURSS scraped data from public websites and stored it in internal databases without regular updates. This meant that if a criminal charge was dropped or an eviction sealed *after* TransUnion had scraped the initial filing, the screening report would still show the original negative record. The settlement mandated that TURSS must refresh its data more frequently to ensure that the disposition of cases (e. g., dismissed, acquitted, sealed) is accurately reflected in the reports.

The Persistence of “Mixed Files”

The *Beard* case highlights a persistent problem in the credit reporting industry known as the “mixed file.” A mixed file occurs when data from Consumer A is commingled with the file of Consumer B. In the context of credit reports, this frequently involves financial accounts. In tenant screening, it involves criminal records and eviction filings. For renters, a mixed file is far more dangerous than a credit error. A drop in credit score might result in a higher interest rate; a false sex offender flag results in homelessness. The *Beard* litigation demonstrated that these mixed files are not accidental anomalies the statistical inevitability of “probabilistic matching.” Probabilistic matching assigns a likelihood score to a chance match. If the score crosses a certain threshold, the system links the records. By setting the threshold too low, TURSS increased its “recall” (finding all chance records) at the expense of “precision” (ensuring the records actually belong to the subject). The *Beard* settlement forces a recalibration of this balance, pushing the threshold higher to favor precision.

Broader for Tenant Screening

The victory in *Beard v. TURSS* serves as a warning to the broader background check industry. It establishes that “possible” matches are insufficient under the FCRA. A screening company cannot simply dump raw, unverified court data into a report and disclaim responsibility for its accuracy. The law requires them to act as a filter, not just a funnel. yet, the settlement also reveals the limitations of private litigation in fixing widespread problem. While TransUnion agreed to change its practices, the underlying business model—selling inexpensive, instant data derived from public records—remains intact. The challenge of distinguishing between two people with the same name in a country of 330 million people is a data science problem that requires expensive solutions, such as manual verification by human investigators. For the millions of renters subject to automated screening, *Beard* provides a legal precedent that inaccurate matching is actionable. It affirms that a person’s identity cannot be reduced to a last name and a general location. The distinction between “Gary” born in 1953 and “Richard” born in 1977 is not a minor detail; it is the difference between a successful rental application and a wrongful denial. By forcing TransUnion to acknowledge this distinction, the *Beard* case imposed a necessary check on the algorithmic overreach that threatens the housing security of innocent tenants. The settlement funds were distributed to eligible class members, categorized into groups based on the type of error they suffered—such as the “Age Mismatch Group” for those like Beard who were linked to records with different dates of birth. This categorization itself serves as an admission of the specific ways the algorithms failed: they ignored age, they ignored middle names, and they ignored the obvious reality that two different people can share a surname. Moving forward, the established by *Beard* require TURSS to maintain an audit trail of its matching logic. If a consumer disputes a record, the company must be able to demonstrate *why* the system linked the record to the file. This transparency is a serious departure from the “black box” operations of the past, where proprietary algorithms were shielded from scrutiny. The case proves that when the “black box” is forced open, the inside frequently reveals not sophisticated intelligence, crude and negligent pattern matching.

The “Maximum Possible Accuracy” Mandate

The core of the settlement rests on the enforcement of Section 607(b) of the Fair Credit Reporting Act (FCRA). For decades, TransUnion argued that “accuracy” meant simply reporting what was on the court docket, regardless of whether that docket was outdated or misleading. The 2023 order rejected this passive definition. TransUnion is legally required to implement “reasonable procedures to assure maximum possible accuracy” of eviction information. This shifts the load of verification onto the bureau. They can no longer blame the court clerks or third-party vendors for data decay. If a record is reported, TransUnion must possess the technical capability to verify its current status, ensuring that the data reflects the final disposition of the case rather than the initial filing.

Prohibition on Reporting Sealed and Resolved Records

One of the most aggressive components of the order is the explicit ban on reporting sealed records. Previously, TURSS algorithms would scrape court data at the moment of filing. If a judge later sealed that eviction file, frequently the result of a settlement where the tenant paid what was owed, TransUnion’s database frequently failed to overwrite the original entry. The sealed record would from public view remain visible in the TURSS report sold to landlords. The consent order mandates that TransUnion design specific procedures to prevent this. They must establish a synchronization cadence that catches expungements and sealing orders. This requirement forces a technological overhaul. TransUnion must query court databases not just for new hits, for status updates on existing hits. The “snapshot” method of data collection, where a record is grabbed once and stored forever, is a violation of federal law.

Eliminating the “Multiple Entry”

The “multiple entry” glitch was a primary focus of the FTC’s investigation. This error occurred when TransUnion’s automated systems treated every procedural update in a single eviction case as a new, separate eviction event. A tenant fighting one eviction case could end up with four or five “records” on their report, signaling to landlords that they were a serial offender. Under the new compliance regime, TransUnion must employ logic that collapses these multiple data points into a single case file. The order requires the bureau to match case numbers, dates, and parties to ensure that a single legal proceeding is reported as one event. This mandate directly attacks the “quantity over quality” method that previously allowed TURSS to risk scores based on redundant data.

Ban on Misleading “Judgment” Amounts

Perhaps the most financially damaging practice for tenants was TransUnion’s reporting of “claim amounts” as “judgment amounts.” When a landlord files for eviction, they frequently claim an arbitrary amount of money is owed. The court may later rule that the tenant owes nothing, or significantly less. TURSS frequently reported the initial claim amount in the “judgment” field, falsely inflating the tenant’s debt profile. The enforcement order imposes a strict prohibition on this practice. TransUnion is forbidden from including monetary amounts other than final judgments rendered by a court. They cannot report the landlord’s wish list as a legal debt. This change prevents the weaponization of the screening report, where a landlord’s initial, unverified allegation could permanently damage a tenant’s financial reputation even if the court later rejected the claim.

the Third-Party Vendors

For years, consumers attempting to dispute an eviction record were met with a bureaucratic dead end. TransUnion would claim the data came from a “public record source” would refuse to identify the specific vendor, frequently LexisNexis or CoreLogic, that actually supplied the file. This prevented tenants from fixing the error at the source. If they disputed with TransUnion, the bureau would simply re-verify with the same vendor, creating a closed loop of inaccuracy. The post-settlement compliance standards obliterate this secrecy. TransUnion must disclose the identity of any third-party vendor that provided the criminal or eviction record. The “file disclosure” provided to consumers must list the name and contact information of the data supplier. This transparency allows tenants to bypass TransUnion’s internal delays and challenge the accuracy of the data directly with the aggregator that scraped it. It forces TransUnion to acknowledge that it is part of a supply chain, not an entity.

The “Adverse Action” Notification Requirement

The settlement also addresses the landlord’s role in the denial process. frequently, landlords would deny a tenant based on a TURSS report fail to provide the required adverse action notice, leaving the tenant guessing why they were rejected. The consent order requires TransUnion to make a sample adverse action notice available on its website. This notice must prompt the landlord to share the tenant screening report with the applicant and explain the reason for the denial. While TransUnion cannot force landlords to use the letter, the requirement creates a paper trail of compliance. It shifts the industry standard toward transparency, making it easier for regulators to identify when landlords, and by extension, TransUnion, are failing to inform consumers of their rights.

The Compliance Monitor and Future Penalties

To assure these reforms are not temporary, the order subjects TransUnion to continued oversight. The $15 million penalty serves as a baseline; future violations of this consent order would trigger significantly higher fines and chance criminal contempt charges. The CFPB has signaled that it is monitoring TransUnion’s dispute resolution logs to verify that the “maximum possible accuracy” standard is being met in practice, not just in policy documents. The operational cost of this compliance is substantial. TransUnion must spend resources verifying data that it previously sold with zero overhead. The “churn and burn” model of tenant screening, where speed was prioritized over truth, is no longer legally viable. The bureau must function as a steward of data accuracy, a role it resisted until the federal government forced its hand.

Conclusion of the Review

The investigation into TransUnion’s tenant screening practices reveals a corporate entity that systematically prioritized algorithmic efficiency over human reality. For decades, the bureau allowed sealed, dismissed, and inaccurate eviction records to destroy the housing prospects of millions of Americans. The “multiple entry” glitches, the reporting of false judgment amounts, and the refusal to disclose data sources were not accidental bugs; they were features of a low-cost, high-volume business model. The 2023 settlement forces a correction, yet the damage to historical victims remains largely unaddressed. While the mandated reforms pledge a more accurate future, the method of profit-driven tenant screening remains inherently risky. As long as housing access depends on automated scraping of fragmented court databases, the chance for error. TransUnion is under a federal microscope, compelled to prove that it can distinguish between a filing and a fact. The era of “blind” reporting is over; the era of accountability has just begun.