The nation’s system of unemployment insurance (UI)—a vital lifeline for workers as well as local and regional economies during an economic downturn—faced unprecedented levels of benefit fraud committed by criminal enterprises during the COVID-19 pandemic. While state efforts to eliminate fraud through programs such as ID verification (IDV) have made tremendous strides, states will need to continue to improve their IDV protocols, as criminal enterprises will continue to attack the UI system relentlessly. But as of yet, there has been scant research or consensus on which approaches have worked best, and which have been easiest to implement. Furthermore, states often struggle to maintain strong lines of communication with each other regarding combating fraud or even establishing a common terminology about the challenges they are facing. This report—the product of interviews with a range of UI stakeholders across the country—will review current state IDV processes and practices as well as emerging trends in combating fraud; it will then detail what is needed to ensure that all states can adequately protect the UI system from current and future threats.

Why ID Verification (IDV) Still Matters

At the height of the COVID-19 pandemic, criminal enterprises found their way into pandemic unemployment benefits, which at the time had greatly enhanced payments but few requirements regarding identity verification. Congress had added a requirement to the Continued Assistance Act of 2020 to require states to “include procedures for identity verification or validation and for the timely payment, to the extent reasonable and practicable, by January 26, 2021,”1 but it took states months to use the allocated federal funding to shore up their ID verification efforts. At the end of the enhanced pandemic programs, state ID verification efforts were achieving results, but the criminals have not given up. Fraud is now a persistent and ongoing threat2 to the unemployment insurance (UI) system.

One of the most critical protections against fraud continues to be verifying that claimants are real people and who they say they are. Because of the robust work states already have done, ID fraud has fallen to nearly zero in almost every state.3 Fraudsters continue to hone their tactics, so states have to respond in kind by improving their practices. For example, as artificial intelligence (AI) becomes more sophisticated, fraudsters can use this technology to make more convincing claims, invent more believable fake employers, and hijack legitimate claims more efficiently. The U.S. Department of Labor continues to communicate to states that it expects them to conduct robust, universal, and frequent ID verification.

This current state stance on identity verification is dramatically different from the one prior to the pandemic. Before 2020, states would accept a claim application once it was submitted. In fact, most states would tend to encourage individuals to file a claim as soon as they were no longer working. In large part, this was based on state statute or policy, and the fact that claims-filed is directly tied to federal funding amounts. Also, in the years leading up to the pandemic, many states were steering individuals to electronic-only filing methods rather than through in-person filing, due to historically low staffing (due to historically low claim rates), decreased staff presence in One-Stop Centers, and state efforts to modernize their benefit system.

Although states remain focused on serving legitimate claimants, they cannot go back to a pre-pandemic stance, as fraudsters are unrelenting in their attacks on UI systems. IDV is now core to state functions and has multiple impacts on state UI systems, ranging from speed of service delivery, benefit rights information that individuals receive and at which point during the claim-filing process, how and when other issues are adjudicated, the customer experience, staffing levels and use of staff within state agencies, workload distribution, and funding. Also, more government agencies and other entities are increasingly involved in combating UI fraud.

ID verification also matters to claimants. It is important for UI to reach the right people, on time. That is why it is critical for IDV services to be user-friendly, accurate, and fair. Overly complex application steps can cause claimant confusion and excessive contacts to state agency contact centers. If IDV is not smooth, it can create workloads for state agencies that they cannot manage while also paying claims on time. A cumbersome process can also exacerbate the stress of a larger economic downturn on individuals and the economy.

Project Overview: The Need for Information on the State of IDV

While IDV is seen as a critical bulwark against fraud in UI programs, not enough is known about the impacts of identity verification on fraud prevention. Implementation of IDV varies widely across states, including which vendors are used, which claimants are screened and when, and what combination of digital and non-digital identity verification is used. U.S. Department of Labor policies and oversight of state operations are still based on a pre-pandemic reality and are not optimized to incorporate and comprehensively track identity verification, and states are eager for guidance that better fit their reality. While fraud prevention is dynamic, there is a need to establish standards and expectations for identity verification, by identifying and normalizing evidence-based best practices.

This report gathers frontline information about the effectiveness of current practices, how states are measuring impact, and the implementation of compensating controls due to limits in IDV and cost considerations. In addition, the report discusses operational and policy issues, such as concerns about facial recognition for UI applicants with darker skin tones and ways states integrate IDV into their process in light of regulations. For the purpose of this analysis, representatives from UI offices in eleven states were interviewed,4 as well as other experts in fraud prevention and technology, vendors, and claimant advocates. Interviews were conducted with state leaders with intimate knowledge of current and past program operations across five of the six Employment and Training Administration (ETA) regions among a mix of ten large and medium states and one small state. These UI programs represent a range of service delivery models, operational challenges, vendors contracted, quality of relationship amongst program stakeholders, identity verification tools, and strategies for detecting and preventing fraud.

Current State IDV Processes, Practices, and Challenges

As with every other area of UI, each state has different processes and practices for ID verification. This section covers the technology, processes, and practices states use, while protecting details that could aid fraudsters. To minimize the risk of providing useful information to fraudsters, this report avoids technical detail about some interventions that states use, as well as where and when they are used.

As several stakeholders noted, there is no single solution or silver bullet for ID verification.

As several stakeholders noted, there is no single solution or silver bullet for ID verification. IDV is a major component in a multilayered fraud prevention program, with states often engaging with multiple vendors that interact in different ways in different states. States utilize IDV at various points throughout the claims process and vendors also expressed the need to have anti-fraud capabilities and different tools across the claim lifecycle. Notably, even some vendors appear to believe that layering with other vendor products is part of the answer, as part of a belt-and-suspenders approach. Several states even credited the relatively lower number of system attacks on the use of multiple vendors.

States and vendors vary in terms of at what point or points in the claims process they conduct ID verification. The U.S. Department of Labor issued guidance in early 2021 informing states that all claimants must be provided a written determination that can be appealed,5 and most of the states interviewed said they allow applicants to file an application before ID verification is completed to ensure their ID verification process is not making a de facto eligibility determination.

Some states treat a claimant’s nonresponse to ID verification requests as fraud while others issue appealable determinations or appealable failure-to-respond determinations. Still others have no way of knowing who has abandoned ID verification.

Some states have implemented—or are expected to implement—a statewide login that necessitates ID verification before navigating to the UI application, or otherwise requires ID verification before starting a claim. This affects the entire claims process, including when claimants receive information about their benefit rights, when other eligibility issues are adjudicated, what the customer experience looks and feels like, how staff is deployed, what workloads look like, and how much funding states need to operate. Vendors have acknowledged this rethinking is ongoing, with multiple vendors noting that states continue to reassess pandemic-era IDV policies and practices. Other states allow claimants to start a claim but do not begin the employer notification process until the IDV is complete to avoid unnecessary workload or aggravating employers of workers who have been impersonated, but this negatively affects payment timeliness. Moreover, even though states acknowledge that the various IDV steps contribute to how quickly a claim is processed, they all agree that there is no negative impact on overall timeliness goals.

The tools, tactics, and approaches best suited to effective IDV vary considerably, depending on factors such as claim rate, population characteristics, available funding, and the nature of the fraud threats a state faces. Smaller states can engage in a more personalized but labor-intensive approach. One small state reported that it has fifty-five internal fraud-prevention interventions it executes before ID verification to minimize burden on claimants. By contrast, another larger state has invested in automated tools that identified a particular fraud attack after approximately 150 claims were affected, compared to a previous attack identified by human staff only after several thousand claims had been impacted. Of course, interventions may need flexibility and a plan to scale in case of major economic downturn and dramatic increase in claims intake.

Often, states do not require ID verification for all claimants. Several vendors provide sophisticated products that provide a risk score on each claimant, and claimants that meet a certain threshold of risk are then referred to an ID verification service. At least one state has built its own in-house risk identification solution. These risk scores vary across states and require continuous refinement. This can save both states and claimants time by sometimes reducing the number of claimants who must go through every IDV step. States also report that these flags need frequent intervention to narrow down overflagging. To that end, states reported the need to be able to make these adjustments on the fly based on operational needs and according to the unique profiles of the given state’s populations. This was especially important for states using vendor scoring models because they were concerned about the lack of flexibility of the vendor’s product. States have an expectation that the product should be maximally customizable to meet their specific needs instead of limiting their requested changes to meet the needs of all other vendor customers. Further, vendors tended to express high confidence in their ability to prevent fraud while being flexible enough to customize to state needs.

Beginning in 2022, the U.S. Department of Labor set up a partnership with the U.S. Postal Service (USPS) to provide in-person identity verification with participating state agencies. States report positive feedback on the USPS partnership because it is an established and trusted entity. At least one state indicated that it has served as a reliable alternative, particularly for claimants who are resistant to providing biometric data required by digital platforms. Access to USPS has also been particularly helpful to claimants in rural areas who may lack Internet access. In some states, it had been the front door for a significant percentage of claimants. Claimant advocates reported that it has been crucial for claimants who struggle with digital literacy. They also pointed out that anti-fraud protocol increasingly leans on claimants being able to access and maneuver an electronic portal. Further, several states reported that USPS locations can be easier to get to, especially in areas where there is little access to One-Stop Centers, and at least one state said the availability of USPS helps the state to successfully manage their timeliness measures. Another state requires all out-of-state claimants to go through USPS. Yet, new guidance from U.S. Department of Labor now requires that claimants establish a Login.gov account6 before they can engage with USPS services. Meanwhile, maintaining USPS as a direct resource for claimants in the IDV process is a popular stance among states interviewed for this report. Also, at least one state expressed concern for whether expanded use of Login.gov could overload the Login.gov call center by claimants seeking IDV or claim resolution.

Because ID verification alone cannot make a determination of eligibility for a UI benefit, states provide multiple opportunities to clear issues as they arise. Since the pandemic, state staff have received training in how to spot false IDs. One state said that they strive for 80 percent in-person verification, while at least two other states couple in-person verification with Reemployment Services and Eligibility Assessment (RESEA) program participation. In addition to in-person verification at either USPS or a state One-Stop Center, some states have a process for verification over the phone. One state’s process for failing the initial online ID verification step is referral to USPS and will ask for a selfie for the UI agency’s review. Claimant advocates also noted that in-person off-ramps should be more prevalent and more publicized. Advocate groups also said there needs to be more flexibility on acceptable document types.

Vendor selection is a challenge, and states have been seeking help from the federal government in this regard, which is why several states appreciated the USPS partnership. Cost is a huge factor, as is mitigating friction for claimants and state agency staff. The public’s familiarity with the product also can sway an agency’s decision to adopt. Conversely, claimant unfamiliarity with a third-party verification entity and concerns about the privacy implications of engaging with yet another platform can create friction and erode trust. Stakeholder pressure can also shape vendor selection in significant ways. Claimant advocates have raised concerns about equitable access that steered states away from particular vendors. In general, states are sensitive to the potential downfalls of facial recognition technology, and multiple states cited the potential for racially disparate outcomes in making their decisions about whether to deploy that kind of technology. Others felt more confident in steps taken by vendors using facial recognition that they made a good faith effort to address those concerns.

States are sensitive to the potential downfalls of facial recognition technology, and multiple states cited the potential for racially disparate outcomes in making their decisions about whether to deploy that kind of technology.

States’ relationships with their IDV vendors are a significant factor in program effectiveness, and interviews revealed a complex dynamic. For example, one state raised concerns about their vendor’s reliance on historical identity data that may itself have been compromised in prior data breaches, raising questions about the reliability of that data as a fraud-detection signal. Interviews also revealed that states with the same vendor can have meaningfully different experiences depending on the specific product or service level they have purchased, which makes the case for greater communication and collaboration between states on vendor selection. This also demonstrates some of the reasons that states must meet with vendors regularly regarding flag accuracy and other performance conversations.

Since interviews were conducted, the federal government has taken several steps to expand its role in IDV,7 including by building the website unemployment.gov, so it may be the case that the vendor role and vendor selection process will shift dramatically in the coming years. However, states’ concerns about data security stemming from early efforts in this administration may cause reluctance in federal interventions involving data. One state specifically said that they do not like the concept of the federal government being the “end-all-be-all” in ID verification, as eligibility determination is the state’s role; rather, the federal role should consist of things like assisting with plain language and translations. Another state viewed having (up to) fifty-three front doors a strategic means of deterring fraudsters. Multiple states expressed a preference for receiving adequate federal funding with maximum flexibility to design their own IDV programs, rather than being steered toward particular vendors or standardized approaches.

Most states use a wide variety of crossmatches, particularly within the offerings provided by the National Association of State Workforce Agencies (NASWA) Integrity Data Hub (IDH). States are now looking to streamline crossmatches, particularly as a part of broader technology modernization.

Connections between agencies can be crucial. A couple of states reported that their staff have direct access to DMV records which they can review against submitted ID documents, and one noted that collaboration has caused a dramatic reduction in false positives.

A significant gap identified across multiple interviews is the absence of reliable methods for measuring IDV-related fraud. Multiple stakeholders noted that states generally lack good tools for quantifying the scope of IDV fraud, making it difficult to assess whether their investments in IDV tools are achieving their intended effect. Most states do their best to gather and evaluate data about when claimants abandon the process during the ID verification stage, and some of the leading states use that drop-off data to inform how to update their efforts to improve accuracy. However, it is very difficult to draw conclusions from drop-off data as to whether those who abandon the process are frustrated claimants or fraudsters. Some states are not even able to access their drop-off data to analyze it. Vendors could provide better data to states about their product’s effectiveness, including important metrics such as demographics and drop-offs. It is critical not just to know whether a drop-off was a fraudster, but also to understand what segments of the population are legitimate claimants who are having the most difficulty accessing benefits because of the hurdles imposed by a fraud intervention.

Emerging Trends in Combating Fraud

As states have become more sophisticated at identifying fraud, criminals have turned to other social engineering methods to attack the UI system. A common manifestation is claim hijacking, in which a fraudster gains access to a legitimate claimant’s account and changes the bank account to which benefits are directed. In response, states often require claimants to go back through identity verification in order to change their benefit payment method. Also, NASWA has added a bank account validation feature to the IDH.

Increasingly, sophisticated AI could also create more believable fake identification, which could require states to also develop a more sophisticated response to this. AI is also getting better at deep fakes, which means that fraudsters can use AI technology to create more believable “claimants” that are more difficult to distinguish from legitimate claimants.

Increasingly, sophisticated AI could also create more believable fake identification, which could require states to also develop a more sophisticated response to this.

As vendor contracts signed in response to the surge in fraud during the pandemic have been expiring, states are faced with whether to continue with their pandemic vendors, pivot to a new vendor, or focus on modernization and attempt to build in-house solutions. These decisions have broad implications on the state’s claim filing process and IDV protocol. Vendors interviewed indicated they are working with their state customers to help demonstrate current ROI.

In addition, more states appear to be moving in the direction of statewide single sign-on (SSO) across participating state agencies. One vendor recommended having a statewide IDV front door, with a statewide fraud score, while also saying that statewide SSOs are better if they allow for different types of risk profiles, allowing the state UI agency to maintain control. This echoes one state that was concerned that having a statewide SSO would possibly not meet the more stringent validation needs of UI applicants. Moreover, the vendor suggested that SSO helps manage budget constraints via cost-sharing and noted the customer service value in having a statewide front door. On a similar topic, two fraud experts shared the opinion that claimants who already verified with one government entity should be able to carry that verification to another government entity.

How to Make IDV More Seamless and Effective

Establishing and communicating commonly understood terminology across stakeholders will help to avoid confusion amongst partners and claimants alike. The wide variations between state practices and terminology is a hindrance to collaboration. For example, the field lacks a shared vocabulary. Interviews conducted for this project revealed meaningful variation in how different stakeholders define and refer to the same concepts. This definitional inconsistency is not merely semantic. At the outset of one state interview, agency staff asked interviewers to define what “ID verification” meant. Even within the state UI community, practitioners may be working from different assumptions about the scope and purpose of identity verification. Interviews also suggested that terminology has different meanings among federal entities, including the U.S. Department of Labor and the National Institute for Standards and Technology (NIST).

Sustainability is a major concern. ID verification is a relatively new responsibility that states must plan and budget for, and some recent grants from the U.S. Department of Labor to fight fraud have been rescinded, while the department has recently threatened state administrative funding over fraud-prevention efforts. States need predictable ongoing funding to pay for and maintain current efforts to combat fraud. This is going to be increasingly important in light of any potential future economic downturn because, as claims increase, so does the potential for fraudulent claims. States with limited resources may find the need to redirect funds committed to combating fraud to crisis response in paying new claims instead. Importantly, states are still funded on the same model as before the pandemic, when the nature of fraud was far less pervasive and sophisticated.

ID verification is a relatively new responsibility that states must plan and budget for.

There will always be a tension between preventing fraud and paying benefits in a timely manner. While fraud prevention ideally would be seamless and instant, that is not the case with current technology and practice. The current timeliness standard was set in 1971, when technology and the claim filing method was far more limited. That standard should not slip in 2026, when we should be able to find technology solutions to help speed those processes. The problem is that ID verification failure can trigger a manual investigation that is onerous and time-consuming. A promising practice that emerged from these interviews is ensuring that claimants have enough opportunity to verify ID within twenty-four hours, so as not to slow the overall claims. Fortunately, many states are now reporting that IDV is no longer having an effect on timeliness.

At least one state reported that working with U.S. Department of Labor Tiger Teams to adopt plain language and form simplification practices has helped to better sort actual claimants from fraudsters. When claimants do not understand what is being asked of them, why, or within what timeframe, they are more likely to get questions wrong, miss deadlines, abandon the verification process, and lose benefits to which they are entitled. Ironically, complicated forms are something that sophisticated fraud rings can master while filtering out first-time filers unfamiliar with jargon.

To the extent that states can minimize how invasive their interventions are without sacrificing security, saving unemployed claimants time and effort that should be focused on reemployment activities, states should seek to minimize the time claimants spend on bureaucracy. To that extent, a few states raised difficulties in document submission and management, and a couple claimant advocates identified factors around electronic document submission as being among the most challenging requirements for claimants. Developing solutions for this challenge that multiple states could implement could be helpful. Security practices such as multi-factor authentication, CAPTCHA, device fingerprinting, and other interventions that are minimally cumbersome for claimants can help root out fraud without creating hurdles that will slow claim processing, particularly during a potential future economic downturn.

One state noted that it would be helpful if third-party assistance, such as USPS, would retain a copy of IDs in case the state wants to further vet them. However, this could present an unacceptable liability on the behalf of those entities. Another state suggested that it would be optimal for states to build in-house services that are now purchased from vendors, which is unfortunately expensive at the outset but probably more cost-effective over time. To that end, when a state develops an effective yet cost-friendly fraud prevention and detection strategy, it should do its best to share that with other states.

Collaboration amongst stakeholders is key to success. State agencies that have productive relationships with their DMV counterparts tend to have better access to data than states with limited or no relationship.

Better communication across states would be helpful on a number of fronts. Trends have not slowed with regard to fraud. The emerging AI threat is a solid example of how states learning from each other is increasingly important. Similarly, information integration across state programs and systems could also spread out the cost and liability of identity verification. The same bad actors that attack UI systems are finding their way into other federal, state, and private programs. Just as sharing information across states is helpful, sharing information across state silos can be useful where it is legal and appropriate.

Appeals of ID verification determinations should have their own expedited process, which appears to be the case in many states, but claimant advocates in more than one state have indicated that the appeals process for these cases are either not separated from more complicated denials or that they are otherwise protracted. Many also have processes that intercept a formal appeal by offering a re-determination, which can be a less adversarial path that may better serve claimants who simply missed a deadline or need additional assistance. In addition, one state discussed training staff for special situations, citing months when public sector employees are regularly laid off and the challenges that are presented by information about them such as their names and salaries being available to the public, making fraudsters’ jobs easier.

Looking Ahead

Combating UI benefit fraud during the pandemic presented an enormous new challenge for states. In the ensuing years, they have become more sophisticated at detecting and preventing fraud, and ID verification in particular has been a sweeping success. However, combating fraud has also created hurdles that legitimate claimants must clear to access their UI benefits. In order for states to establish resilient IDV systems that do not prove overly onerous to legitimate claimants, stakeholders should consider pursuing the following:

  • Predictable funding. With the arrival of widespread UI claims fraud, states have a new fixed and ongoing cost that needs accounting for. States are going to have trouble being proactive and planning for the future without the acknowledgement of and support that meets this new reality. States are going to be increasingly skeptical of one-time or episodic funding after grants through the American Rescue Plan Act (ARPA) were rescinded last year. State leaders seem confident in their ability to protect state UI systems from IDV fraud, but maintain that federal funding support must match the ongoing threats.
  • Sharing information within states. While multiple states raised the advantage of information sharing with the DMV, it might also be helpful for various state-level entities to share information, such as a suspicious actor repository, across various state programs. Data sharing could help decrease vendor-related dependencies and expenses.
  • Sharing information between states. NASWA routinely convenes states to share patterns and things that states have learned. Importantly, sharing promising practices among states who share vendors is the most obvious use case that arose in interviews. States should be empowered to get feedback from other states who share the same vendor, whether they make a direct connection, or it is facilitated by the vendor or another third-party.
  • Sharing definitions and processes across states. This is an issue more broadly with UI administration, but particularly when new issues arrive, it would be helpful to establish standards and definitions for states.
  • Helping states understand what “good” looks like. At least two vendors noted that states often lack a clear picture of what effective identity verification looks like in practice, such as which metrics to track, which outcomes to target, or how to evaluate tradeoffs between fraud prevention and access. Federal technical assistance that helps states articulate and measure success could significantly improve program design.
  • Improvement in pathways for legitimate claimants who fail IDV. Claimants should have multiple paths to proving they are who they say they are if the ID verification fails. This includes an easier appeal path if the only disqualification is failing IDV.
  • Better data. State agencies need to have more information about who is abandoning claims and why, and also have demographic data about who is successfully navigating IDV and who is not; if vendors do not supply this data, agencies cannot know whether a claimant drop-off is a fraudster denied, or a legitimate claimant prevented from getting benefits. This information also would help states target outreach and assistance.
  • Guardrails on IDV technology so states are not on their own in navigating. The UI program will continue to be exposed to rapidly evolving technologies and vendor products, particularly in regards to advancements in AI. States should have a clear understanding about data security standards, including how claimant data is stored, shared, and protected, along with a framework for ensuring accountability. States need these supports while maintaining operational flexibility in administering UI programs.

Notes

  1. John Pallasch, “Advisory: Unemployment Insurance Program Letter No. 28-20, Change 1: Additional Funding for Identity Verification or Verification of Pandemic Unemployment Assistance (PUA) Claimants and Funding to Assist with Efforts to Prevent and Detect Fraud and Identity Theft as well as Recover Fraud Overpayments in the PUA and Pandemic Emergency Unemployment Compensation (PEUC) Programs,” U.S. Department Of Labor, Employment and Training Administration, January 15, 2021, https://www.dol.gov/sites/dolgov/files/ETA/advisories/UIPL/2021/UIPL_28-20_Change_1.pdf.
  2. Michele Evermore and Kevin Parker, “State Approaches to Unemployment Insurance Program Integrity: Early Findings of a National Survey of State UI Agencies,” National Academy of Social Insurance, August 2026, https://www.nasi.org/wp-content/uploads/2026/08/State-Approaches-to-Unemployment-Insurance-Program-Integrity-Brief.pdf.
  3. Data received by authors from the Employment and Training Administration (ETA), Office of Unemployment Insurance.
  4. The states interviewed may be disclosed upon request. States interviewed varied in population size. The regional distribution includes five of the Department of Labor’s six regions.
  5. Suzan G. Levine, “Unemployment Insurance Program Letter No. 16-21: Identity Verification for Unemployment Insurance (UI) Claims,” Employment and Training Administration, U.S. Department of Labor, April 13, 2021, https://www.dol.gov/sites/dolgov/files/ETA/advisories/UIPL/2021/UIPL_16-21.pdf.
  6. Henry Mack, “Unemployment Insurance Program Letter No. 10-26: Announcement of Enhancements to Identity Verification Services and Funding Opportunity to Support State Costs Associated with Participation in Such Services,” Employment and Training Administration, U.S. Department of Labor, March 31, 2026, https://www.dol.gov/sites/dolgov/files/ETA/advisories/UIPL/2026/UIPL%2010-26/UIPL%2010-26.pdf.
  7. Ibid.

Tags: , ,