Alert Memorandum — ETA Needs Data Analytics Capability to Improve Oversight of the UI Program

Record facts

Summary

An alert memorandum issued September 25, 2023 by the U.S. Department of Labor Office of Inspector General to the Principal Deputy Assistant Secretary for Employment and Training, Report Number: 19-23-012-03-315. It reports nearly $1.3 billion in potentially fraudulent unemployment insurance benefits paid from March 2020 through April 2022 in two high-risk age categories: Social Security numbers of children under the age of 14, $1,225,663,851, and of persons 100 years of age or older, $66,541,872. It states this brings the cumulative total in six high-risk areas to more than $46.9 billion, after more than $45.6 billion previously reported. The memorandum says the Employment and Training Administration lacks direct access to state UI claims data and the capability to analyze it, and notes UI programs paid $888 billion in benefits from March 2020 through September 2021.

Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used

Full text

U.S. Department of Labor 
Office of Inspector General 
 
 
Washington, DC 20210 
Working for America’s Workforce 
September 25, 2023 
 
 
MEMORANDUM FOR: 
BRENT PARTON 
 
 
 
 
Principal Deputy Assistant Secretary  
 
 
 
 
  for Employment and Training 
 
                                        
 
FROM: 
 
 
CAROLYN R. HANTZ 
 
 
 
 
Assistant Inspector General  
  for Audit 
 
 
SUBJECT: 
Alert Memorandum: ETA Needs to 
Incorporate Data Analytics Capability to 
Improve Oversight of the Unemployment 
Insurance Program  
Report Number: 19-23-012-03-315 
 
The purpose of this memorandum is to alert the Employment and Training 
Administration (ETA) to a concern the Office of Inspector General (OIG) has 
determined needs immediate attention. The OIG has identified nearly $1.3 billion 
in potentially fraudulent unemployment insurance (UI) payments made during the 
pandemic in two high-risk age categories, to individuals with Social Security 
numbers: (1) of children under the age of 14 and (2) of elderly persons 100 years 
of age or older. The OIG previously identified more than $45.6 billion in 
potentially fraudulent UI pandemic benefits paid in four other high-risk areas.1 
This memorandum builds on our previous work, identifying additional risk within 
state UI claims data for agency action. 
 
We are concerned ETA currently does not have direct access to state UI claims 
data. Additionally, we are concerned ETA does not have the capability to analyze 
said data, which would allow it to better identify fraud and other improper 
payments, as well as other trends or emerging issues, such as timeliness or 
equity.  
 
                                            
1 Alert Memorandum: Potentially Fraudulent Unemployment Insurance Payments in High-Risk 
Areas Increased to $45.6 Billion, Report No. 19-22-005-03-315, (September 21, 2022), available 
at: https://www.oig.dol.gov/public/reports/oa/2022/19-22-005-03-315.pdf  

-2- 
To provide relief to American citizens experiencing pandemic-related 
employment issues and challenges, the Coronavirus Aid, Relief, and Economic 
Security Act and subsequent legislation authorized seven new temporary UI 
programs.2 These temporary pandemic-related UI programs, coupled with the 
state UI programs, paid $888 billion in benefits during the heightened period of 
the COVID-19 pandemic—March 2020 through September 2021. This rapid 
expansion and substantial increase in benefits significantly increased the risk for 
fraud, waste, and abuse in UI programs. 
 
For more than 20 years, the OIG has reported on the U.S. Department of Labor’s 
(DOL) challenge to reduce improper payments in the UI program, which has 
experienced some of the highest improper payment rates across the federal 
government. In addition to the OIG’s prior work, the Government Accountability 
Office (GAO), in its 2021 report on the federal response to the pandemic, 
expressed concern about overpayments and potential fraud in the UI system.3 
GAO subsequently added the UI system to its “High Risk List” and recommended 
that DOL develop a plan for transforming this system.4 In 2023, GAO estimated 
the pandemic related fraud rate for the UI programs was about 11 to 15 percent 
for the period April 2020 to May 2023, and estimated up to $135 billion was lost 
to fraud.5 GAO also noted that the full extent of UI fraud during the pandemic will 
likely never be known with certainty. 
 
The improper payment rate estimate for the UI program, as reported to the Office 
of Management and Budget, has been above 10 percent for 16 of the last 
19 years (see Figure 1).6 In 2021 and 2022, ETA estimated an improper payment 
rate of 18.71 percent and 21.52 percent, respectively. Further, ETA estimated a 
                                            
2 These new programs included the Pandemic Unemployment Assistance program, which 
extended UI benefits to individuals not traditionally eligible; the Federal Pandemic Unemployment 
Compensation program, which provided supplemental payments to individuals receiving 
traditional and other eligible UI benefits; and the Pandemic Emergency Unemployment 
Compensation program, which provided up to an additional 13 weeks of unemployment 
compensation to individuals who exhausted their regular UI benefits. 
3 GAO, COVID-19: Sustained Federal Action is Crucial as Pandemic Enters Its Second Year, 
Report No. GAO-21-387 (March 31, 2021), available at: https://www.gao.gov/products/gao-21-
387 
4 GAO, Unemployment Insurance: Transformation Needed to Address Program Design, 
Infrastructure, and Integrity Risks, Report No. GAO-22-105162 (June 7, 2022), available at: 
https://www.gao.gov/products/gao-22-105162  
5 GAO, Unemployment Insurance: Estimated Amount of Fraud during Pandemic Likely Between 
$100 Billion and $135 Billion, GAO-23-106696 (September 2023), available at: 
https://www.gao.gov/assets/gao-23-106696.pdf  
6 The improper payment reporting year is the 12-month period ending June 30 of the reporting 
year. 

-3- 
fraud rate of 8.57 for 2021—a 170 percent increase over the prior year’s fraud 
rate of 3.17 percent.7 
 
Figure 1: UI Improper Payment Rates, 2004 – 20228 
 
 
Source: OIG analysis of improper payment data published by ETA 
 
During our previous audit9 of DOL’s response to the UI program’s expansion 
under the Coronavirus Aid, Relief, and Economic Security Act, we performed 
comprehensive data analyses on state UI claims data for March 2020 through 
October 2020. Our analyses identified more than $5.4 billion of potentially 
fraudulent UI pandemic benefits paid in four specific high-risk areas. The four 
originally identified high-risk areas were benefits paid to individuals with Social 
Security numbers: (1) filed in multiple states, (2) of deceased persons, (3) of 
federal prisoners, and (4) used to file UI claims with suspicious email accounts.10  
                                            
7 ETA estimated the fraud rate as part of its Benefit Accuracy Measurement program for the 
period July 1, 2020, through June 30, 2021. The Benefit Accuracy Measurement program is 
designed to determine the accuracy of paid and denied claims in three major UI programs: state 
UI, Unemployment Compensation for Federal Employees, and Unemployment Compensation for 
Ex-Servicemembers. 
8 The 2020 improper payment rate of 9.7 percent was calculated based on 9 months of data, from 
July 2019 through March 2020, and was reflective of only 8 percent of total program year 
expenses due to the exclusion of Coronavirus Aid, Relief, and Economic Security Act UI 
expenditures and the fourth quarter of the program year. The collection of improper payment data 
from April 2020 to June 2020 was suspended due to the pandemic. 
9 COVID-19: States Struggled to Implement CARES Act Unemployment Insurance Programs, 
Report No. 19-21-004-03-315 (May 28, 2021), available at: 
https://www.oig.dol.gov/public/reports/oa/2021/19-21-004-03-315.pdf  
10 Alert Memorandum: The Employment and Training Administration (ETA) Needs to Ensure 
State Workforce Agencies (SWA) Implement Effective Unemployment Insurance Program Fraud 
Controls for High Risk Areas, Report No. 19-21-002-03-315 (February 22, 2021), available at: 
https://www.oig.dol.gov/public/reports/oa/2021/19-21-002-03-315.pdf  
0%
5%
10%
15%
20%
25%
Improper Payment Rate

-4- 
 
Our subsequent analyses of those high-risk areas, using UI claims data through 
April 2022, brought the cumulative total of potential fraud to more than 
$45.6 billion.11 We provided ETA with our underlying methodology as well as 
specific claimant information, and ETA shared the information with states for 
appropriate action. We are currently performing four separate audits to examine 
the extent to which ETA and states have taken action on a sample of potentially 
fraudulent claims in each of the four high-risk areas. 
 
Since our earlier analyses, we have identified almost $1.3 billion in potentially 
fraudulent UI pandemic benefits paid from March 2020 through April 2022 in two 
additional high-risk areas. The two additional high-risk areas involve benefits paid 
to individuals with Social Security numbers: (1) of children under the age of 14 
and (2) of elderly persons 100 years of age or older (see Table 1).12  
 
Table 1: UI Pandemic Benefits Paid to High-Risk Age Groups, 
March 2020 through April 2022 
 
High-Risk Area 
Total Potential Fraud Identified 
Under Age 14 
$1,225,663,851 
Age 100 or Older 
$66,541,872 
Total 
$1,292,205,723 
Source: OIG data analysis of state UI claims data 
 
Although these could be legitimate claims, they merit additional oversight and 
scrutiny as workers typically do not fall into these age categories. For example, 
the Fair Labor Standards Act of 1938 sets a minimum age of 14 for employment 
in non-agricultural occupations covered by the act, which effectively limits 
employment for children under 14 to work that is exempt from the act, such as 
delivering newspapers and acting.  
 
In addition to the potential fraud risk inherent with children under 14 receiving UI 
benefits, there is a risk and concern that these children could be victims of child 
labor exploitation. Since 2018, DOL has seen a 69 percent increase in children 
being employed illegally by companies. In Fiscal Year (FY) 2022, DOL found 
835 companies it investigated had employed more than 3,800 children in 
violation of labor laws. The OIG is planning to do audit work in this area—which 
falls under the purview of DOL’s Wage and Hour Division—in the future. 
 
                                            
11 Alert Memorandum: Potentially Fraudulent Unemployment Insurance Payments in High-Risk 
Areas Increased to $45.6 Billion, Report No. 19-22-005-03-315 (September 21, 2022), available 
at: https://www.oig.dol.gov/public/reports/oa/2022/19-22-005-03-315.pdf  
12 To prevent double counting, these results do not include any claims identified in our analyses 
of the four high-risk areas covered by our previous alert memoranda.  

-5- 
Further, while there are no legal restrictions placed on elderly workers, it is rare 
for individuals over the age of 100 to be in the workforce. The U.S. Bureau of 
Labor Statistics reported the labor force participation rate for people age 75 and 
older was 8.9 percent in 2020.13  
 
Our analysis identified four states that paid UI claims to 18 percent or more of 
individuals aged 100 years or older from March 2020 to April 2022.14 If 
these claims were legitimate, this would indicate a remarkably large percentage 
of centenarians were still working and eligible for UI benefits in these four 
states—a trend that was not observed in other states. For example, the State of 
Michigan paid claims to over 58 percent of its centenarian population during this 
period, while other states with significantly larger centenarian populations paid 
claims to less than 1 percent. Excluding the four outlier states with percentages 
exceeding 18 percent, an average of only 1.2 percent of centenarians in the 
remaining 46 states received UI pandemic benefits during this period (see 
Figure 2).  
 
Figure 2: Percentage of Persons 100 Years of Age or Older with UI 
Pandemic Benefits Paid from March 2020 through April 2022 
 
 
Source: OIG data analysis of state UI claims data and 2020 U.S. Census data 
 
                                            
13 Bureau of Labor Statistics, U.S. Department of Labor, The Economics Daily, “Number of people 
75 and older in the labor force is expected to grow 96.5 percent by 2030” (November 4, 2021), 
last accessed August 15, 2023, available at: https://www.bls.gov/opub/ted/2021/number-of-
people-75-and-older-in-the-labor-force-is-expected-to-grow-96-5-percent-by-2030.htm 
14 For this comparison, we used state population data sourced from the 2020 U.S. Census. U.S. 
Census Bureau, “Exploring Age Groups in the 2020 Census” (May 25, 2023), available at: 
https://www.census.gov/library/visualizations/interactive/exploring-age-groups-in-the-2020-
census.html  
1.2%
58.5%
22.9%
21.1%
18.8%
0%
10%
20%
30%
40%
50%
60%
70%
Average State
Michigan
Rhode Island
California
Georgia
Percentage of Centenarians Receiving UI Benefits

-6- 
Due to the results of our analysis in these two high-risk areas, we remain 
concerned about the amount of benefits paid in the aforementioned age groups. 
Further, our latest analysis brings the cumulative total amount of potentially 
fraudulent payments to more than $46.9 billion15 in the six high-risk areas (see 
Figure 3),16 showing how effective and beneficial data analytics can be for 
providing effective program oversight and combatting fraud. 
 
Figure 3: Six High-Risk Areas for Potential UI Fraud Identified by the OIG 
 
 
Source: OIG data analysis of state UI claims data 
 
In “A Framework for Managing Fraud Risks for Federal Programs,” GAO 
recommends using data analytic tools and techniques to prevent and detect 
fraud. It states:  
 
Data analytics activities can include a variety of techniques to 
prevent and detect fraud. For example, data mining and 
data-matching techniques can enable programs to identify potential 
fraud or improper payments that have already been awarded, thus 
assisting programs in recovering these dollars, while predictive 
analytics can identify potential fraud before making payments.17  
 
                                            
15 Total amounts for these analyses do not include duplicates that were identified in one or more 
areas. 
16 Of the more than $46.9 billion total identified, the OIG had previously reported in prior OIG 
reports that $45.6 billion in potential fraudulent payments had been identified within four high-risk 
areas. This report identifies almost $1.3 billion in additional funds put to better use within two 
additional high-risk areas that were not claimed in our prior reports. See Attachment II for detailed 
information, including our calculation. 
17 Appendix III in the GAO’s “A Framework for Managing Fraud Risks in Federal Programs,” 
Report No. GAO-15-593SP (July 2015), available at: https://www.gao.gov/assets/gao-15-
593sp.pdf 

-7- 
In addition, the Office of Management and Budget recommends that 
federal agencies work closely with their data analysts to establish robust 
analytics capabilities that allow an agency to identify potential improper 
payments, including fraud, before they occur. Data analytics can include 
analysis for trends, patterns, anomalies, and exceptions within data sets. 
The Office of Management and Budget also provides an overview of 
common data analytic techniques that federal agencies should consider to 
reduce the risks of improper payments.18  
 
According to ETA, it has developed a preliminary draft of a UI Fraud Risk Profile 
to align DOL’s ongoing fraud risk management activities with GAO’s Fraud Risk 
Framework. ETA also issued guidance in April 2023 to communicate the 
antifraud strategy for the UI program to the states. This guidance provides states 
with information on both required and strongly recommended strategies, tools, 
and services for UI fraud risk mitigation and improper payment reduction.19 
 
ETA’s FY 2023 Agency Management Plan states that improving overall program 
integrity in the UI program is one of the agency’s top priorities. Strategy 3.2, 
“Strengthen Program Integrity by Preventing and Detecting Fraud, and Reducing 
Improper Payments,” indicates that ETA will continue to “promote UI program 
integrity actively and aggressively by identifying and reducing fraud, waste, and 
abuse in the program.” One of the activities for this strategy is to “invest in 
technology and data analytics,” and specifically: 
 
Invest in tools and strategies to help states verify [the] identity of UI 
claimants, promote states’ use of these tools/strategies, and 
perform data analytics to prevent and detect fraud, and reduce 
investigations backlog. These tools and strategies will be balanced 
to ensure equitable access and delivery of UI services.  
 
Another activity identified in ETA’s FY 2023 Agency Management Plan is to work 
with the UI Integrity Center “to promote the consistent use of the Integrity Data 
Hub (IDH) by states to cross-match with all available datasets and promote the 
UI Integrity Center’s Integrity Knowledge Exchange as a resource for program 
integrity information, tools (including a data analytics tool), and promising 
practices to states.” 
 
                                            
18 Payment Integrity Alert: The Use of Automation and Data Analytics From the Office of 
Management and Budget (OMB) Office of Federal Financial Management and the Pandemic 
Response Accountability Committee (PRAC); posted by the CFO Council on July 21, 2021, 
available at: https://www.pandemicoversight.gov/media/file/joint-payment-integrity-alert-use-
automation-and-data-analytics-omb-and-prac  
19 Unemployment Insurance Program Letter No. 22-21, Change 2, available at: 
https://www.dol.gov/agencies/eta/advisories/uipl-22-21-change-2  

-8- 
The UI Integrity Center is a joint federal-state initiative funded by DOL and 
operated by the National Association of State Workforce Agencies.20 It serves as 
a resource to assist states in their efforts to improve integrity in the UI program, 
focusing particularly on the prevention, detection, and recovery of improper 
payments including fraud. One of the services provided by the UI Integrity 
Center, the IDH, is a multistate data system used for advanced data 
cross-matching and analysis on submitted claims to help detect and prevent UI 
fraud and improper payments. 
 
While tools and resources like the IDH can improve program integrity and 
increase detection and prevention of improper payments, including fraudulent 
payments, there are significant limitations. First, while ETA encourages states to 
use the tools available from the UI Integrity Center, ETA maintains that it lacks 
the authority to require states to participate. To this end, the OIG previously 
recommended in our first high-risk alert memo,21 issued in February 2021, that 
ETA work with Congress to establish legislation requiring state workforce 
agencies to cross-match high-risk areas, including the four originally identified 
areas. Consistent with our recommendation, in the FY 2024 President’s Budget, 
ETA included a legislative proposal that would give DOL authority to require 
states to perform cross-matches using the IDH. Additionally, states that 
participate in the IDH have been inconsistent in their use of the tool and in the 
level of claims data shared. For example, as of June 2023, although all 53 state 
workforce agencies had IDH participation agreements in place, ETA reported 
only 50 of the 53 state workforce agencies were participating in the IDH 
multistate cross-matches. 
 
We also found the IDH is limited in its effectiveness in identifying potential 
improper payments and fraud. We recently reported that the IDH was not able to 
identify the same amount of potentially improper multistate claims when 
compared to the OIG’s analysis.22 The IDH used a threshold that was higher than 
what the OIG would use to flag improper multistate claims. However, even when 
we used the IDH’s threshold, we found the IDH only identified 39.7 percent of 
potential improper multistate claims using data from September 2020, when 
compared to our analysis. This occurred because the states who participated did 
not share complete data with the IDH on a consistent basis.  
 
                                            
20 The National Association of State Workforce Agencies is the national organization representing 
all of the state workforce agencies and whose mission is to enhance the state workforce 
agencies’ abilities to accomplish their goals, statutory roles, and responsibilities. 
21 Alert Memorandum: The Employment and Training Administration (ETA) Needs to Ensure 
State Workforce Agencies (SWA) Implement Effective Unemployment Insurance Program Fraud 
Controls for High Risk Areas, Report No. 19-21-002-03-315 (February 22, 2021), available at: 
https://www.oig.dol.gov/public/reports/oa/2021/19-21-002-03-315.pdf 
22 COVID-19 – ETA Can Improve its Oversight to Ensure Integrity over CARES Act UI Programs, 
Report No. 19-23-011-03-315 (September 22, 2023), available at: 
https://www.oig.dol.gov/public/reports/oa/2023/19-23-011-03-315.pdf  

-9- 
Finally, even without these limitations, the tools and resources do not relieve ETA 
from its program oversight responsibilities. ETA has recognized the critical 
importance of data analytics in detecting and preventing improper payments, 
including fraud, and improving program integrity. To this end, ETA has invested 
in providing states with tools through the UI Integrity Center and encouraged 
states to use data analytics to strengthen program integrity. However, ETA does 
not have direct access to UI claims data, nor has it developed its own data 
analytics capability at the federal level as part of its oversight of the UI program. 
 
In its FY 2023 and FY 2024 Congressional Budget Requests for State 
Unemployment Insurance and Employment Service Operations, ETA included a 
legislative proposal to allow it to collect and store states’ UI claimant data. This 
proposal would provide DOL “direct access to all claim and wage data used by 
state agencies in administrating the state’s unemployment compensation 
program solely for the purposes of Federal unemployment compensation 
administration and to conduct research, evaluation, and performance 
assessments of unemployment compensation programs and federally funded 
employment-related programs.”23 Therefore, ETA is aware of its need for access 
to state UI claims data to provide proper oversight of the UI program. 
 
ETA officials indicated the noted language in its Congressional Budget Requests 
was included as a component of future UI reform, but it would require statutory 
authority allowing access to the data, a new records management system, and 
individual data sharing agreements with states. Because this would take a 
significant increase in funding and years to plan and achieve, ETA indicated it is 
focusing its energies and resources on improving integrity controls in the federal-
state partnership that reflect the current UI program, such as investing in 
enhancements to, promoting increased states participation in, and conducting 
evaluations of the IDH. 
 
Although ETA has stated it needs additional authority to obtain access to state UI 
claims data, current federal regulations provide such authority for purposes 
described in this memorandum. Specifically, 20 C.F.R. § 603.6 (a) states that 
“the Department of Labor interprets Section 303(a)(1), [Social Security Act of 
1935], as requiring disclosure of all information necessary for the proper 
administration of the [UI] program.” Additionally, 29 C.F.R. § 96.41 authorizes 
DOL to obtain UI data from the states to evaluate and improve program integrity. 
This regulation states that DOL “shall have access to any books, 
documents…and records (manual and automated) of the entity receiving funds 
                                            
23 FY 2023 Congressional Budget Justification for State Unemployment Insurance and 
Employment Service Operations, available at: 
https://www.dol.gov/sites/dolgov/files/general/budget/2023/CBJ-2023-V1-07.pdf; FY 2024 
Congressional Budget Justification for State Unemployment Insurance and Employment Service 
Operations, available at: https://www.dol.gov/sites/dolgov/files/general/budget/2024/CBJ-2024-
V1-07.pdf 

-10- 
from DOL…for the purpose of making surveys, audits, examinations, excerpts, 
and transcripts.” 
 
Conclusion 
 
Ready access to UI claims data from all states and territories would enable ETA 
to ensure proper administration and provide sufficient oversight of the UI 
program.  
 
Further, in addition to the tools and resources it provides to the states for fraud 
detection, such as the IDH, establishing a data analytics capability with a 
dedicated team of data scientists at the federal level would allow ETA to monitor 
and analyze UI claims data on an on-going basis. ETA would then be able to 
identify high-risk areas across multiple states and quickly flag potentially 
fraudulent claims that can be referred to the OIG and states for further action, 
which could help prevent future losses to fraudsters. Likewise, incorporating a 
data analytics capability into its program oversight function would improve ETA’s 
ability to detect trends and emerging issues that could negatively impact the 
timeliness or equity of UI payments—before the issues grow into critical 
problems.  
 
Ultimately, a data access and analytics capability would allow ETA management 
to make better informed decisions about the UI program and help mitigate the 
risk of improper payments including fraud, while also preparing ETA for future 
emergency UI programs where hundreds of billions in federal dollars could be at 
risk. 
 
Recommendations 
 
We recommend the Principal Deputy Assistant Secretary for Employment and 
Training: 
 
1. Obtain direct access to unemployment insurance claims data from all state 
workforce agencies.  
 
2. Create an integrity program that incorporates a data analytics capability 
and regularly monitors state unemployment insurance claims data to 
detect and prevent improper payments, including fraudulent payments, 
and to identify trends and emerging issues that could negatively impact 
the unemployment insurance program.  
 
3. Establish effective controls, in collaboration with state workforce agencies, 
to mitigate fraud and other improper payments to ineligible claimants in 
high-risk age categories. 
 

-11- 
Summary of ETA’s Response 
 
On September 6, 2023, ETA provided us its formal response to the draft alert 
memorandum and recommendations (see Attachment III). ETA agreed with 
Recommendation 3 and stated it had already implemented controls to mitigate 
fraud and improper payments to ineligible claimants in high-risk age categories. 
However, ETA disagreed with Recommendations 1 and 2, stating that they are 
not achievable at this time without significant new and ongoing appropriations.  
 
For Recommendation 1, ETA stated that it lacks resources to develop a data 
warehouse that would be required to implement this recommendation. For 
Recommendation 2, ETA stated that developing a data analytics capability at the 
federal level would duplicate DOL’s ongoing investment in the UI Integrity Center, 
including the IDH. To address these two recommendations, ETA suggested an 
alternative approach of working with the UI Integrity Center to improve IDH data 
analytics capabilities to better identify fraud trends. This would include meeting 
regularly with the UI Integrity Center to receive information about identified fraud 
trends and to discuss efforts to mitigate fraud and reduce improper payments.  
 
ETA raised two additional concerns with our alert memorandum, related to our 
analyses of payments to elderly individuals 100 years of age or older and ETA’s 
legal authority to access state UI claims data.  
 
First, ETA indicated that one of the four outlier states we identified in our analysis 
had manipulated the date of birth field in the process of creating “pseudo 
records” for previously identified fraudulent claims. This was done in accordance 
with ETA guidance to ensure victims of fraud were not unfairly prevented from 
accessing benefits. According to ETA, 90 percent of the claims we identified for 
this state were “pseudo records” and not actual payments to individuals 
100 years of age or older. ETA also stated that “it advised the OIG of this 
information and...is disappointed that the OIG did not include this context in the 
draft alert memorandum....”  
 
Second, ETA raised a concern that the alert memorandum exaggerates DOL’s 
legal authority to obtain data from states. Specifically, ETA stated it can only 
require the states to disclose claims information that is necessary for the proper 
administration of the program, and it cannot require states to provide DOL 
access to all claims data.  
 
OIG Response to ETA Management Comments  
 
For Recommendation 1, the OIG recognizes the resource concerns expressed 
by ETA and acknowledges there may be other ways that ETA can initially obtain 
direct access to state UI claims data. For example, with a more complete and 
consistent data set, the data maintained by the UI Integrity Center could be a 

-12- 
potential source for this data until such a time that ETA can allocate funding to 
establish its own data warehouse. 
 
For Recommendation 2, the OIG disagrees that developing a data analytics 
capability at the federal level would be a duplicative effort. As we noted in our 
memorandum, data analytics is an essential tool for program oversight—which is 
ETA’s responsibility—that would allow ETA to improve its detection and 
prevention of improper payments, including fraud, perpetrated across multiple 
states, while also preparing ETA for future emergency UI programs where federal 
money could be at risk.  
 
In relation to ETA’s suggested alternative approach, while the suggested actions 
would likely benefit the UI program, this approach appears to continue shifting 
program oversight responsibilities to the states and the UI Integrity Center. We 
noted in our memorandum that states are not required to participate in the IDH, 
and those that do participate do so to varying degrees—which has limited the 
IDH’s effectiveness. However, even if these limitations were resolved, it would 
not relieve ETA from its program oversight responsibilities. As such, ETA’s 
suggested alternative approach would not alone meet the intent of the OIG’s 
recommendations. 
 
In relation to the additional concern on our analysis of claims paid to claimants 
100 years of age or older, the OIG acknowledged in our memorandum that these 
claims could be legitimate but they required additional oversight and scrutiny. 
ETA did not notify us that the claims data contained “pseudo records” with 
altered dates of birth until it provided that information in its response to our draft 
memorandum.  
 
Additionally, ETA did not provide us with evidence to support its statement that 
90 percent of the claims the OIG identified for this state were “pseudo records.” If 
accurate, this state paid UI claims to 5.9 percent of its centenarians, which was 
almost 5 times higher than the average of the 46 non-outlier states. The OIG 
stands by our assertion that all claims paid in high-risk age categories require 
additional oversight and scrutiny.  
 
We further note that the UI program has experienced historic levels of improper 
payments. The OIG has issued multiple alert memoranda identifying tens of 
billions of dollars in potentially fraudulent UI payments identified through our use 
of data analytics. In addition, GAO recently reported that up to $135 billion could 
have been lost to UI fraud during the pandemic. 
 
In relation to DOL’s legal authority to obtain data from states, the OIG did not 
recommend ETA acquire unnecessary data but rather, we recommended ETA 
obtain direct access to UI claims data from all states to improve the accuracy and 
reliability of results when performing data analytics and oversight of the UI 

-13- 
program. ETA should access data that is necessary for, and relevant to, program 
administration and oversight. 
 
We consider Recommendations 1 and 2 as open and unresolved. For 
Recommendation 1, ETA should obtain access to the data, ideally developing its 
own data warehouse; however, until ETA receives sufficient funding, it may be 
able to obtain access via the UI Integrity Center. For Recommendation 2, ETA 
should not rely totally on the UI Integrity Center for data analytics capability. 

Attachment I 
 
-14- 
UI Improper Payments Rates, 2004 – 2022 
 
This table provides accessible data showing the reported improper payment rate 
estimates for the UI program over the last 19 years, from 2004 to 2022. The 
improper payment rate exceeded 10 percent in all years except 2008 
(9.96 percent), 2013 (9.32 percent), and 2020 (9.17 percent).24 
 
Table 2: Accessible Data Table for  
Figure 1 - UI Improper Payment Rates, 2004 – 2022 
 
Year 
Improper Payment Rate 
2004 
10.34% 
2005 
10.13% 
2006 
10.90% 
2007 
10.30% 
2008 
9.96% 
2009 
10.30% 
2010 
11.20% 
2011 
12.00% 
2012 
11.42% 
2013 
9.32% 
2014 
11.58% 
2015 
10.73% 
2016 
11.65% 
2017 
12.50% 
2018 
13.05% 
2019 
10.61% 
2020 
9.17% 
2021 
18.71% 
2022 
21.52% 
Source: OIG analysis of improper payment data 
published by ETA 
                                            
24 The actual improper payment rate for 2020 is likely higher than 10 percent, based on OIG 
pandemic audit and investigative work. On August 6, 2021, we reported that the improper 
payment rate of 9.17 percent for 2020 was calculated based on 9 months of data, from July 2019 
through March 2020, and was reflective of only 8 percent of total program year expenses 
($22.6 billion out of $281.8 billion) due to the exclusion of Coronavirus Aid, Relief, and Economic 
Security Act UI expenditures and the fourth quarter of the program year. The collection of 
improper payment data from April 2020 to June 2020 was suspended due to the pandemic. For 
more information, see “The U.S. Department of Labor Complied with The Payment Integrity 
Information Act for FY 2020, but Reported Unemployment Insurance Information Did Not 
Represent Total Program Year Expenses,” Report No. 22-21-007-13-001 (August 6, 2021), 
available at: https://www.oig.dol.gov/public/reports/oa/2021/22-21-007-13-001.pdf.  

Attachment II 
 
-15- 
Potential Funds for Better Use25 
 
Table 3: Total Net Funds for Better Use 
 
Description 
Amount 
Total Funds for Better Use 
$46,931,696,330 
Funds for Better Use Claimed in 
Prior OIG Alert Memoranda 
($45,639,490,607) 
Net Funds for Better Use 
$1,292,205,723 
Source: OIG data analysis of state UI claims data 
 
The table shows the total net funds for better use for the six high-risk areas 
identified in this and previous alert memoranda. In this alert memorandum, we 
identified $46.9 billion in total cumulative potential fraud for the period 
March 2020 to April 2022. To prevent double counting, we subtracted the 
$45.6 billion in potentially fraudulent payments identified in a previous 2022 alert 
memorandum.26 As a result, we are claiming nearly $1.3 billion as total net funds 
for better use in this alert memorandum. 
 
 
 
 
                                            
25 As defined by the Inspector General Act of 1978, as amended, “funds for better use” means 
funds that could be used more efficiently or achieve greater program effectiveness if 
management took certain actions. These actions include reduction in future outlays and 
deobligation of funds from programs or operations. 
26 Alert Memorandum: Potentially Fraudulent Unemployment Insurance Payments in High-Risk 
Areas Increased to $45.6 Billion, Report No. 19-22-005-03-315, (September 21, 2022), available 
at: https://www.oig.dol.gov/public/reports/oa/2022/19-22-005-03-315.pdf  

Attachment III 
 
-16- 
 
 
 

Attachment III 
 
-17- 
 
 
 

Attachment III 
 
-18- 
 
 
 

Attachment III 
 
-19- 
 
 
 

Attachment III 
 
-20- 
 
 
 

Attachment III 
 
-21- 
 
 
 

Attachment III 
 
-22- 
 
 
 

Attachment III 
 
-23- 
 
 
 

Attachment III 
 
-24-

File and source

File

REPORT_DOL-OIG_alert-memorandum-eta-needs-to-incorporate-data-analytics-capability-to_2023-09-25.pdf

Size

891,142 bytes

SHA-256

f5f53638ce65930e26e07358d26b7844fb9838948054c9324331cc07b762e7b2

Our copy

REPORT_DOL-OIG_alert-memorandum-eta-needs-to-incorporate-data-analytics-capability-to_2023-09-25.pdf

Original

www.oversight.gov