Proposed unemployment insurance integrity performance measure

UIPL17-11ACC.pdf (https://www.dol.gov/sites/dolgov/files/ETA/advisories/UIPL/2011/UIPL17-11ACC.pdf)

                                                          CLASSIFICATION
   EMPLOYMENT AND TRAINING ADMINISTRATION                 UI
               ADVISORY SYSTEM                            CORRESPONDENCE SYMBOL
                                                          OUI/DPM
          U.S. DEPARTMENT OF LABOR
                                                          DATE
             Washington, D.C. 20210
                                                          June 9, 2011



ADVISORY:          UNEMPLOYMENT INSURANCE PROGRAM LETTER NO. 17-11

TO:                STATE WORKFORCE AGENCIES

FROM:              JANE OATES /s/
                   Assistant Secretary

SUBJECT:           Proposed Performance Measure for Unemployment Insurance
                   (UI) Integrity

1. Purpose. To describe and solicit comments on a proposed UI Performs Core Measure for UI
Integrity.

2. References. Improper Payments Information Act of 2002 (31 U.S.C. 3321 note); Executive
Order (E. O.) 13520, 74 Fed. Reg. 62,201 (November 20, 2009); Improper Payments Elimination
and Recovery Act of 2010 (31 U.S.C. 3321 note); U. S. Department of Labor Fiscal Year (FY)
2010 Agency Financial Report (AFR) (November 15, 2010),
http://www.dol.gov/_sec/media/reports/annual2010/2010annualreport.pdf.

3. Background. As required by the Improper Payments Information Act of 2002 (IPIA), the
U. S. Department of Labor (Department) estimates and reports the annual rates and amounts of
improper payments in the UI program. These estimates are based on the results of the Benefit
Accuracy Measurement (BAM) survey of paid and denied UI claims in the State UI,
Unemployment Compensation for Federal Employees (UCFE), and Unemployment
Compensation for Ex-Service Members (UCX) programs.

On November 20, 2009, President Obama signed E. O. 13520, “Reducing Improper Payments.”
The purpose of this E. O. is “to reduce improper payments by intensifying efforts to eliminate
payment error, waste, fraud, and abuse in the major programs administered by the Federal
Government, while continuing to ensure that Federal programs serve and provide access to their
intended beneficiaries.” The order implements a comprehensive set of policies, including
“coordinated Federal, State, and local government action in identifying and eliminating improper
payments.”

E. O. 13520, in part, works in tandem with the Improper Payments Elimination and Recovery
Act of 2010 (IPERA), which amended the IPIA. IPERA established several criteria that Federal
agencies must meet in order to be in compliance with the law, including the requirement that the
agency has reported an improper payment rate of less than 10 percent for each program and


 RESCISSIONS                                             EXPIRATION DATE
 None                                                    Continuing

activity to which the IPIA and IPERA apply. For the 2010 IPIA reporting period, the
Department reported an improper payment rate of 11.2 percent (10.6 percent overpayment rate
and 0.6 percent underpayment rate) in its FY 2010 AFR, p. 179.

Because the UI program is not in compliance with IPERA, the Department has developed a
Strategic Plan that includes several initiatives to address the root causes of UI improper
payments and bring the program into compliance. One element of the Strategic Plan is the
development of a new performance measure under the UI Performs performance management
system to address the leading cause of UI improper payments -- claimants who return to work
and who continue to claim and collect UI benefits. In FY 2010, these benefit year earnings
(BYE) overpayments totaled $2.1 billion, which represented 3.4 percent of total UI benefits paid
and nearly 30 percent of all overpayments.

Initially, this measure would be defined in terms of the percentage of BYE overpayments due to
fraud, estimated by the BAM survey. However, the Department plans to request authorization
from the Office of Management and Budget under the Paperwork Reduction Act to collect data
on the distribution of BYE overpayments established by state Benefit Payment Control activities
by the number of weeks that were overpaid. Once this data collection has been implemented, the
performance measure would be defined in terms of reducing BYE overpayments of five weeks or
more.

4. Proposed Measure and Acceptable Level of Performance (ALP).

Measure Definition: Percentage of UI benefits overpaid due to BYE fraud.

Benchmark Period: The benchmark would be the state’s average BYE fraud rate for the
previous three calendar years (CY). This would result in more accurate estimates due to the
larger samples and would smooth out year-to-year variances due to macroeconomic factors. A
table of BYE fraud rates by state for the CY 2008 - 2010 baseline period is provided in
Attachment A.

Acceptable Level of Performance (ALP): Reduce the percentage of UI benefits overpaid due
to BYE fraud from the baseline level by 35 percent at the end of the first year of implementation
and by 50 percent by the end of the second year of implementation. Performance would be
measured by the BYE fraud rate, estimated by the BAM survey, rather than the amount overpaid,
to control for year-to-year changes in benefit outlays. A table of ALP targets by state is provided
in Attachment A.

Calculation: The measure would be calculated from BAM data using the following data
elements:

      Dollar Amount of Key Week Error (BAM data element ei1 - defines the overpayment
       amount for the key (sampled) week of benefits)
      Key Week Action (BAM data element ei2 = 10 - defines fraud overpayments)
      Error Cause (BAM data element ei3 = 100-119 and 150-159 - defines BYE
       overpayments)
      Original Amount Paid (BAM data element f13 - defines amount paid to claimant in key
       week)

The rate (expressed as a percentage) would be the weighted (by the number of paid UI weeks in
the BAM survey population) estimate of:

Amount overpaid due to BYE Fraud
———————————————— X 100
Amount of UI benefits paid

Performance Period: In order to include the results in the State Quality Service Plan (SQSP),
the performance period would be based on BAM data for the CY. Per the BAM State
Operations Handbook (ET Handbook 395, 5th edition), 98 percent of BAM cases must be
completed by 120 days after the end of the CY (April 30, or April 29 in leap years). The first
measurement period would be CY 2012.

Sampling Error: Because this measure would be based on sample data, the sampling error
would be taken into account in determining whether a state meets its ALP.

Failure to Meet the ALP: States failing to meet the ALP would be expected to develop a
Corrective Action Plan as part of the SQSP. The first SQSP performance period would be CY
2012.

Future ALPs: As states reduce their BYE fraud overpayment rates, percentage reductions for
performance periods after CY 2013 would be adjusted to reflect ALPs that are cost effective for
state agencies to achieve.

Data Collection Costs: Because the performance measure would initially use data collected
through the BAM survey, there would be no data collection start-up costs for this performance
measure.

5. Action Requested. State Administrators are requested to:

      provide the above information to appropriate staff; and
      provide comments within 30 days from the date of this advisory on the proposed
       definition, ALP, and method of calculation. Comments should be sent to Andrew Spisak
       by e-mail ([email protected]), fax (202-693-3975), or mail:

       Mr. Andrew Spisak
       U. S. Department of Labor
       ETA / Office of Unemployment Insurance
       200 Constitution Avenue, NW, Room S-4231
       Washington, DC 20210

6. Inquiries. All inquiries should be directed to the appropriate Employment and Training
Administration Regional Administrator.





7. Attachments.

Attachment A - State baseline performance and ALP targets.
Attachment B - Benefit Year Earnings Performance Measure Technical Analysis.

UIPL17-11ATT-A.pdf (https://www.dol.gov/sites/dolgov/files/ETA/advisories/UIPL/2011/UIPL17-11ATT-A.pdf)

                                                                      Attachment A
State Benefit Year Earnings Fraud Rates and Acceptable Levels of Performance
                       Baseline Period CY 2008 to 2010

 ST     BYE Fraud       BYE Fraud        Percent of    35% Rate    50% Rate
            Rate         Amount          Total BYE     Reduction   Reduction
       (Percent of UI   Overpaid*          Fraud        CY 2012     CY 2013
       Benefits Paid)                   Overpayments
 AK        1.75%         $2,941,461        0.32%        1.14%        0.88%
 AL        1.72%         $8,472,575        0.91%        1.12%        0.86%
 AR        5.19%        $25,174,572        2.71%        3.37%        2.60%
 AZ        2.96%        $20,621,815        2.22%        1.92%        1.48%
 CA        2.02%        $180,019,495       19.38%       1.31%        1.01%
 CO        0.64%         $5,167,189        0.56%        0.42%        0.32%
 CT        1.38%        $14,086,765        1.52%        0.90%        0.69%
 DC        4.47%         $7,611,013        0.82%        2.91%        2.24%
 DE        2.43%         $3,880,367        0.42%        1.58%        1.22%
 FL        0.78%        $18,734,665        2.02%        0.51%        0.39%
 GA        0.80%         $9,704,799        1.04%        0.52%        0.40%
 HI        0.10%          $335,895         0.04%        0.07%        0.05%
 IA        0.06%          $388,641         0.04%        0.04%        0.03%
 ID        1.77%         $5,306,623        0.57%        1.15%        0.89%
 IL        1.07%        $35,695,867        3.84%        0.70%        0.54%
 IN        1.51%        $19,957,187        2.15%        0.98%        0.76%
 KS        1.51%         $8,065,649        0.87%        0.98%        0.76%
 KY        0.67%         $4,941,035        0.53%        0.44%        0.34%
 LA        7.27%        $26,891,993        2.89%        4.73%        3.64%
 MA        1.26%        $26,928,388        2.90%        0.82%        0.63%
 MD        2.05%        $18,253,682        1.96%        1.33%        1.03%
 ME        0.72%         $1,486,522        0.16%        0.47%        0.36%
 MI        0.74%        $19,443,703        2.09%        0.48%        0.37%
 MN        1.18%        $13,414,613        1.44%        0.77%        0.59%
 MO        2.09%        $16,879,423        1.82%        1.36%        1.05%
 MS        3.34%         $8,371,203        0.90%        2.17%        1.67%
 MT        0.50%          $660,194         0.07%        0.33%        0.25%
 NC        1.79%        $36,152,567        3.89%        1.16%        0.90%
 ND        0.93%          $643,910         0.07%        0.60%        0.47%
 NE        1.18%         $1,853,650        0.20%        0.77%        0.59%
 NH        1.15%         $2,126,951        0.23%        0.75%        0.58%
 NJ        0.72%        $21,231,488        2.29%        0.47%        0.36%
 NM        3.09%         $8,337,030        0.90%        2.01%        1.55%
 NV        4.77%        $38,680,880        4.16%        3.10%        2.39%
 NY        1.40%        $54,666,166        5.88%        0.91%        0.70%
 OH        1.99%        $37,463,596        4.03%        1.29%        1.00%
 OK        1.05%         $3,996,826        0.43%        0.68%        0.53%
 OR        1.91%        $21,596,394        2.32%        1.24%        0.96%
 PA        2.06%        $78,866,327        8.49%        1.34%        1.03%

                                       A-1

PR        3.40%       $9,261,672           1.00%       2.21%   1.70%
RI        2.09%       $6,794,132           0.73%       1.36%   1.05%
SC        2.41%       $16,003,760          1.72%       1.57%   1.21%
SD        0.61%         $280,864           0.03%       0.40%   0.31%
TN        1.89%       $13,292,396          1.43%       1.23%   0.95%
TX        0.52%       $14,177,209          1.53%       0.34%   0.26%
UT        0.88%        $3,228,108          0.35%       0.57%   0.44%
VA        1.74%       $14,300,604          1.54%       1.13%   0.87%
VT        1.13%        $1,656,272          0.18%       0.73%   0.57%
WA        0.96%       $16,673,952          1.79%       0.62%   0.48%
WI        1.51%       $20,589,945          2.22%       0.98%   0.76%
WV        0.60%        $1,487,164          0.16%       0.39%   0.30%
WY        2.19%        $2,261,733          0.24%       1.42%   1.10%

US        1.58%       $929,058,930         100.0%      1.03%   0.79%

* Annual average amount overpaid, CY 2008 – CY 2010.




                                     A-2

UIPL17-11ATT-B.pdf (https://www.dol.gov/sites/dolgov/files/ETA/advisories/UIPL/2011/UIPL17-11ATT-B.pdf)

                                                                                    Attachment B

             Benefit Year Earnings (BYE) Performance Measure Technical Analysis

Weeks Overpaid vs. Fraud

The proposed performance measure would target those BYE errors that are considered the most
serious. To meet this objective, the BYE performance measure can be defined either in terms of
the number of weeks overpaid or by the proportion of BYE overpayments classified as fraud.

The Benefit Accuracy Measurement (BAM) survey samples paid and denied UI claims each
week and conducts a thorough audit to determine the propriety of the decision to pay or deny
benefits. Although BAM improper payment estimates refer to the single paid week that was
selected for audit (referred to as the key week), BAM captures limited information concerning
overpayments established outside the key week. BAM data for CY 2008 to CY 2010 show that,
nationally, a little over half of all BYE errors are for five or more weeks of benefits that were
improperly paid.

                                Distribution of Number of Weeks Overpaid
                                    BYE Overpayments - CY 2008-2010


                                                  3 Weeks
                                                                     4 Weeks
                          2 Weeks                   7.9%
                                                                       5.8%
                            9.9%



    1 Week
     19.4%




                                                                                    5+ Weeks
               < 1 Week
                                                                                      51.7%
                 5.3%




In deciding between weeks overpaid versus fraud to define the measure, both data quality and
data availability need to be considered. Currently, the only source of data available for this
measure is the BAM survey. In terms of data quality, the fraud definition is preferable to weeks
overpaid due to the following significant limitations of the BAM data.

      BAM sample sizes at the state level are too small to produce accurate estimates of the
       distribution of BYE improper payments by the number of weeks overpaid. Given the
       current sample allocation for BAM paid claims (360 cases per year in the ten smallest
       states in terms of UI weeks paid and 480 cases in all other states), the number of cases
       coded as BYE overpayments is limited, even for multiple years.
                                                  B-1

                                                                                    Attachment B

       For CY 2008 to CY 2010, cases coded as BYE overpayments ranged from 24 in Georgia
       to 181 in Louisiana. While sample sizes are adequate (although only marginally so in
       Georgia and a few other states) to estimate a binomial distribution of fraud vs. nonfraud,
       samples are inadequate in most states to estimate a distribution of weeks of overpayment
       from 1 to 26 or more.

      BAM is designed to estimate the accuracy of the single compensated week that is
       selected for the BAM audit. Although BAM documents overpayments for multiple
       weeks that are attributable to issues identified in the BAM audit in the Total
       Overpayments data element (H3 in the b_master table), these data are not suitable for this
       proposed performance measure because:

                 According to the BAM coding guidance, only overpayments officially
                  established as a result of BAM are documented in the Total Overpayments
                  data element. Accordingly, many states do not count overpayments
                  established outside the key week that were detected through state Benefit
                  Payment Control activities.

                 Issues detected on State UI, Unemployment Compensation for Federal
                  Employees (UCFE), and Unemployment Compensation for Ex-Service
                  Members (UCX) program claims may result in the establishment of
                  overpayments for benefits paid from the Extended Benefits program and
                  temporary programs such as Emergency Unemployment Compensation,
                  which are included in the Total Overpayments data element in the BAM
                  database. Because these overpayments for the episodic and temporary
                  programs are not reported separately from overpayments in the State UI,
                  UCFE, and UCX programs, the BAM Total Overpayments data element will
                  overstate State UI, UCFE, and UCX overpayments.

      In the case of overpayments with multiple causes, it is not possible to separate the
       overpayments reported in the BAM Total Overpayments data element by cause. BYE
       overpayments cannot be differentiated, for example, from overpayments attributable to
       separation, active work search, and availability issues.

Because of the limitations of the BAM survey data, the U. S. Department of Labor (Department)
would begin the process of obtaining authorization from the Office of Management and Budget
(OMB) to collect population data on the number of overpayments and amounts overpaid
established by state agencies by the number of weeks overpaid for BYE issues. However, in
order to begin focusing on and measuring progress in reducing BYE fraud, we propose using the
existing BAM data collection for this performance measure on an interim basis.

It is estimated that the development of an alternative data source would take 18 months to two
years, and include the following requirements:

      Analysis and identification of an alternative metric and data collection source;
      Employment and Training Administration and Department approval of proposed data
       collection;
                                               B-2

                                                                                   Attachment B
      Solicitation of states to participate in pilot testing;
      Development of pilot test procedures and instructions;
      Pilot testing by states;
      Analysis of pilot test results;
      Preparation of the Paperwork Reduction Act package for OMB approval;
      Preparation of final methodology and instructions for state implementation; and
      State agency programming of additional data collection.

Analysis - BYE Fraud Overpayments

The BAM methodology requires all overpayment errors to be classified as fraud or nonfraud. By
defining the performance measure in terms of fraud BYE overpayments, the objective of
reducing overpayment errors by targeting the most serious BYE violations can be met. For the
period CY 2008 to CY 2010, 54.2 percent of BYE overpayments were due to fraud, slightly
higher than the proportion of BYE overpayments that are five weeks or more (51.7 percent).
The following table shows the U. S. aggregate and ranges by state for various metrics.

BYE Fraud CY 2008-2010                  Lowest         Highest           U. S.
Average Annual Amount Overpaid         SD - $0.28    CA - $180.02       $929.06
(In millions of $)
Percentage of UI Benefits              IA - 0.06%    LA - 7.27%         1.58%
Percentage of BYE Overpayments         IA - 2.92%    MS - 92.05%        54.18%

For CY 2010 alone, the U. S. percentage and amount of BYE fraud overpayments were slightly
above the CY 2008 - CY 2010 averages.

BYE Fraud CY 2010                       Lowest*        Highest           U. S.
Amount Overpaid                        MT - $0.82    CA - $164.20      $1,038.45
(In millions of $)
Percentage of Benefits                ME - 0.04%      LA - 9.08%        1.77%
Percentage of BYE Overpayments        MT - 1.38%      MS - 97.5%        53.30%

* Iowa reported no BYE fraud overpayments in CY 2010. The state reported $9.7 million in
nonfraud BYE overpayments, which was 1.61% of UI benefits paid.

Estimated Affect on Annual Report and Operational Rates

The following chart shows the reduction in the Annual Report and Operational rates for various
levels of reduction in the BYE fraud rate, using the preliminary CY 2010 rates of 11.51 percent
(Annual Report) and 6.13 percent (Operational) as the baseline. All BYE fraud overpayments
are included in both the Annual Report and Operational rates and represent 1.77 percent of UI
benefits paid. If all states meet their reduction targets, the U.S. BYE fraud rate would decrease
by 35 percent in the first year, and both the Annual Report and Operational rates would decrease
by approximately 0.7 percentage points. After two years of implementation, the U. S. BYE
fraud rate would decrease by 50 percent, and both the Annual Report and Operational rates
would decrease by approximately 1.0 percentage point.


                                               B-3

                                                                             Attachment B

      Estimated Reduction In Annual Report and Operational Rates
                           CY 2010 Baseline
Overpayment
   Rates
         12.0%
         11.0%
         10.0%
          9.0%
          8.0%
          7.0%
          6.0%
          5.0%
                 0%          10%       20%         30%       40%       50%
                         Percent Reduction In BYE Fraud Overpayments


                      Est. Annual Report       Current Ann. Rpt.
                      Est. Operational         Current Oper.




                                             B-4