Implementation of Unemployment Insurance (UI) Benefit Accuracy Measurement (BAM) Denied Claims Accuracy (DCA) and the Inclusion of Paid Interstate Claims

U.S. DEPARTMENT OF LABOR Employment and Training Administration Washington, D. C. 20210	CLASSIFICATION UI
	CORRESPONDENCE SYMBOL
	ISSUE DATE November 23, 2000
RESCISSIONS None	EXPIRATION DATE November 30, 2001

DIRECTIVE

:

UNEMPLOYMENT INSURANCE PROGRAM LETTER NO. 10-01

TO

:

STATE EMPLOYMENT SECURITY AGENCIES

FROM

:

GRACE A. KILBANE
Administrator
Office of Workforce Security

SUBJECT

:

Implementation of Unemployment Insurance (UI) Benefit Accuracy Measurement (BAM) Denied Claims Accuracy (DCA) and the Inclusion of Paid Interstate Claims

Purpose. To provide State Employment Security Agencies (SESAs) with information on the implementation of BAM DCA and the inclusion of interstate claims in the BAM paid claims samples. Pending Office of Management and Budget (OMB) information collection approval, SESAs should plan to select DCA samples and include interstate claims in the BAM paid claims samples beginning with batch 200123 (batch ending date June 9, 2001).
References. Federal Register Notice (FRN) 52 FR 33520 (September 3, 1987); FRN 65 FR 59021 (October 3, 2000); Unemployment Insurance Program Letter (UIPL) No. 4-86 (December 20, 1985); UIPL No. 41-95 (August 24, 1995); UIPL No. 15-96 (April 2, 1996); UIPL No. 3-97 (November 20, 1996); UIPL No. 1-98 (October 20, 1997); Code of Federal Regulations (CFR), Title 20, Part 602; and Benefit Accuracy Measurement State Operations Handbook, ET Handbook No. 395.
Background. Since 1987 all SESAs, except the Virgin Islands, have been required by regulation at 20 CFR Part 602 to operate a quality control program to assess the accuracy of benefit payments in three programs: State Unemployment Insurance (UI), Unemployment Compensation for Federal Employees (UCFE), and Unemployment Compensation for Ex-Servicemembers (UCX). This program, implemented under the Department's authority at Sections 303(a)(1), (6), and (b)(2) of the Social Security Act, was initially called Benefits Quality Control (BQC), but was renamed Benefit Accuracy Measurement (BAM) in 1996.

During the public consultation process which preceded the establishment of the BQC/BAM program, several public interest groups representing employers, employees, and State government agencies proposed underlying principles to govern the program. The Department adopted these consensus principles in UIPL No. 4-86.

In the final rule establishing the BQC program, published in the Federal Register at 52 FR 33520, one of the consensus principles states that, "QC would be expanded to include the review/investigation of claims that had been denied." In addition, 20 CFR 602.2 states that:

Other elements of the QC program (e.g., interstate, extended benefits programs, benefit denials, and revenue collections) will be phased in under a schedule determined by the Department in consultation with State agencies.

In 1986-87, five States conducted a one-year pilot to measure the accuracy of determinations to deny unemployment compensation (UC) eligibility for monetary, separation, and nonseparation reasons. These States tested three sampling designs and used the BAM case investigation methodology. Although the pilot identified significant rates of error in the denial determinations that were investigated, national implementation of a program to measure the accuracy of denied claims was deferred because of resource constraints and other program priorities, such as the implementation of Benefit Timeliness and Quality and the Tax Performance System. Since the 1986-87 denied claims pilot, several groups, including organized labor, employee rights legal support groups, the Department of Labor's Office of Inspector General, and the Vice President's National Performance Review have urged the Department to measure the accuracy of determinations that deny eligibility for UC claims.

In 1995 the Performance Enhancement Work Group (PEWG), which consisted of senior SESA managers and Federal staff, recommended several changes in the way UC operational performance was measured and improved. The Department accepted most of the recommendations and has implemented them as UI PERFORMS. UIPL No. 41-95 describes in detail the UI PERFORMS performance management system. Among the PEWG recommendations with respect to BAM were: (1) the implementation of a system to measure the accuracy of determinations that deny eligibility for UC claims; (2) reductions in the BAM paid claims sample allocations; and (3) a modification of data collection methods to provide the States more flexibility in using program resources. Resources freed up by the implementation of (2) and (3) would be redirected to support UI PERFORMS continuous improvement activities, including investigating the accuracy of denied claims.

UIPL No. 15-96 described the proposed changes and solicited comments from the SESAs. According to the Attachment to the UIPL, "Measuring UI Benefit Payment Accuracy Under UI PERFORMS: Proposed Changes to Benefits QC":

Under the proposal, staff freed up because of sample reductions or changes in how verifications are conducted will be available for investigating denied claims and other UI Performs [sic] activities, including taking other performance measurements.

The implementation of the changes and the reallocation of BAM resources were reported in UIPL No. 3-97.

Because significant time had elapsed since the initial denied claims pilot, the Department conducted a new pilot to guide implementation of DCA and the inclusion of interstate claims. Five States -- Nebraska, New Jersey, South Carolina, West Virginia, and Wisconsin -- participated in this new pilot. Sampling and investigation of denied UC claims were conducted from September 1997 through September 1998. The principal objectives of the pilot were:
- to test the operational feasibility of applying the BAM paid claims investigation methodology to denied UC claims, including both intrastate and interstate claims; and
- to evaluate whether the quarterly reviews of nonmonetary determination quality (ET Handbook 301) adequately assess the accuracy of determinations that deny eligibility for UC claims for separation or nonseparation reasons or whether the accuracy of these nonmonetary determinations can be measured only through the more comprehensive BAM fact-finding process.
The final DCA pilot evaluation report was issued in May 1999 and is available on the ETA Office of Workforce Security (OWS) Web site at: http://workforcesecurity.doleta.gov/unemploy/pdf/toc.pdf.

The five States that participated in the new pilot demonstrated that the BAM case investigation methodology was easy to implement for both intrastate and interstate claims, was successful in detecting erroneously denied claims, and identified valuable information on the cause, responsible party, point of detection, and prior agency action for improperly denied claims that SESAs could use to improve UC operations. In general, the new pilot confirmed the results of the 1986-87 pilot that significant percentages of UC claims in all three eligibility areas (monetary, separation, and nonseparation) were incorrectly denied, although accuracy rates varied both among States and among the three eligibility areas.

The new pilot also demonstrated that the results of the quarterly evaluations of nonmonetary determination quality are not a reliable predictor of the accuracy of determinations that deny eligibility for UC. Because DCA focuses on whether the outcome was proper, and the quarterly quality review focuses on the degree to which proper adjudication procedures were followed, it is possible to pass one and fail the other. The quarterly quality review evaluates several elements of the adjudication rocess based on agency records. DCA is more narrowly focused to determine the accuracy of the denial, based on both agency records and new fact-finding. DCA is not designed to be a substitute for the quarterly quality review. Rather the two ystems complement each other through their different emphases and contribute important but distinct information within UI PERFORMS.

The Department, as part of its continuing effort to reduce paperwork and respondent burden, conducts a preclearance consultation program to provide the general public and Federal agencies with an opportunity to comment on proposed and/or continuing collections of information in accordance with the Paperwork Reduction Act of 1995 (PRA95) [44 U.S.C. 3506(c)(2)(A)]. ETA is soliciting comments concerning the proposed new collection of information on the accuracy of determinations that deny eligibility for UC and the inclusion of interstate claims in the BAM paid claims samples (65 FR 59021). The comment period extends through December 4, 2000. ETA is seeking OMB approval under the PRA95 to establish Denied Claims Accuracy (DCA) as a component of the quality control (QC) program in the Federal-State unemployment insurance system (20 CFR Part 602) and include interstate claims in the BAM paid claims samples. A copy of the Federal Register notice is available on the OWS Web site, http://workforcesecurity.doleta.gov/unemploy/pdf/dca_frn.pdf. The proposed DCA data collection instrument is available at http://workforcesecurity.doleta.gov/unemploy/pdf/dca_dci.pdf.
DCA Methodology.
1. Sample Design and Sample Sizes: Each week, States will select systematic random samples from three separate sampling frames constructed from the universes of UC claims for which eligibility was denied for monetary, separation, or nonseparation reasons. States will use the BAM population edit and sample selection software program, which was distributed to all SESAs in January 1998, to select the weekly samples. This software uses a systematic random sampling algorithm. The Department will distribute a table of random start numbers to use with the BAM/DCA sample selection program.
  
  All States will sample a minimum of 150 cases of each type of denial in each calendar year. The Department considers the annual sample allocation of 150 cases for each of the three types of denials to be the minimum necessary to produce a sufficient number of completed cases to estimate DCA rates with acceptable precision and a sufficient number of error cases to produce program improvement information. The proposed DCA sample allocations take into account the difficulty of obtaining information from claimants who were denied benefits and the possibility that DCA case completion rates will be less than the nearly 100 percent completion rate achieved for BAM paid claims. This will result in fewer than the allocated 150 sample cases that will be available to estimate accuracy rates and to provide information on error causes, responsibility, and other information that can be used for program improvement.
  
  Samples for BAM paid claims accuracy are unchanged from the current allocation of 360 cases per year in the ten States with the smallest UI program workloads and 480 cases per year in the other States.
2. Scope: Denied intrastate and interstate claims in the State UI, UCFE, and UCX programs will be included in DCA. In addition, the BAM paid claims sampling frames will begin to include interstate claims in the UI, UCFE, and UCX programs, effective with the implementation of DCA. Both intrastate and interstate paid claims will be included in the annual BAM paid claims samples of 360/480 cases; separate samples of paid intrastate and interstate claims will not be selected.
3. Operational Definitions of Sampling Frames: Unless otherwise stated, definitions refer to those used in ET Handbook 401, 3^rd edition. ETA report cell references are those used in ET Handbook 402, 4^th edition.
4. Frequency and Timing: State agencies will create a sampling frame file each week. The sampling frame includes all determinations that deny eligibility for UC issued during the period 12:00 a.m. Sunday to11:59 p.m. Saturday. The date of the determination is the date printed on the determination notice. If no notice is issued, it is the date that the denial action was entered into the agency's record system or that a permanent stop payment order was issued.
  
  In the 1997-98 DCA pilot, several claims for UC were sampled which were initially denied for insufficient wages but were subsequently determined to be monetarily eligible upon the addition of wages from out-of-State employers (combined wage claims) or Federal wages (UCFE and/or UCX programs). In order to allow time for States to request and receive UCFE wages and wage transfers from other States, the sampling frame for monetary denials will be constructed two weeks after the week ending date of the initial claim. For example, the sampling frame for batch 200123 (June 3 - 9, 2001) will consist of new initial and transitional claims filed on or before May 26 for which the most recent determination issued between May 20 and June 9 denies monetary eligibility.
5. Case Investigation: DCA will be conducted according to the methods and procedures documented in ET Handbook 395; case investigation procedures are described in detail in chapter VI. Investigators will review agency records and contact the claimant, employer(s), and all other relevant parties to verify information in agency records or obtain additional information pertinent to the determination that denies eligibility. Unlike the investigation of paid claims, in which all prior determinations affecting claimant eligibility for the compensated week selected for the sample are evaluated, the investigation of denied claims will be limited to the issue upon which the denial determination was based. For example, if a continued week claim is denied because the agency determined the claimant was not available for work, only the availability issue will be investigated. The monetary, separation and any other nonmonetary determinations which could have affected eligibility for the week claimed will not be investigated. Like the investigation of paid claims, States have the flexibility to conduct the investigation of denied claims by in-person interview, telephone, mail, e-mail or fax, as they deem appropriate.
6. Data Recording and Reports: States will record the results of their investigations using a standard data collection instrument and suite of software supplied by the Department. The Department will collect this information from the State databases and store the data in a database in the National Office in Washington, D.C. DCA is a UI PERFORMS Tier II measure; the Department will publish accuracy rates for each of the three types of denied claims for each State in the UI PERFORMS Annual Report.
Staff Resources. When BAM paid claims sample sizes were reduced to the 480/360 levels and flexibility in verification methods was permitted in accordance with PEWG recommendations in 1996, it was with the understanding that State staff freed up by these changes would be sufficient to conduct BAM denied claims accuracy when implemented and UI PERFORMS continuous improvement activities. Two full-time equivalent (FTE) staff positions allocated to continuous improvement activities are estimated to be required to support DCA in each State. States will decide how to allocate/apportion BAM paid and denied claims sample cases among staff.
ADP Tasks and Resources.
1. State Programming Tasks: UIPL No. 1-98 included the documentation for the revised BAM/DCA population edit and sample selection (COBOL) program, which was distributed to all SESAs in January 1998, and the specifications for programming required to construct the DCA sampling frame files, for which the SESAs are responsible. States can download a copy of UIPL 1-98 from the Document Management System on the OWS Web site at http://workforcesecurity.doleta.gov/dmstree/index.html.
  
  Several States have completed the programming and testing required to implement DCA sampling. While work related to the implementation of DCA cannot be required prior to OMB approval, it would be prudent for States that have not yet completed the programming to schedule the ADP tasks needed to implement DCA and to include interstate claims in the sampling frame files for BAM paid claims so that they are completed by the DCA implementation date. The BAM/DCA COBOL program will be modified to include paid interstate claims in the BAM samples and to reflect revisions that were identified in the DCA pilot, such as the two-week lag in sampling monetary denials. States will have to recompile the revised COBOL program on their ADP systems prior to the implementation of DCA and the inclusion of interstate claims in BAM paid claims samples.
  
  States that intend to use the automatic case conversion option of the DCA software are referred to pages 34-37 and Attachments B and C of UIPL No. 1-98 for specifications of the rec1.dat file, which consists of the BAM/DCA sample cases and is created by the State, and the sfsum.dat file, which consists of sample and populations characteristics and is created by the BAM/DCA COBOL program. These two files must be downloaded from the State's ADP system to the State's Sun Ultra 10 computer. States are responsible for the programming to create the rec1.dat file, which must be formatted exactly as described in UIPL No. 1-98. Each sample case consists of 10 fixed-length records (lines). Positions 1 to 79 of each record must be coded; positions that do not contain data should be coded as a space (octal '040', hex '20'). The system-generated new line character (octal '012', hex '0a') terminates each record.
2. DCA Applications Software: For the DCA pilot, the Department developed several DCA software modules, which are installed on each State's Sun Ultra 10 computer. Although this software is functional, the Department will release a revised version of the software in advance of the national implementation of DCA. The revised software will:
  - include Change Control Information and Unlock Cases modules, which were not available for the DCA pilot;
  - modify the edit criteria for several data elements in the Update Cases, Reopen Cases, and Stamp modules to reflect the experiences of the pilot States in coding DCA case investigation information; and
  - modify the Case Conversion module so that States do not have to select samples for all three types of denials for a batch.
  The software released prior to implementation of DCA will not include any output reports, due to the considerable time and resources required to develop them. Output reports will be included in a future DCA software release. The National Office will provide SESAs with Informix ACE report programs and instructions to produce a case time lapse report and a report that includes the accuracy rates for each of the three types of denials included in DCA - monetary, separation, and nonseparation. Accuracy rates will be expressed as the percentage of completed cases (BAM/DCA supervisor sign-off) which the DCA investigation determined were proper denials.
Training and Handbooks. The Department will conduct training for State staff prior to national implementation of DCA and will issue information on the times, locations, and content of the training in advance of the sessions. The Department will also revise the Benefit Accuracy Measurement State Operations Handbook, ET Handbook No. 395, prior to implementation.
Action Required. State Administrators are requested to:
1. provide copies of these instructions to the appropriate staff; and
2. ensure that, contingent on approval by OMB, all required ADP tasks described in this directive and UIPL No. 1-98 are completed prior to DCA implementation.
Inquiries. Questions should be directed to the appropriate ETA Regional Office.

DIRECTIVE	:	UNEMPLOYMENT INSURANCE PROGRAM LETTER NO. 10-01

TO	:	STATE EMPLOYMENT SECURITY AGENCIES

FROM	:	GRACE A. KILBANE Administrator Office of Workforce Security

SUBJECT	:	Implementation of Unemployment Insurance (UI) Benefit Accuracy Measurement (BAM) Denied Claims Accuracy (DCA) and the Inclusion of Paid Interstate Claims