National Accuracy Clearinghouse (NAC) Evaluation

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

National Accuracy Clearinghouse

(NAC) Evaluation

Final Report

Revised October, 2015

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Contents

Acknowledgements ....................................................................................................................................... 4

Executive Summary ....................................................................................................................................... 5

Background / Project Overview .................................................................................................................... 9

The NAC Evaluation ..................................................................................................................................... 12

NAC Business Rules and Processes ............................................................................................................. 15

Key Research Questions .............................................................................................................................. 20

Has the NAC resulted in a reduction in dual participation? ................................................................... 20

How effective have states been in utilizing the NAC to prevent dual participation? ............................. 23

How does the NAC compare to the use of PARIS? ................................................................................. 25

What is the NAC’s Return on Investment? ............................................................................................. 30

Potential Impact of Expansion ................................................................................................................ 35

Recommendations ...................................................................................................................................... 39

General Recommendations for All States ............................................................................................... 39

Expansion-Specific Recommendations ................................................................................................... 41

Opportunities for Improvement ............................................................................................................. 43

Glossary ....................................................................................................................................................... 45

Appendix ..................................................................................................................................................... 47

A – Common Business Rules ................................................................................................................... 47

B – Methodology – Percentage of Matches that Became Collisions ...................................................... 50

C – Big Bang Statistics Organized by Top 5, 6+ and SSN-only ................................................................. 51

D – Process for Using the NAC for Prevention of Dual Participation as of May 31, 2015 ...................... 52

E – Matching State Process for Responding to Request from Initiating State as of May 31, 2015 ........ 55

F – Results of State Survey for Cost/Savings Analysis ............................................................................. 56

Ongoing Costs ..................................................................................................................................... 56

Start-up Costs ...................................................................................................................................... 57

G – Percentage of Household Size Methodology ................................................................................... 58

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

List of Tables

1. Dual Participation as a Percentage of SNAP Participants, May 2014

2. Project Timeline

3. Summary of State Business Processes

4. Match Code Key

5. Match Code Combination Strength Rankings

6. Treatment of Match Code Combinations by State

7. Comparison of Dual Participation, Pre-Pilot and Pilot Periods

8. Dual Participation as a Percentage of SNAP Participants, May 2014 and May 2015

9. NPD Matches at Big Bang

10. Match Code Validity, Combinations without Exact SSN Match

11. PARIS-NAC Comparison

12. Florida Collision Volume as of June 2, 2014

13. Spell of Dual Participation from First Month of Benefit Overlap through June 2014 (Florida)

14. Percentage of Entries Remaining Dual Participants in Succeeding Months, Pre-Pilot and Pilot

Comparison

15. NAC Net Impact

16. NAC Start-up Costs

17. Comparison of Match Distribution, PARIS and NAC

18. Estimated Annual Program Savings, Nationwide NAC Implementation

19. Percentage of Dual Participants Residing in Households in which all Members are Not Dual

Participants

20. Explanation for SSN-Only Matches Identified at the Big Bang

21. Claims Data on Cases with Dual Participants Identified at the Big Bang

List of Figures

1. Percentage Reduction in Entries and Continuations, Early and Late Stages of Pilot

2. Percentage of Top 5 Matches the Became Collisions, August 2014 – May 2015

3. PARIS Timeline Scenario

4. NAC Timeline Scenario

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Acknowledgements

The efforts of staff in the NAC pilot states have been invaluable to the evaluation. The state coordinators

– Ann York (preceded by Johnnie Cox) in Alabama, Lynn Rossow in Florida, Sandra Frederick in Georgia,

Mike Morris in Louisiana, and Joel Savell in Mississippi – responded to numerous requests to support the

evaluation over two-plus years and provided insights and constructive comments throughout. Many other

state staff assisted in data collection and other evaluation activities, including Brandon Early (AL), Carmen

Young, Sonya Ward (GA), J Funderburk (LA), Karen Powell (MS and FNS), and Tim Meeks (MS).

Their contributions to the project are truly appreciated.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Executive Summary

The Supplemental Nutrition Assistance Program (SNAP) is one of the largest components of the social

safety net in the United States – providing more than $69 billion in benefits in fiscal year 2014. Ensuring

that program participants are not receiving assistance in two states at once – also referred to as dual

participation – is an important factor in maintaining program integrity and public support for SNAP.

Throughout the history of the program, states have been limited in their ability to access information on

applicants’ eligibility in other states, making prevention of dual participation difficult. Furthermore, the

processes employed by states to communicate with other SNAP agencies regarding an applicant or

recipients’ eligibility status have been inefficient, and can result in a less than optimal level of customer

service.

The Mississippi Department of Human Services contracted with Public Consulting Group, Inc. (PCG) to

evaluate the National Accuracy Clearinghouse (NAC) — a technology-based solution designed to curb

interstate dual participation in SNAP through a shared database of eligibility information that is updated

daily. The software for the NAC was developed by LexisNexis, also under contract with the State of

Mississippi. From May 2013 to August 2015, PCG worked with five pilot states — Mississippi, Louisiana,

Florida, Alabama, and Georgia – to assess the technical capacity of the NAC, states’ success in utilizing the

tool and implementing the accompanying business rules, and the cost savings — if any — associated with

adoption of the NAC solution.

This final report is the fifth in a series of documents produced by PCG as part of the NAC evaluation, and

integrates key findings from previous reports. The timeline below documents key project phases and

summarizes evaluation reports to date.

The evaluation focused on four central research questions:

1. Has the NAC resulted in a reduction in dual SNAP participation?

2. How effective have states been in utilizing the NAC to prevent dual SNAP participation?

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

3. How does the NAC compare to the use of PARIS?

4. What is the NAC’s return on investment?

Impact on Dual Participation. Comparisons of dual participation prevalence before and after the pilot

show a reduction occurred in all five pilot states, but with large variations in magnitude. Alabama and

Mississippi have had the largest declines (81 percent); Louisiana has also had a significant reduction (71

percent) in dual participation. Florida and Georgia have experienced a (relatively) small reduction (27

percent).

Effectiveness in Utilization of the NAC. The ability of pilot states to utilize NAC data to prevent dual

participation from occurring has varied greatly. Alabama and Mississippi’s prevention efforts have been

extremely successful, with consistently less than 10 percent of possible instances resulting in dual

participation. These percentages translate into the reductions in prevalence the states have realized.

While Louisiana’s efforts have not been as consistent, less than 20 percent of matches on average turn

into dual participation. Georgia and Florida have had lower levels of success at preventing dual

participation with 30 – 45 percent of matches resulting in dual participation.

Comparison of NAC and PARIS. The Public Assistance Reporting and Information System (PARIS)

presents states with some tools to reduce SNAP dual participation; however, the NAC is a more robust

and effective means for curbing SNAP dual participation. The advantages and limitation of PARIS were

identified by PCG as a part of site visits that occurred during the pre-pilot phase and are also documented

in previous reports by the General Accounting Office and Health Systems Research, Inc.

Specifically, the

NAC surpasses PARIS in the capacity to support the prevention of dual participation, supports constructive

interstate communication, allows for easier identification of “false positives,” and identifies individuals

that would not be flagged by matching only on Social Security Numbers.

Return on Investment. The net impact of the NAC during the pilot phase totaled approximately $5.6

million in SNAP overpayment avoidance (100% federal dollars). This estimate is conservative, as it only

focuses on the impact of prevention of dual participation and not the early detection of dual participation

that the NAC can also support. In addition, the estimate assumes that an individual will remain eligible in

one of the two states involved, and the calculation uses only matches that are almost certain to represent

actual dual participation. Most importantly, the estimate reflects the business processes that were in

place during the pilot. As states modify their approaches and improve system integration, the degree that

they are able to stop dual participation before it begins is expected to increase. Estimates of the savings

realized if the NAC were implemented nationwide average more than $114 million annually – less than

two-tenths of one percent of total SNAP benefits issued annually, but a significant amount nonetheless.

Recommendations

The five NAC pilot states have implemented the tool in significantly differently ways, and have realized

different levels of success. Those that have achieved superior outcomes provide a set of best practices

that should be considered as use of the NAC continues in the current states and as expansion beyond the

pilot is explored. Furthermore, the pilot states have learned lessons that should be heeded by any state–

current or future–intending to use the NAC. The best practices, lessons learned, and opportunities for

improvement may be identified under three general categories: technical recommendations, business

processes, and staff utilization.

GAO-01-935, “PARIS Project Can Help States Reduce Improper Benefit Payments.” September 2001; and “Evaluation to Determine the

Effectiveness of the Public Assistance Reporting and Information System Final Report.” Health Systems Research, Inc., June 30, 2007.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Technical Recommendations

 States should automate to the greatest extent possible. This includes integration with state

eligibility systems, use of web services for state-NAC interfaces, automated emails requesting case

action, and system edits to ensure staff take appropriate action prior to case authorization.

 Checking on individuals being added to an open SNAP case is an important component of

prevention. Data suggests that nearly half of dual participation falls into this category, and the

procedures for inquiring about benefits in another state that apply to new applicants may not be

followed in these instances. The integration of the NAC with state eligibility systems should

include automated look-ups that occur before the new household member is added to an existing

case.

 States should ensure that IP address issues do not limit access to the NAC. Some state staff,

especially those working from home, experienced difficulty accessing the NAC. These issues

occurred because not all IP addresses were accounted for, and users were denied access by the

NAC’s firewall. A thorough accounting of all potential users and their IP addresses is an important

step in NAC implementation planning.

 “Social Security Number-only” matches should be treated differently. When the only matching data

element between state data is a Social Security Number, it is likely that dual participation has not

occurred. These matches often occur because one digit of the SSN has been entered incorrectly

in the state eligibility system. In these situations, automation in place for other types of matches

is not advisable, and states should confirm the accuracy of data entry before contacting another

state or the client.

Business Processes

 States should have a robust process for addressing dual participation when prevention efforts fail.

In some instances, the business processes implemented by pilot states or worker error result in

dual participation occurring despite the availability of the NAC. The establishment of an internal

state work group responsible for follow-up and the automation of reports supports early

detection and reduction in the length of dual participation that does occur when dual

participation is not prevented.

 Additional standardization of business processes would support improved outcomes. These include

consistent treatment across states of different match types, timely submission of contributory

files, and common naming conventions for state email addresses.

 The requirement to notify a state of case closure in all instances should be reconsidered. The initial

business rules limited the number of email exchanges required between states to support timely

case closure and removal of individuals. Due to inconsistency in states’ abilities to meet required

timeframes for case action, these rules were revised. As states improve their capacity to utilize

NAC data, the revision should be reconsidered.

 States should explore a wider use of “passive matches.” The NAC notifies states when a query of

the system has been conducted by another consortium member. Although these notifications

were not used to a large degree during the pilot, they present an opportunity for additional

streamlining of communication.

Staff Utilization

 Separate staffing models are needed for the initial match and ongoing operations. When states join

the NAC consortium, the initial matching process will identify many instances of apparent dual

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

participation. States should dedicate staff to the effort to address these situations, which includes

the establishment of overpayment claims.

 Comprehensive front line staff training is essential to successful implementation. Regardless of the

specific processes a state implements to access and use NAC data, front-line staff must have an

understanding of their responsibilities and the information needed to communicate effectively

with other states.

In conclusion, the evaluation found that the National Accuracy Clearinghouse has supported a decrease

in dual SNAP participation in the five pilot states and gives states the capacity to prevent dual participation

to an extent not available through the PARIS match. A strong argument can be made that the pilot savings

and the NAC’s potential warrant continuation and expansion of the project.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Background / Project Overview

The Supplemental Nutrition Assistance Program (SNAP) is a key component of the social safety net in the

United States, and ensuring that program participants do not receive benefits in more than one state in a

single month is an important component of program integrity and maintaining public support for the

program.

When someone does receive SNAP in two or more states in the same month or in two or more

households within the same State

, it is referred to as dual participation.

Individuals working at all levels of the SNAP program have long suspected that dual participation could be

more effectively limited through an improved process for data sharing between states. SNAP caseworkers

also express frustration with the time required to communicate with other states to verify eligibility status,

which can impact applicants’ ability to receive benefits quickly.

Several southern states tested the concept of data sharing through the “buddy state” model as early as

2008 as a result of lessons learned operating D-SNAP programs following Hurricane Katrina. The

establishment of the Partnership Fund for Program Integrity Innovation by the Office of Management and

Budget (OMB) in 2010 created the opportunity for funding a more comprehensive solution.

The following year, OMB awarded the United States Department of Agriculture (USDA) Food and Nutrition

Service (FNS) $2.5 million with the goal of reducing improper payments that occur due to dual

participation in SNAP. This grant funded the development of a searchable database – the National

Accuracy Clearinghouse (NAC) – to support near real-time sharing of eligibility information. Subsequently,

Mississippi was awarded the funding to lead the project on behalf of a consortium of contiguous states

(also including Alabama, Florida, Georgia, and Louisiana).

Scope of the Problem

With limited information on the scale of dual participation, the size of the problem that the NAC was

designed to address was not entirely clear at the outset of the project. Table 1 demonstrates one approach

for quantifying the scale of dual SNAP

participation prior to the implementation of

the NAC. Here, the number of dual

participants

in May 2014 is divided by the

total number of active SNAP recipients in the

same month (using data only made available

after the development of the database). Note

that the dual participant statistics only

represent the instances identified with the

other four pilot states and do not include data

from the other 45 states, Washington D.C.,

and the territories.

SNAP regulations at 7 CFR 273.3 (a) dictate that “No individual may participate as a member of more than one

household or in more than one project area, in any month…” An exception exists if the individual is a resident of a

shelter for battered women and children and was a member of a household containing the person who had abused

him or her.

State eligibility systems generally support the ability to identify dual participation within state borders (intrastate);

the NAC was conceived to address interstate dual participation.

Count of dual participants based on “Top 5” match code combinations. Reference “NAC Business Rules and

Processes” section for additional information.

Table 1

Dual Participation as a Percentage of SNAP Participants

May 2014

Eligible

individuals

Dual

participants

Alabama

898,301

1534

0.171%

Florida

3,487,797

3534

0.101%

Georgia

1,847,395

3464

0.188%

Louisiana

866,941

755

0.087%

Mississippi

650,853

789

0.121%

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

The statistics reflect a relatively low occurrence of dual participation – ranging from less than one-tenth

of one percent of Louisiana’s eligible individuals in May 2014 to just below two-tenths of one percent of

Georgia’s. However, in a program as large as SNAP - with total allotments exceeding $69 billion in FY 14 -

even small percentages of benefits issued in error translate into a significant improper expenditure of

taxpayer dollars.

The NAC Solution

The premise of the NAC is simple – states contribute daily files of their active SNAP participants in a

common format to a centralized database. The states then submit information requests to the database

on program applicants, and the NAC looks for overlapping information on a range of data points, such as

Social Security Numbers, names, and dates of birth (DOB), to determine if the individual is already a SNAP

recipient in another state.

The following example demonstrates how the NAC can

prevent dual participation (and what happens when it

does not):

 John Doe is receiving SNAP in Alabama. He moves

out of state without notifying his caseworker, and

his SNAP case remains active in Alabama.

 John applies for SNAP in Louisiana. When his

application is registered there, an inquiry to the

NAC is completed to determine if he is receiving

SNAP in any of the other four pilot states. A match

(see box) is returned that confirms he is an active

SNAP recipient in Alabama.

 John’s caseworker in Louisiana receives

the information from the NAC as part of the eligibility determination process. SNAP benefits

in Louisiana are not approved until case closure in Alabama is confirmed, and dual

participation is prevented.

John’s caseworker in Louisiana ignores the information in the match and approves SNAP

benefits without notifying Alabama. Dual participation (a collision) occurs.

The NAC provides participating states with data on both matches (to support the prevention of dual

participation) and collisions (to end dual participation as soon as possible when it does occur).

The Matching Process

When a state submits an individual to the NAC to perform a search

, the database first determines if a

LexID is associated with the input information. The LexID is a unique, 12-digit identifier assigned after a

successful identity resolution, generated by comparing input information provided by Participating States

across billions of unique public records. Through identity analytics, input information is resolved to a single

The submission occurs most commonly through an automated process in which all demographic information is

utilized, and less often via manual portal query in which a user enters selected demographic elements for matching.

Match: An instance in which a state identifies an

individual who is already receiving SNAP or D-

SNAP benefits in another state; generally in the

context of dual participation prevention (prior to

approval of benefits in a second state).

Collision: An instance in which dual participation

– receipt of SNAP or D-SNAP in two or more states

in the same month - has occurred. The NAC

supports next-day identification or “early

detection” of dual participation when it does

occur.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

individual with an extremely high degree of accuracy

. The level of accuracy is captured in the score of 0-

100, with 100 representing the highest confidence. If the LexID is associated with the input information,

it compares that ID to all the LexIDs currently associated with the NAC. When matches for dual

participation are identified, LexisNexis adds applicable match codes, such as Full/Partial Name, SSN, DOB

and Address to provide additional match information and to ensure that all matches include match code

combinations (rather than a LexID score only).

Because identities for a large portion of SNAP-eligible individuals cannot be assigned a LexID (children in

particular, who have limited public records), LexisNexis matches directly with Full/Partial Name, SSN, DOB

and Address to return dual participation results, also in the form of match code combinations. For

example, a match code combination of “NSD” indicates the name, Social Security Number, and date of

birth are the same in both states. In practice, states use both the match code combinations and – to a

much lesser degree – LexID scores to dictate the process to be followed in addressing a match, or whether

a match is investigated at all.

Project Timeline

The NAC project can be characterized as taking place over three partially-overlapping phases–pre-pilot,

“Big Bang,” and pilot operations. Table 2 provides a high-level summary of key dates and timeframes over

the life of the project.

Table 2

Project Timeline

Phase

Date

Pre-pilot

April 2013

LexisNexis awarded development contract to design, develop,

implement, host, and provide ongoing operation of the NAC

June 2013 – May 2014

Pilot states develop common business rules for interstate

communication, establish procedures for client contact, and clarify

applicable SNAP policies

August 2013 – May 2014

Pilot states and LexisNexis develop and modify file layouts and test

submission of data files

Big Bang

June 5, 2014

States receive list of all potential active dual participants as of June 2,

2015

June – August 2014

States contact, take action on dual participants identified at Big Bang

Pilot

Operations

June 5, 2014

States initiate use of the NAC to prevent new instances of dual

participation (pilot operations begin)

June 2014 – May 2015

Ongoing pilot activities (includes prevention and identification of dual

participation that is not prevented)

May 31, 2015

Conclusion of pilot operations

The LexID is generated independently by LexisNexis, and not by the participating states or federal partners.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

The NAC Evaluation

Summary of Previous Reports

This final report is the fifth in a series of documents produced by Public Consulting Group (PCG) as part of

the NAC Evaluation. The Pre-Pilot Report, submitted in July 2013, found that the processes in place for

identifying and acting on dual participation prior to the NAC were not particularly advanced. Consortium

states generally relied on client self-attestation and caseworker experience to prevent dual participation.

Use of the optional quarterly PARIS match of common Social Security Numbers to identify possible

instances of dual participation varied drastically, with two states not utilizing the match at all for SNAP

recipients. The report also documented states’ different policies and procedures for verifying out of state

participation, ranging from placing the onus for obtaining verification mostly on the client to relying

heavily on state staff to confirm eligibility status.

Perhaps most significant from an evaluation standpoint was that the extent to which potential dual SNAP

participation was being prevented – either through the client’s own reporting or states’ pre-authorization

business practices–was not captured in any consistent manner within or among states. In states where

denial and closure codes from eligibility systems were analyzed, more dual participation actions were

associated with closures than denials, suggesting a lack of effectiveness in pre-authorization identification

of eligibility in another state.

The second evaluation document (Three Month

Report) was submitted in September 2014. This

report described the business rules and processes

states put in place to utilize the NAC, illustrated

the prevalence and characteristics of dual

participation in the pilot states at the point the

NAC became operational, and provided initial

observations on the successes, challenges, and

lessons learned in the months leading up to the

availability of the NAC tool and the first three

months following its implementation.

The third evaluation document (Six Month

Report) was submitted in December 2014.

It included analysis of the changes in dual SNAP participation between a pre-implementation period

(March-May 2014) selected to represent the prevalence of dual participation immediately preceding

implementation and post-implementation period (September-October 2014), made preliminary

observations regarding linkages between implementation success and business process design, and

identified areas of improvement in the communication between states.

The fourth evaluation document (Twelve Month Report), submitted in June 2015, provided an update of

pilot state business processes at the conclusion of the pilot, an analysis of the likelihood that matches

with certain common or similar demographic information do in fact represent the same individual;

statistics regarding the frequency of dual participation for the months of August 2014 to March 2015; and

a preliminary analysis of pilot state effectiveness in preventing dual participation for the months of August

2014 through March 2015.

A Note on the NAC and D-SNAP

The need to determine if individuals applying for

Disaster SNAP benefits were already active in another

state was an important impetus for the development of

the NAC. However, no disasters warranting

implementation of a D-SNAP effort occurred during the

NAC pilot. The evaluation was therefore unable to

draw any conclusions with D-SNAP data regarding the

NAC’s ability to prevent dual participation under

disaster circumstances.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Analyses included in the Three, Six, and Twelve Month Reports are inserted and expanded throughout

this document. The most relevant findings and observations are also integrated into the

Recommendations section of the report.

Data Sources

Numerous analyses have been conducted as part of the evaluation, and specific methods used are

addressed where applicable throughout the report. The following provides a summary of the data and

information sources utilized to support evaluation observations and findings.

Pre-Pilot Files: PCG received files containing data on individuals identified as possible dual participants

by the NAC’s matching algorithm for the benefit months of September 2013 through May 2014 (the last

month before the pilot began). Statistics from these months provide meaningful baseline information as

it reflects the prevalence of potential dual participation in the months prior to the use of the NAC and can

be used to document how successful pre-NAC efforts to combat dual participation were in relation to the

pilot period.

Big Bang Files: On June 5, 2014, PCG received an Excel workbook created by LexisNexis containing all

collisions identified by the NAC for the month of June 2014 as of “go-live” (June 2

) and the demographic

information associated with them

. This data, in conjunction with the supplemental data referenced

below, was used for a numerous calculations, including, but not limited to, the average/median length of

dual participation and the value of the SNAP allotment attributed to the dual participant.

Supplemental Big Bang Data: PCG modified the file provided by LexisNexis, preparing separate

documents for each of the five pilot states and adding data elements for the states to enter for each

individual identified as a possible dual participant. This data collection by the states occurred in two

“tiers”:

 Tier One data collection consisted of:

 SNAP allotment for entire case in benefit month

 Household size in benefit month

 Next recertification date

 Total SNAP redemption for case during prior benefit month

 Client eligibility date (to confirm information provided via administrative data)

 Action taken/disposition

Additional dual participation would occur for June 2014 on applications approved and individuals added to cases

after the 2

of the month. Those instances of dual participation were not a part of the Big Bang file.

Options provided for description of the action taken:

a) Client responded to contact letter -Individual removed and case remains open;

b) Client responded to contact letter - Individual closed and case closed;

c) Client responded to contact letter - Individual remains eligible;

d) Client DID NOT respond to contact letter - individual removed and case remains open;

e) Client DID NOT respond to contact letter - individual closed and case closed;

f) Case already closed OR individual removed; and

g) Match not valid.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

 Tier Two data collection included two sets of data:

 Claims-related information on the individuals for whom Tier 1 data was collected; and

 For individuals identified at the Big Bang whose only common data element is a Social Security

Number, an explanation/reason for the discrepancy in demographic information.

Match Search History (MSH) Files: Beginning in June 2014 and continuing through July 2015

, pilot

states provided PCG with daily Match Search History (MSH) files. Information on MSH files is organized

into four “Activity Types” documenting the previous days’ NAC activity for four categories of information:

1. Single: a record of requests (and resulting matches if applicable) for match information made by

states via NAC portal queries or via state eligibility systems connected to the NAC by batch or

real-time web service;

2. Batch: a record of requests (and resulting matches if applicable) for match information made via

batch process;

3. Passive: a notification informing a matching state that an initiating state conducted a search that

generated a match; and

4. Build-Time Collisions: A record of new collisions.

These files were used to identify the prevalence of, and state success in preventing, dual participation

during the pilot.

Qualitative Data: PCG collected qualitative data throughout the evaluation, primarily to gain

understanding of the business processes in place to address dual participation both before and after the

NAC became available.

 In the pre-pilot phase, PCG conducted site visits in each of the five consortium states to explore the

processes states had in place for identifying and addressing dual participation prior to the NAC.

 During pilot operations, PCG compiled regular updates from the pilot states on their internal rules

and processes. In addition, two “face-to-face” meetings were held. In these forums, states reported

on challenges and accomplishments that informed the ongoing evaluation work. In addition, SNAP

regulations, business rules developed and published jointly by pilot states, and technical requirement

documents authored by LexisNexis were referenced to support understanding of NAC

implementation.

Biweekly conference calls held throughout the project provided ongoing updates from pilot states, USDA-

FNS, and LexisNexis regarding policy, business process, and technical issues.

In order to determine if matches received during the pilot period became instances of dual participation, data

was collected for 2 months following the conclusion of the pilot.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

NAC Business Rules and Processes

From the outset of the project, the pilot states understood the importance of establishing a set of

consistent business processes across the pilot that, if implemented successfully, would allow them to use

information received via the NAC to improve customer service, reduce worker effort, and prevent dual

participation. The pilot states and USDA-FNS collaborated to establish business rules that sought to

achieve those aims while maintaining and promoting access to SNAP and D-SNAP benefits for eligible

participants and compliance with regulatory requirements. States also recognized that due to

organizational capacity and the desire for some autonomy in decision-making, it would be necessary to

allow for the establishment of NAC-specific business rules and processes for some components of the

project. For example, how staff would access NAC information, the degree that NAC information would

be available through integration with existing eligibility systems, and the prioritization of certain types of

information provided by the NAC.

This section addresses the business rules and processes developed prior to pilot operations and their

maturation and adaptation that occurred during the pilot.

Common Business Rules

The common business rules - intended to be followed identically by all five pilot states - were designed to

address how states were to handle two general scenarios:

1. Active dual participation – when NAC data indicates an individual is receiving SNAP in two pilot

states (a collision); and

2. Dual participation prevention–when NAC data indicates an individual is receiving SNAP in one

state and is applying, but has not yet been approved, for SNAP benefits in a second state (a

match).

Active dual participation. One set of business rules addresses the process for communicating with clients

and states when individuals are identified through the NAC as potentially receiving SNAP in one or more

of the pilot states, also referred to as a “collision.”

Pilot states initially intended for information received through the NAC on active dual participants to be

treated as “verified upon receipt.” However, while that information would seemingly verify that an

individual was an active SNAP recipient in two states, it could not serve as verification of the state in which

he/she was actually residing. Therefore, the NAC information could not be considered verified upon

receipt. Accordingly, the business rules were written to direct the states involved to send a Request for

Contact Notice to the head of household. This notice was to include 1) the reason for the contact (that a

computer match indicated that one or more household members may be active in another state); and, 2)

a request for proof of residency and verification that SNAP benefits for the person/persons in question

have been terminated in the other state.

The process dictating that a contact notice be sent prior to taking action on the case represents a

departure from SNAP Simplified Reporting rules established in Section 6(c)(1)(d) of the Food and Nutrition

Act of 2008 (the Act). To address this issue, waiver requests were submitted by the pilot states and

approved by USDA (under Section 17(b)(1) of the Act) to allow adoption of the NAC business rules.

Without the waivers, states would have been prevented from acting on information that is not verified

upon receipt (NAC collisions specifically) until the next client contact.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Dual participation prevention. A second set of business rules was established for communication

between states when a match is received indicating an individual is already an active SNAP participant in

one of the pilot states but prior to authorization in the state receiving the match.

To reduce the overall work effort required to resolve a match, states developed business rules that

assumed certain actions would be taken within specified timeframes by their fellow pilot states.

Specifically, if the match was received prior to the 15

of a month the initiating state (see box) was

directed to assume that the matching state would

take action to close the case/remove the

individual prior to the issuance of the following

month’s benefits. This was a key feature of the

rules as they were originally implemented.

However, the pilot states discovered at a relatively

early stage of the pilot that the matching state was

not always able to take the action as dictated in

the business rules and dual participation occurred as a result. Accordingly, the rules were modified so as

to require an email response from the matching state prior to approval of SNAP in the initiating state.

The Common Business Rules in effect as of the conclusion of the pilot are found in Appendix A.

State-Specific Business Processes

The common business rules dictate that states participating in the pilot:

“…must submit ALL household member data

to the NAC prior to certification of benefits,

including any new household members. The only exception would be on expedited cases

where States may follow their own policy or procedure regarding the processing of the

initial months’ benefits.”

States have in fact implemented different processes in response to the flexibility offered in the common

business rules relative to expedited case processing. The processes that have been instituted to address

other business needs, and the level of automation developed to support the NAC, also vary significantly

from state to state.

Detailed descriptions of state-specific business processes are found in Appendices D and E. In Table 3

below key components of these processes – each with impact on project outcomes – are highlighted. The

impacts of incorporating these business processes – or not – are addressed in the Recommendations

section of this report.

Member data includes (but is not limited to) name, address (home and mailing), Social Security Number, gender,

race, ethnicity, and date of birth for the participant and additional information on the head of household.

Initiating State: the state in which an individual is

applying and has not yet been approved for SNAP.

Matching State: the state in which an individual is

already receiving SNAP benefits.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Table 3

Summary of State Business Processes

Process to check NAC on date of application when same-day processing required

Process to check NAC on same day for individuals being added to an existing case

NAC data imported directly into state eligibility system

Eligibility system requires resolution of NAC match prior to authorization

NAC portal access for eligibility field staff

Automation of contact emails to matching state

Treatment of Matches

A key question for states implementing the NAC concerns how matches are treated depending on their

“strength.” At a basic level, match strength represents the degree that commonalities in a data element

or elements suggest that an individual receiving SNAP or D-SNAP in one state is the same person receiving

the benefit in another. The matches that a state chooses to investigate and/or act upon impacts both the

efficiency and efficacy of the NAC in preventing and

identifying dual participation.

Table 4 documents the ten matching elements that, either

alone or in combination, suggest that individuals reported by

two or more states may be the same person.

Because some matching elements are stronger indicators

than others, and some combinations of these elements

when matched indicate a strong likelihood that the records

represent the same individual, match code combination

rankings (see Table 5) provide a means to prioritize and

organize data produced by the NAC. For example, the “NSD”

match indicates the full name, full SSN, and date of birth are

identical. If two states submit records in which an applicant

or recipient’s data matches on all three of these elements, it

is very likely that this is the same person.

NAC data is imported to the Clearance Summary, which is outside the eligibility system

SNAP application processing timeframes cannot be waived pending conformation of eligibility status.

Table 4

Match Code Key

Match Code

Description

Full Name

Last Name + Partial First

Last Name

Full (exact) SSN

Probable SSN

Date of Birth

Possible Date of Birth

Street Address

City/State Address

Zip Address

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

“Top 5” match code combinations (in bold italic

in table 5) make up a significant majority of the

output from the NAC. For the period of August

2014 through May 2015, they accounted for

84.8% of matches.

With virtually no exceptions,

matches with these five code combinations have

been established as valid per feedback from the

pilot states and analysis. A review of Top 5

matches in which Mississippi was the second

authorizing state found that of 235 Top 5

matches, only 2 (<1%) were identified as invalid

(i.e. not representing a true match of the same

individual).

An analysis of the other 15.2% percent of

matches – referred to as “6+” – found that about 6 in 10 of those were valid. These include matches

generated based on the LexID that are not included in the combinations listed in Table 5. In fact, in some

instances there may only be one matching demographic element, but other public records data accessed

by LexisNexis generates a score indicating a likely match. Thus, when states choose to ignore the LexID

score and only address matches with a strength of 1-5, there is some dual participation that is missed.

Table 6 documents pilot states’ treatment of NAC data based on the match code combination.

Table 6

Treatment of Match Code Combinations by State

As of May 31, 2015

State

Match Code Combinations/LexIDs

Deemed “Valid”

Treatment of Other Matches

Alabama

Prevention: 1-9, and/or LexID of 80+

Collisions:1-3 and LexID 95+

All other hits are investigated.

Florida

1-5, plus SD and LexID of 100

All other match combinations and LexID scores are

ignored.

Georgia

1-5 generally considered valid

Process for determining validity is manual–claims

managers review a daily report and select matches to

forward to matching state for appropriate action.

Louisiana

1-5

Workers are instructed not to take action on other match

code combinations.

Mississippi

1-5 (automated email sent to matching

state)

Match Code Combinations 6 + require review; worker

prompts automated email to matching state if warranted.

“SSN Only” matches excluded from this calculation. These are matches in which the only common element is the

Social Security Number and are often the result of a data entry error in one of the states involved. They generally do

not reflect an instance of actual dual participation.

“Valid” generally meaning that actions such as automated emails or contact letters are initiated without further

investigation of validity.

Alabama is the only pilot state that treats match code combinations differently depending on whether the match

is received for prevention purposes or represents a collision.

Table 5

Match Code Combination Strength Rankings

Rank

Code

Combination

Rank

Code

Combination

NSD

NDAZ

VSD

VDACZ

NSB

VDAC

VSB

VDAZ

NPD

NBACZ

VPD

NBAC

NPB

NBAZ

VPB

VBACZ

VBAC

NDACZ

VBAZ

NDAC

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Analyses in the evaluation typically use Top 5 match code combinations when tallying frequency of

matches and collisions.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Key Research Questions

The National Accuracy Clearinghouse (NAC) evaluation focuses on four central research questions:

1. Has the NAC resulted in a reduction in dual SNAP participation?

2. How effective have states been in utilizing the NAC to prevent dual SNAP participation?

3. How does the NAC compare to the use of PARIS?

4. What is the NAC’s return on investment?

The analyses conducted to answer the research questions utilize data related ONLY to inter-state

participation. The NAC does identify instances of intra-state dual participation as well. These are

generated when the NAC matches an individual who appears to be active in two cases in the same state.

However, intra-state dual participation is outside the scope of the evaluation and is not included in the

following analyses.

Has the NAC resulted in a reduction in dual participation?

Several months elapsed between the first submissions to the database and actual use of the NAC by the

states to prevent and identify dual participation. This provided the evaluation with pre-pilot data, which

allows for comparison of the frequency of dual participation before and after the NAC was available for

use by the pilot states.

Table 7 compares the prevalence of dual participation (Top 5 match code combinations) prior to pilot

operations to the final four months of the pilot (February - May 2015).

Table 7

Comparison of Dual Participation

, Pre-Pilot and Pilot Periods

Top 5 Match Code Combinations

Monthly average,

pre-pilot*

Monthly average,

final 4 pilot months

Change from

pre-pilot

Alabama

1592

301

-81.1%

Florida

3383

2446

-27.7%

Georgia

3323

2427

-27.0%

Louisiana

862

249

-71.1%

Mississippi

882

166

-81.2%

*September 2013-May 2014; excludes November 2013 due to data abnormalities

Reductions in dual participation have occurred in all five pilot states, but with large variations in size. It is

important to stress that these dual participation statistics are duplicative – i.e. an instance reported for

one state is also included in the tally for the other state in which the individual is an active SNAP recipient.

A previous report referenced the “symbiotic” nature of dual participation in Georgia and Florida, and the

nearly-identical statistics in Table 7 reinforce that observation. On the other hand, the similarities in the

percentage reduction in Alabama and Mississippi (both 81%) appear to be the product of their practices

and not a reflection of their caseloads mirroring each other (note that the raw numbers are quite

different).

Number of dual participants calculated by adding entries and continuations.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

The analysis uses the final four months

of the pilot for comparison to the pre-

pilot months, assuming that a late-pilot

timeframe is more likely to demonstrate

how the NAC will impact dual

participation on an ongoing basis (as

states would have had time to adjust

and hone their processes). To test this

assumption, Figure 1 compares dual

participation entries and continuations

(see box at right) in the middle and late

stages of the pilot.

Figure 1

Change in Entries and Continuations,

Middle (October 2014 thru January 2015) and Late (February – May 2015) Stages of Pilot

Average, October 2014 – January 2015

Average, February – May 2015

The bar graphs show all five states reduced the number of entries and continuations as the project

matured. For example, Alabama averaged fifty fewer entries on average in the late phase compared to

earlier in the pilot (233, down from 283). Louisiana’s continuations fell 40 percent, from 199 to 120, over

the comparison periods. Improvement between middle and late stages of the pilot suggest that states

gained a better understanding of how to utilize the NAC over the pilot period, and supports the use of

late-pilot data in Table 7.

-17.7%

-9.4%

-10.7%

-26.2%

-22.8%

200

400

600

800

1000

AL FL GA LA MS

Entries

-35.2%

-14.9%

-17.4%

-39.7%

-26.3%

200

400

600

800

1000

1200

1400

1600

1800

2000

AL FL GA LA MS

Continuations

Entries: Collisions that are present in the target month but were not

present in the previous month. An entry represents an instance in

which dual participation was not prevented.

Exits: Collisions that were present in the previous month but not the

target month.

Continuations: Collisions that are present in both the target month and

were present in the previous month. A continuation represents an

instance in which the early detection of dual participation made

possible by the NAC was not utilized.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Reductions in dual participation could occur simply because fewer individuals are receiving SNAP benefits

overall. Table 8 illustrates another measure of the impact of the NAC on dual participation by comparing

the number of dual participants

as a percentage of eligible individuals in the month before pilot

operations began (seen previously in Table 1) to the same statistic captured in the last month of the pilot.

Table 8

Dual Participation as a Percentage of SNAP Participants, May 2014 and May 2015

May 2014

May 2015

SNAP

Participants

Dual

participants

SNAP

Participants

Dual

participants

Alabama

898,301

1534

0.171%

881,147

310

0.035%

Florida

3,487,797

3534

0.101%

3,630,463

2424

0.067%

Georgia

1,847,395

3464

0.188%

1,785,403

2354

0.132%

Louisiana

866,941

755

0.087%

854,073

230

0.027%

Mississippi

650,853

789

0.121%

628,737

146

0.023%

In each of the five pilot states, both the raw number and percentage of the caseload made up of dual

participants declined, suggesting that the drop is not primarily due to a reduction in the overall SNAP

caseload (which occurred in four of the five pilot states). For example, before NAC implementation, just

over one-tenth of one percent (.121) of SNAP recipients in Mississippi were dual participants; by May

2015, that percentage had dropped to .023.

Top 5 Match code combinations only.

Eligible individuals statistics reported by FNS is initial data for May 2015.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

How effective have states been in utilizing the NAC to prevent dual participation?

The primary goal of the NAC is to provide states with information that supports prevention of dual

participation in SNAP. The analysis in the previous section does not address which of the two states

involved in an instance of dual participation was in the best position to prevent it from occurring, and

whether they were successful. Here, matches received prior to authorization of SNAP benefits in a second

state are identified and tracked over a three month period to determine if they became collisions. This is

an indicator of how successful the second state has been in utilizing the NAC to prevent dual participation.

The relative success that pilot states have had in achieving the goal of prevention is assessed by identifying

the matches (opportunities to prevent dual participation) received in a given month and “following” those

same individuals in the state in which the match was received. By tracking these matches to see if a

collision eventually occurred, a determination is made of whether or not prevention efforts were

successful or not.

Figure 2 illustrates, for each pilot state, the percentage of Top 5 matches received in August 2014 through

May 2015 that became collisions. Unlike the statistics in Table 7, which are duplicative (i.e. a collision in

one state is also tallied as a collision in another), these percentages reflect the success rate of only the

state in which the individual is applying and not yet eligible (the initiating state). See Appendix B for

methodology used in this calculation

Figure 2

Percentage of Top 5 Matches that Became Collisions: August 2014–May 2015

To summarize, two states – Alabama and Mississippi – have been extremely effective in preventing dual

participation, with consistently less than 10 percent of matches resulting in collisions. Louisiana’s

success rate, while not as consistent, has achieved similar results in some pilot months. Florida and

This analysis assumes that the 2nd authorizing state (initiating state) is primarily responsible for dual participation.

A limited number of collisions presumably occur because the 1st authorizing state (matching state) failed to take

action to close a case or remove an individual as requested by the initiating state.

0.00%

5.00%

10.00%

15.00%

20.00%

25.00%

30.00%

35.00%

40.00%

45.00%

Alabama

Florida

Georgia

Louisiana

Mississippi

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Georgia have not been as successful in preventing dual participation, with 30-45 percent of matches

becoming collisions during the pilot.

The relative rates of success illustrated in Figure 2 align closely with what would be expected based on

the approaches states have taken to utilize NAC information. Mississippi has implemented a range of

processes that would tend to support better outcomes (such as integration with the eligibility system and

use of web services), whereas states with less successful outcomes have not put similar practices in place.

For instance, in Florida “there is no process in place to check the NAC if disposition on an application is

taken prior to processing of the overnight batch

” and, with respect to the addition of new household

members in Georgia, the “caseworker typically does not become aware of the match until after benefits

have been authorized.” The differences in business processes and systems integration not only provide at

least a partial explanation for the varied outcomes achieved by states, but also support a set of practices

that may be adopted to improve upon and maximize the effectiveness of the NAC.

For example, an application may need to be approved on the day it is registered in the eligibility system in order

to meet expedited processing timeframes. Absent a portal query or web services connection to the NAC, benefits

would be approved before the results of an overnight match with the NAC database is received.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

How does the NAC compare to the use of PARIS?

The National Accuracy Clearinghouse is not the first effort to match public assistance data across states.

In 1997, the Public Assistance Reporting Information System, or PARIS, was initiated to identify individuals

who may be active participants in the same program in more than one state. In this section, the

differences between the NAC and PARIS are explored.

NAC and PARIS – key differences

Multiple factors influence the degree that the NAC and PARIS are able to reduce dual participation

overpayments effectively and efficiently. The characteristics and limitations of the PARIS match were

identified during PCG’s pre-pilot site visits and in two separate evaluations of PARIS, conducted by the

General Accounting Office and Health Systems Research, Inc. (an Altarum Company)

. The distinctions

between the two solutions are considerable, and are detailed below.

Frequency of the data match. States may submit data to be matched for PARIS on a quarterly basis,

whereas the NAC database is updated daily with information on beneficiaries’ status (and may be queried

at any time).

Identification of the benefit month. A match generated via PARIS indicates an individual was eligible in two

states within a three month period. And, while the results of the PARIS match do document the dates that

the benefits in question were received, the process of determining if an actual overlap occurred is left up

to the states. The standard format for NAC contributory files includes the benefit month, and a match is

generated only if the match occurred for a specific month. So, unlike PARIS, the NAC provides states with

confirmation that a match represents overlapping SNAP receipt, and not just receipt in the same quarter.

The GAO Report includes the following relative to PARIS’ limitations:

“…participating states do not have adequate protocols or guidelines to facilitate critical

interstate communication. As a result, some states have reported problems that

compromise the effectiveness of the project, such as difficulty determining whether an

individual identified in a match is actually receiving benefits in another state.”

For example, if an individual is active in Mississippi in only the first month of a quarter and active in

Alabama in only the last month of a quarter, a match is generated and states must investigate further to

determine that in fact no SNAP eligibility overlap occurred.

Prevention vs. pay-and-chase. Perhaps the most significant difference between the NAC and PARIS is that

PARIS only identifies potential dual participation after it occurs, and sometimes several months afterward.

The GAO report specifically cited this as a limitation of PARIS, stating “…because the PARIS match is only

designed to identify people after they are already on the rolls, it does not enable the states to prevent

improper payments from being made in the first place.” Conversely, the NAC allows states to prevent dual

participation before it occurs by supporting the submission of data on SNAP recipients and applicants.

The graphic below illustrates the timeline associated with using PARIS to identify possible dual

participation. In this hypothetical scenario, an individual applied for and was approved for benefits in

GAO-01-935, “PARIS Project Can Help States Reduce Improper Benefit Payments.” September 2001; and

“Evaluation to Determine the Effectiveness of the Public Assistance Reporting and Information System Final Report.”

Health Systems Research, Inc., June 30, 2007.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Mississippi in January, moved to Alabama in March, and was approved for SNAP there immediately upon

application (months highlighted in yellow represent months in which a PARIS match is conducted).

Figure 3

PARIS Timeline Scenario

Client

approved for

SNAP in MS

PARIS match

conducted

Client applies,

approved for

SNAP in AL

PARIS match

conducted

case closed in

MS eff. 6/30

January

February

March

April

May

June

July

SNAP receipt in MS

SNAP receipt in AL

Four months of dual participation

The four month spell of dual participation in this example (the red bars) assumes a quick turnaround on

the utilization of the PARIS data received in May; in many instances the dual participation would continue

beyond four months as the PARIS data is processed and/or the client’s actual circumstances are

investigated. This investigation includes a determination of whether overlapping eligibility occurred at all

– as noted above, PARIS generates matches based on eligibility in two states within a quarter.

The same scenario is illustrated in Figure 4, except it illustrates the impact of near real-time data matching

for SNAP applicants that the NAC supports. When the client applied for SNAP in Alabama, the NAC was

available to confirm the individual already received SNAP in March in Mississippi. Therefore, SNAP

benefits in Alabama were authorized effective in April.

Figure 4

NAC Timeline Scenario

Client

approved for

SNAP in MS

Client applies

for SNAP in AL

Client

approved for

SNAP in AL

January

February

March

April

May

June

July

SNAP receipt in MS

Case closed in

MS eff. 3/31

SNAP receipt in AL

No dual participation

In addition, if a state’s business process and/or system integration prevents a NAC inquiry from being

conducted before an application is authorized (such as the need to approve expedited benefits on the day

the application is registered), the timing of the match is such that the spell of dual participation would be

one month instead of four.

Administrative Cost Avoidance. In addition to supporting the prevention of duplicate SNAP issuance, the

NAC model allows states to reduce costs associated with fraud/overpayment investigations, processing

claims, and recovering benefits, because the improper payment never occurs in the first place.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Demographic matching points. The PARIS match exclusively

uses the Social Security Number submitted by participating

states to identify possible dual participation – if an exact SSN

match is not produced, no results are communicated to the

states involved. The NAC uses multiple demographic

elements and public records to establish matches that states

are able to prioritize based on the level of confidence that a

match truly represents the same individual.

Table 9 demonstrates the impact of the one Top 5 match

combination

that does not include an exact SSN match.

Because PARIS matches are generated only when SSNs are

exactly the same, none of the ninety-two matches identified

here would have been reported by PARIS (NPD matches

indicate an exact match on name and date of birth, and a

close - but not exact - match on SSN).

Several other matches – albeit with lesser rates of validity than the Top 5 match code combinations–are

identified by the NAC and are not available to states through PARIS because they do not include an exact

SSN match. Table 10 provides information on the frequency and validity identified for these combinations

at the Big Bang. Note that instances in which the match codes in the first column are not included in

Table 5, the match was actually generated by additional public records information that support the

assignment of a LexID.

Table 10

Match Code Validity, Combinations without Exact SSN Match

Match

Code*

# of

Collisions

(not

valid)

(valid)

% Valid

VPD

47.1%

NPB

47.8%

VPB

25.0%

58.6%

26.3%

46.7%

28.6%

0.0%

WPD

25.0%

0.0%

100.0%

VPDACZ

100.0%

0.0%

*See Table 5 for code translations

Top 5 match combinations have been found to be valid with very limited exceptions as noted in the Twelve

Month Report.

Table 9

NPD Matches at Big Bang

Two-state

combo

NPD matches

at Big Bang

AL-FL

AL-GA

AL-LA

AL-MS

FL-GA

FL-LA

FL-MS

GA-LA

GA-MS

LA-MS

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

In addition to identifying matches that PARIS does not, the range of data utilized by the NAC provides a

high degree of certainty regarding the large majority of matches. This reduces the time and resources

required for follow up between states.

Table 11 summarizes these key differences between NAC and PARIS.

Table 11

PARIS-NAC Comparison

Component

PARIS

NAC

Frequency of data submission for matching

Quarterly

Daily

Supports dual participation prevention

Matches generated for specific benefit months

Demographic matching points

SSN only

Multiple

Multiple program matching

Case Study - Florida

The NAC evaluation Pre-Pilot Report included a review of the five pilot states’ processes for identifying

and addressing dual participation in SNAP prior to implementation of the NAC. The report described the

degree that the states were utilizing the PARIS match as part of those efforts. The observations showed a

wide variation, from no use of PARIS whatsoever (in Alabama and Georgia) to - in Florida - a systematic

process for utilizing PARIS data. Florida’s methods were described as follows:

Upon receipt of quarterly match data via PARIS, an automated process filters out cases in

which benefit overlap did not occur. The remaining cases, which require action of some

kind, are then divided between active and inactive. For the active cases, the system

automatically generates a letter to the client informing them that Florida has information

indicating they appear to be receiving SNAP in another state. The letter gives the client 10

days to produce verification of residency. A worker assigned to a Case Management Unit

works a report displaying the matches that resulted in a contact letter being delivered.

After 10 days, this worker checks to see if a document has been scanned in response to

the letter and if the client has not produced verification within the allotted time frame

action is taken to close the case.

Data collected at the Big Bang illustrates the prevalence of dual participation in Florida even with a

relatively robust use of PARIS data. The following statistics reflect collision volume just prior to NAC “go-

live” in June 2014:

Table 12

Florida Collision Volume as of June 2, 2014

Collision Type

SSN only

189

Top 5

2373

487

There is also variance in the months a state selects to submit for matching; some select all three months in a

quarter and others select just the month during which the match is conducted.

The NAC has the capacity to expand to other programs.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Furthermore, when dual participation did occur, it continued for several months, as evidenced by the

average and median spells of dual participation for individuals found to be receiving SNAP in Florida and

another pilot state:

Table 13

Spell of Dual Participation from First Month of Benefit Overlap through June 2014 (Florida)

(n=1561)*

Spell Length

(months)

Average

6.2

Median

*Top 5 collisions in which Florida was found to be the 2nd authorizing state

Neither of these results are surprising given that PARIS operates under the pay and chase model and the

variety of data elements it utilizes to match individuals are significantly limited compared to the NAC.

Regardless, the data points to the potential advantages of the NAC compared to PARIS. Each of the

instances of dual participation represented above – and the approximately $135/month

of SNAP

overpayments that accumulated every month duplicate benefits were issued – could have been prevented

had the NAC been available AND utilized to its full potential by Florida and the other pilot states.

A passage from the 2001 GAO report summarizes the differences between the NAC and PARIS by

documenting state officials’ vision at that time:

“Officials from most states we spoke with said they would like a datasharing process that

could be used before benefits are provided—that is, a process that would allow state

caseworkers to check other states’ data to see if an applicant was already receiving

benefits elsewhere before the state approved an application for benefits. Such a process

would have to provide prompt responses (probably within 24 hours) to inquiries—

something very different from the quarterly PARIS matches. One option for this process

includes a national database of clients receiving public assistance in any state. Such a

database would be maintained by the federal government and would consist of records

submitted and regularly updated by the states.”

Representatives from pilot states expressed hesitation to expend resources on PARIS given its limitations

in identifying dual participants effectively and the resulting work required for follow up. Those limitations

would continue to exist barring fundamental changes to PARIS – changes that would likely exceed what

would reasonably be considered an expansion of that existing system.

Based on Florida-specific collisions identified at the Big Bang, including those that would not be identified by

PARIS.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

What is the NAC’s Return on Investment?

Evaluation findings indicate that dual participation has in fact decreased in the pilot states, and that the

NAC gives states the capacity to prevent dual participation. This section addresses the question of

whether, and by how much, that decrease translates into savings in SNAP overpayment avoidance from

each pilot state given the manpower and technical resources required to participate and necessary to act

on the information the NAC provides. The analysis is conducted in three parts, each with multiple steps:

 Monthly savings in SNAP overpayment avoidance (100% federal dollars)

 Monthly costs in resources required to support the overpayment avoidance (all expenses eligible

for 50-50 federal/state SNAP administrative cost split)

 Net Impact

Source data used is referenced in the narrative associated with the relevant step in the calculations.

Limitations/Assumptions

The estimates of cost savings should be considered conservative, for several reasons. First, the focus of

this analysis is solely on the prevention of dual participation; some additional savings is being realized

through earlier detection of ongoing dual participation when it does occur. Table 14 demonstrates the

impact of the NAC with respect to early detection. In this analysis, the decay rate

of dual participation

is calculated by comparing entries from December 2013 (before pilot) and December 2014 (during pilot),

and following these individuals for five months.

Table 14

Percentage of Top 5 Entries Remaining Dual Participants in Succeeding Months, Pre-Pilot and Pilot Comparison

Time Period

December

Entries

Percentage Remaining as Dual Participants in

January

February

March

April

May

Pre-Pilot 2013-14

683

71.4%

54.3%

45.8%

38.4%

33.2%

Pilot 2014-15

248

21.0%

0.8%

0.4%

0.0%

Pre-Pilot 2013-14

1844

74.8%

57.4%

45.7%

41.5%

36.9%

Pilot 2014-15

807

51.4%

29.6%

23.8%

19.5%

17.8%

Pre-Pilot 2013-14

2156

79.0%

62.4%

49.9%

45.1%

41.4%

Pilot 2014-15

824

49.6%

28.5%

22.5%

18.2%

17.1%

Pre-Pilot 2013-14

275

66.2%

56.4%

39.6%

34.2%

30.2%

Pilot 2014-15

232

41.4%

22.0%

13.8%

7.8%

6.5%

Pre-Pilot 2013-14

298

60.1%

43.6%

32.2%

28.2%

25.8%

Pilot 2014-15

126

34.9%

15.1%

11.1%

5.6%

3.2%

Table 14 illustrates that the NAC supports a faster resolution of dual participation when it does occur. For

example, in Louisiana, more than thirty percent of new instances of dual participation in December 2013

continued five months later. Once the NAC was in place, less than seven percent of the December 2014

entries were in dual participation status five months out.

Decay rate refers to decline in the percentage of clients who remain dual participants in the five succeeding

months following entry.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Other factors that have the effect of reducing the actual savings include the following:

 A key variable in the calculation – the difference in entries from the pre-pilot phase to the pilot -

uses only the Top 5 match code combinations. Depending on the state and its approach for

treatment of matches, some additional savings has been realized through the prevention of valid

“6+” matches, which accounted for 15 percent of all matches (excluding “SSN-only” matches)

from August 2014 – May 2015.

 The model is designed to represent the actual impact of the NAC through the end of pilot

operations, not the potential of the solution. Clearly, the pilot states (some more than others)

have room to improve their processes and should be able to increase their prevention rates over

time. The Recommendations section of this report identifies efforts that can support such

improvement.

 The model assumes that when dual participation is prevented, the individual remains eligible in

their actual state of residence.

 The savings does not include the recoupment of overpayments due to dual participation identified

at project “go-live.”

Other methods used in the calculations that may result in under-counting the impact of the NAC are noted

in the narrative.

Savings Calculation

Step 1: Savings per month per instance of prevention

Each month that dual participation is prevented for an individual, a SNAP overpayment is avoided. The

value of this avoidance each month depends on the characteristics of the household–it could represent

an even distribution of the monthly SNAP allotment (in the case where all household members are dual

participants and the case would not have been opened), or a smaller percentage of the allotment (in the

case where an individual is removed but the case remains open). These factors are considered in

calculating the savings per month in each instance of prevention (see Appendix G for methodology relative

to percentage of allotment considered). Source data for this calculation is the Supplemental Big Bang

information provided by the pilot states.

Proportion of matches that prevent dual participation (DP) in

which all HH members would have been dual participants

44.1%

61.8%

45.8%

46.7%

48.8%

Average SNAP allotment per individual when all HH members

are dual participants

$140

$146

$159

$142

$145

Proportion of matches that prevent DP in which portion of case

would have been authorized

55.9%

38.2%

54.2%

53.3%

51.2%

Average SNAP allotment per individual when only portion of

HH is a dual participant

$110

$116

$113

$109

$110

Savings per month per instance of prevention

$123

$135

$134

$124

$127

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Step 2: Savings per spell of avoided dual participation

Using data on individuals identified as dual participants at the Big Bang, the following method was

employed to establish the length of a typical “spell” of dual participation:

 Identified eligibility date in each state

 Selected the latest of the two dates to establish when overlapping eligibility began

 Identified the next recertification date for the individual’s case in each state

 Selected the soonest of the two recertification months, assuming that dual participation

would be identified and addressed at that time.

 The number of months between the start of overlapping eligibility and the next recertification

month establishes the expected length of a dual participation spell.

The median number of months for each state’s spells are reflected below.

Months of dual participation avoided/instance of prevention

6.0

11.0

9.0

10.0

(x Savings per month per instance of prevention)

$123

$135

$134

$124

$127

= Savings per spell of avoided dual participation

$738

$810

$1,474

$1,116

$1,270

Step 3: Savings per month

The impact of the NAC on dual participation is quantified by a comparison of the number of entries per

month before and after implementation of the solution. The difference represents instances of dual

participation that would have occurred absent the NAC solution. The change in entries statistic (-263 for

Alabama below) was calculated using the average number of Top 5 match code entries over eight months

in the pre-pilot period (excluding September and November 2013) and comparing these to the average

number of entries over the last four months of the pilot.

In quantifying impact, it is vital to recognize that an entry is counted in both states involved in dual

participation, and that an individual will generally remain eligible in one of the two states. So, when

dual participation is prevented, only one state should receive “credit” for an avoided overpayment.

Accordingly, the reduction in dual participation entries must acknowledge that while dual participation is

being avoided, SNAP benefits will still be issued in the state in which the individual actually resides.

To avoid double-counting of savings, the change in entries between pre-pilot and pilot is multiplied by the

percentage of each state’s dual participants at the Big Bang that were that state’s “responsibility.” In

other words, the “responsible state” determined eligibility and, upon authorization of SNAP benefits,

created the instance of dual participation. For example, Alabama was the 2

authorizing state for 54.8%

of the dual participation identified at the Big Bang; therefore, applying that percentage to the reduction

in entries (263 x .548 = 144) provides an adjusted change that avoids crediting two states.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Change in entries between pre-pilot and pilot

-263

-361

-378

-114

-149

Percentage of dual participants authorized by state

54.8%

68.8%

32.6%

36.0%

46.7%

Adjusted Change in entries between pre-pilot and

pilot

-144

-248

-123

-41

-70

(x Savings per spell of avoided dual participation)

$738

$810

$1,474

$1,116

$1,270

= Savings per month

$106,272

$200,880

$181,302

$45,756

$88,900

Step 4: Savings per month with “redemption discount”

Analysis of Big Bang Supplemental Data indicated that in approximately 12% of the dual participation

identified at the Big Bang, no SNAP benefits were redeemed in one of the two states involved in the month

prior. In this step, savings are reduced to reflect benefits that would not have been redeemed (88% were

redeemed in both states).

Redemption discount (collisions w/> 0%

redemption)

88%

(x savings per month)

$106,272

$200,880

$181,302

$45,756

$88,900

= Savings per month with redemption discount

$93,519

$176,774

$159,546

$40,265

$78,232

For example, the savings of $200,880 identified in Florida is discounted to $176,774, the amount of

benefits estimated to actually have been redeemed had they been issued. The difference ($24,104)

represents SNAP benefits that were issued, but never redeemed.

Cost Calculation

Step 1: Costs associated with work effort as initiating state

This step captures resources expended when a match is received, including communication with the

matching state and time required to bring the issue to resolution. Note that it does not include “lag time”

between the initial contact with the matching state and the point at which action is taken. Source data

used in this step includes the state survey (staff time and wage) and MSH files (match volume).

Staff hours required to act on each match as

initiating state

0.21

0.10

0.33

0.38

0.02

Monthly match volume (all combinations except

“SSN-Only”)

751

1282

1627

259

311

Average hourly cost (wage + benefits) of staff taking

action

$21.51

$22.31

$20.94

$29.07

$14.87

Costs associated with work effort as initiating state

$3,392

$2,860

$11,243

$2,861

$92

Of note in Step 1 is Mississippi’s low cost in comparison to the other states. This is due in part to the

automation that supports very limited time necessary to act on the information received from the NAC.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Step 2: Costs associated with work effort as matching state

States also expend resources when responding to requests from out-of-state, including communication

with the initiating state and taking action – typically closure of a case or removal of an individual.

Staff hours required to act on each out of state

request (incoming)

0.21

0.14

0.21

0.5

0.08

Monthly volume of out of state requests

479

1974

971

464

342

Average hourly costs (wage + benefits) of staff

working out of state requests

$21.51

$22.31

$20.94

$29.07

$14.87

Costs associated with work effort as matching state

$2,164

$6,166

$4,270

$6,744

$407

Step 3: Other ongoing costs associated with NAC operations

The survey of pilot states asked if there were any other ongoing costs associated with NAC operations. In

addition, the monthly fee paid by states to LexisNexis–based on the number of individuals receiving SNAP

in each state – are included here.

Monthly $ paid to LexisNexis for use of the NAC

$5,000

$10,417

$6,250

$5,000

Other reported monthly admin expenses (e.g. IT staff)

$8,600

$1,447

Other monthly expenses

$13,600

$11,863

$6,250

$5,000

Without actual investment in new staff or equipment, ongoing monthly costs are dependent somewhat

on prioritization. Generally, states integrate technical work associated with the NAC into existing duties.

Step 4: Total costs

The three cost categories calculated in steps 1-3 are summed to generate a monthly ongoing cost for each

state. Ongoing costs vary and are affected by the volume of matches produced by the NAC, inquiries

received from other states, staffing costs, and the degree that processes are automated. Mississippi, on

the low end in all those areas, naturally has the lowest total costs per month as a result.

Costs associated with work effort as initiating state

$3,392

$2,860

$11,243

$2,861

$92

Costs associated with work effort as matching state

$2,164

$6,166

$4,270

$6,744

$407

Other monthly expenses

$13,600

$11,863

$6,250

$5,000

Total costs per month

$19,156

$20,890

$21,763

$14,605

$5,499

Net Impact

The impact of the NAC for each state over a year is calculated by subtracting costs from savings, and is

annualized by multiplying by 12. The calculation does not net out staff costs associated with non-NAC

matching processes, which continue as states receive inquiries resulting from an applicant’s self-

attestation or a PARIS match. In addition, while the savings are characterized as being realized by the

Based on % of individuals each state serves

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

states, the SNAP overpayment avoidance that the NAC supports are 100 percent federally-funded

benefits.

Table 15

NAC Net Impact

Monthly savings

$93,519

$176,774

$159,546

$40,265

$78,232

Monthly costs

$19,156

$20,890

$21,763

$14,605

$5,499

Savings-Costs

$74,363

$155,885

$137,783

$25,660

$72,733

Annualized (Savings-Costs x 12)

$892,360

$1,870,616

$1,653,396

$307,920

$872,792

The pilot-wide net impact of the NAC totals more than $5.6 million. It is important to reiterate that this

estimate is conservative – it focuses on the prevention of dual participation and not early (nearly

immediate) detection that the NAC can also support, assumes that an individual will remain eligible in one

of the two states involved, and is based only on the five match code combinations that are almost certain

to represent actual dual participation. Perhaps most importantly, though, the model reflects the business

processes in place during the pilot period; as these are perfected, the rate of prevention (and savings) is

expected to improve.

Start-up costs

The ongoing expenses reported by pilot states and used in the preceding cost savings analysis do not

include the one-time costs associated with initial start-up. Pilot states were also surveyed regarding the

resources expended prior to NAC operations, and results are presented in Table 16. Details of state

responses are included in Appendix F.

The question specific to NAC start-up costs instructed states to

include costs associated with programming, staff training, file

preparation, and planning. They were asked to exclude costs

specifically related to the evaluation, as those do not reflect what

other states could be expected to spend if the project is expanded

beyond the five consortium states.

One comment included in the survey from Florida noted that

expenditures would likely have been higher had it not been for other

systems-related priorities. Because the NAC project was started during the development of a new

Medicaid eligibility system to support the Affordable Care Act, the state was limited in its ability to create

a more integrated process with the existing SNAP eligibility system.

Potential Impact of Expansion

The cost/savings analysis estimates only the savings the NAC pilot states can expect if no other states join

the consortium. However, as NAC expansion has been explored, two additional questions related to its

potential impact have been posed:

1) How much would pilot states save if they matched with all 50 states; and

2) How much would the program as a whole save if all states participated?

Table 16

NAC Start-up Costs

State

Start-up costs

Alabama

$29,200

Florida

$147,019

Georgia

$35,557

Louisiana

$127,555

Mississippi

$330,000

Average

$133,866

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Savings if pilot states matched with all fifty states

The five pilot states are contiguous but vary significantly in geography and population. They also

experience differing migration patterns from states across the country. Presumably, some pilot states will

experience more frequent dual participation with states outside the current consortium than others.

Data from PARIS matches, which are conducted with nearly all other states, was utilized to estimate the

percentage of all potential savings represented by the other four pilot states. This percentage was then

applied to the savings identified in the above analysis to estimate how nationwide implementation would

translate into savings for each pilot state.

Limitations

The use of PARIS data to extrapolate the prevalence of dual participation involving the pilot and non-pilot

states has several limitations. Optimally, a proxy would be available that allows for an “apples to apples”

comparison. For example, if all states matched all their SNAP participants in PARIS and no other programs,

the distribution of matches could be reasonably expected to be similar for the NAC. Unfortunately, this

is not the case – two primary factors impact the reliability of this method:

 As noted in the Altarum study, some States have reduced the number of SSNs submitted to PARIS

by selecting only individuals that meet certain criteria, or selecting only individuals active in the

final month of the quarter for which the match is being conducted.

 Not all states that do participate in PARIS submit data from the same programs. For example,

California may submit only their eligible SNAP population, while Illinois may submit both the SNAP

and Medicaid population. This results in a higher percentage of all matches occurring with Illinois

than if California also provided their Medicaid population.

These limitations are not insignificant. To provide at least a partial test of validity of this approach, the

distribution of PARIS matches within the pilot states only was compared to the breakdown of NAC matches

at the Big Bang.

Table 17

Comparison of Match Distribution, PARIS and NAC

PARIS

NAC BIG BANG

# (TOP 5)

2341

43.9

422

39.9

2131

40.0

431

40.8

359

6.7

5.7

501

9.4

144

13.6

TOTAL

5332

100%

1057

100%

PARIS

NAC BIG BANG

# (TOP 5)

2514

15.2%

422

17.8%

12571

75.9%

1639

69.1%

829

5.0%

161

6.8%

644

3.9%

151

6.4%

TOTAL

16558

100%

2373

100%

PARIS

NAC BIG BANG

# (TOP 5)

2131

16.2%

431

18.7%

9594

72.9%

1639

71.0%

999

7.6%

137

5.9%

430

3.3%

103

4.5%

TOTAL

13154

100%

2310

100%

PARIS

NAC BIG BANG

# (TOP 5)

359

10.2%

12.0%

1176

33.4%

161

32.1%

999

28.4%

137

27.3%

982

27.9%

143

28.5%

TOTAL

3516

100%

501

100%

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

PARIS

NAC BIG BANG

# (TOP 5)

501

19.5%

144

26.6%

657

25.6%

151

27.9%

430

16.7%

103

19.0%

982

38.2%

143

26.4%

TOTAL

2570

100%

541

100%

With the exception of Mississippi, the distribution of PARIS matches and NAC matches in the months

analyzed were quite similar. This suggests that, at least within the pilot states, the relative

frequency/volume of NAC matches mirrors the distribution of PARIS matches.

Step 1: Percentage of all potential savings represented by the other four pilot states.

PARIS data from March 2013 (for AL, GA, LA, and MS) and November 2014 (FL) was used to calculate the

percentages in Step 1. In calculating the total number of PARIS matches, intra-state matches were

excluded, and the number of matches generated for each pilot states with the other consortium members

was divided by the total number of interstate matches.

PARIS matches with NAC states

5,332

16,558

13,154

3,516

2,570

PARIS matches with all states

10,165

87,502

34,183

11,320

7,433

(# of states w/ PARIS matches)

NAC state %

52.5%

18.9%

38.5%

31.1%

34.6%

Non-NAC state %

47.5%

81.1%

61.5%

68.9%

65.4%

Despite the limitations noted above, the percentages calculated in Step 1 align with anecdotal

assumptions:

 Florida would have the highest percentage of matches with states outside the NAC pilot (81.1)

due to the large number of individuals traveling there from far outside the pilot areas for the

winter and/or tourism; and

 Of the other four states, the one with the fewest non-NAC border states (Alabama, with one)

would have the lowest percentage of matches outside the NAC pilot area (47.5).

Step 2: Annual savings for each pilot state if the NAC were nationwide.

To calculate the annual net savings if the NAC were nationwide, the net savings for each state is divided

by the percentage of all savings represented by the NAC states only (calculated in Step 1).

Annual Net Savings–pilot states only

$892,360

$1,870,616

$1,653,396

$307,920

$872,792

NAC state %

52.5%

18.9%

38.5%

31.1%

34.6%

Annual savings if NAC were nationwide

$1,699,733

$9,897,439

$4,294,535

$990,095

$2,522,250

If some states did not participate, this only would increase the impact of going nationwide. As noted earlier, not

all states submit the same programs to PARIS or all active participants for the programs they do submit.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Overall SNAP savings if NAC was expanded nationwide

The final component of the cost/savings analysis is to extrapolate the program-wide SNAP savings realized

if all states participated in the NAC. For this calculation, a range of savings is estimated based on the

percentage of each state’s total annual SNAP allotment comprised by the savings calculated above. The

same caveats noted above in calculating individual state savings if the NAC were nationwide apply to this

calculation.

Step 1: State savings as a percentage of all SNAP allotments in FY 14.

As noted throughout the report, the pilot states have achieved different degrees of success in utilizing the

NAC. This is underscored when the savings calculated in the previous step are compared to each state’s

total SNAP allotments.

Annual savings if NAC

were nationwide

$1,699,733

$9,897,439

$4,294,535

$990,095

$2,522,250

SNAP allotments, FY 14

$1,318,133,562

$5,472,834,001

$2,827,853,876

$1,288,316,273

$912,985,504

% of allotments saved

0.13%

0.18%

0.15%

0.08%

0.28%

Savings range from less than one-tenth of one percent to nearly three-tenths of one percent of each

state’s total SNAP allotments.

Step 2: Program-wide savings

Total SNAP allotments across the program totaled $69,999,805,422 in FY 14. When the percentages

calculated above ranging from .08% to .28% are applied to total allotments, the potential savings vary

from $53.8 to $193.4 million annually, and average more than $114 million. The significant variations

represented by the estimates in Table 18 underscore the uncertainty of the precise impact if the NAC

were expanded and utilized by all fifty states.

Table 18

Estimated Annual Program Savings, Nationwide NAC Implementation

If overpayments are avoided

on “x” % of all allotments…

Nationwide savings =

0.08%

$53,796,169

0.13%

$90,264,747

0.15%

$106,305,562

0.18%

$126,592,331

0.28%

$193,404,957

AVERAGE

$114,072,753

The average estimate of $114 million is equivalent to .16% of the total SNAP allotments of $69 billion.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Recommendations

The five NAC pilot states have implemented the NAC in significantly differently ways; and have realized

different levels of success. States that have achieved superior outcomes provide a set of best practices

that should be considered as use of the NAC continues in the current states and as expansion beyond the

pilot is explored. Furthermore, the lessons learned by the five pilot states should be heeded by any state

– using or intending to use the NAC. And, the evaluation has identified trends in and characteristics of

dual participation that can support decision-making going forward.

The recommendations below cover a range of areas and are grounded in both the quantitative analyses,

observations, and qualitative information included here and in previous reports. They are organized in

three general categories:

 General Recommendations for All States;

 Expansion-Specific Recommendations; and

 Opportunities for Improvement.

General Recommendations for All States

Automate to the greatest extent possible. Both the statistical evidence and comments from pilot states

demonstrate the importance of automating NAC processes to the greatest extent possible. Options

related to automation include the following:

 Integration of the NAC with the state’s SNAP eligibility system. This is a critical component. Statistics

cited previously relative to the overall prevalence of dual participation indicate that matches will be

generated by the NAC on only a small percentage of applications and new household members. Given

the need to streamline eligibility processes and achieve business process efficiencies, caseworkers

should not be asked to check the NAC portal on every application they process and every person they

add to a case. Instead, NAC data should be integrated into the existing workflow, flagging a

caseworker to take additional steps only in the event a match is produced.

 Use of web services to optimize real-time automation. Web services, which provides states with a

real-time link to the NAC, provides a “best of both worlds” model, in which the NAC is queried in near

real-time in a manner similar to a manual portal query, with the added advantage of limiting

caseworker intervention to only those instances in which a match is generated. For example, if a

caseworker needs to process as application on the same day the application is registered, the web

services concept allows for sending and receiving information from the NAC that same day. In a batch

process model, the return information would not be returned until the following day.

 Automation of emails to the matching state. The level of certainty for match code combinations 1-5

is such that additional worker investigation regarding its validity is not necessary. Accordingly, an

automated email to the matching state eliminates the need for a caseworker to draft and send the

email and ensures that contact is made. Furthermore, the email format can be standardized with all

necessary information for the matching state to take appropriate action upon receipt.

 System edits that require resolution of a match before authorization. To further ensure that staff act

on matches that may impact eligibility, system edits can be programmed that force the user to enter

a resolution code prior to authorization of benefits. These codes also have potential to be used to

track outcomes if the resolution options include an indicator of whether the match was determined

as valid or not.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Establish processes for utilizing the NAC when individuals are added to an open SNAP case. SNAP

eligibility interviews conducted when a new application is processed include standard questions regarding

SNAP receipt in another state. And, the application itself includes this question as well. However, the

same procedures may not be followed when a new individual is added to an existing SNAP case – for

instance, an application is not required when adding a new household member. And, the business

processes in place in Georgia and Florida suggest that the addition of new household members is likely

contributing to those states’ lower rates of prevention success.

Although data collection for the evaluation did not specifically address the percentage of individuals

whose dual participation began when they were added to an existing case, a similar scenario was

identified that provides some insight. This analysis calculates the percentage of dual participants (at the

Big Bang) who were not part of households in which all members were also dual participants.

Statistics

are shown for each two-state combination and totaled.

Table 19

Percentage of Dual Participants Residing in Households in which all Members Are Not Dual Participants

Dual

participants

# residing in HHs where

all members are not

dual participants

AL-FL

418

184

44%

AL-GA

431

262

61%

AL-LA

40%

AL-MS

144

56%

FL-GA

1636

688

42%

FL-LA

160

34%

FL-MS

151

42%

GA-LA

137

61%

GA-MS

103

59%

LA-MS

141

59%

TOTAL

3381

1586

47%

When 100% of SNAP household members are active in two states, it is presumed that the household

moved as a unit from state-to-state and standard application processes were conducted. However, if all

household members are not dual participants – which was the case for almost half (47%) of the individuals

in this analysis - it is more likely that the situation was different. While not a definitive indicator of the

circumstances involving an individual’s approval for SNAP in a second state, these statistics suggest that

a significant percentage of dual participation occurs when a new member is added to an existing case.

Accordingly, states should institute a robust process for using the NAC in those situations. The web

services model described above would be optimal; in the absence of that solution, worker access to the

NAC portal is an alternative to be considered.

Data sources: Big Bang administrative data (# of household members who were dual participants) and information

collected by pilot states (SNAP household size). Also, note that household with one member are included in

calculating the number of cases in which all household members are dual participants.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Treat “Social Security Number-only” matches differently.

Matches are characterized as “SSN only” if the Social Security Number is the only data element that

matches between eligible SNAP participants–name, date of birth, and all other demographic information

are dissimilar. Table 20 captures the state-determined explanations for the Big Bang matches that fell into

this category.

Table 20

Explanation for SSN-Only Matches Identified at the Big Bang

Data entry error

54.8%

18.8%

32.4%

34.0%

SSN verified correct

39.9%

76.3%

58.9%

56.0%

100%

Other

5.3%

4.8%

8.3%

10.0%

Blank/TBD

0.4%

A data entry error typically means that a caseworker transposed one digit when entering a SSN into the

state’s eligibility system; if the SSN has been verified correct, documentation provided by the individual

or head of household, such as a Social Security Card, confirmed the number entered in the state’s

eligibility system as belonging to that individual.

As a lesson learned for future NAC states, an internal review of SSN-only matches to confirm that a data

entry error was not the cause is advised prior to making contact with the other state involved (or the

client). In the event both states confirm the accuracy of the SSN, further investigation will be necessary.

In a few isolated cases, states discovered that the Social Security Administration had issued identical SSNs

to two different people.

Establish processes for addressing dual participation when prevention is not possible or

unsuccessful. For reasons of access or process failure, dual participation will still occur even with the

availability of the NAC. Although prevention of dual participation is the primary goal of the Clearinghouse,

it also provides information necessary to shorten spells of dual participation when they do occur.

 Designate a task force to act on dual participation. Because an overpayment has occurred in these

situations, it may be efficient to integrate the task into a centralized state claims/investigations unit

to limit duplication of efforts.

 Automate MSH reports for delivery to entity responsible for action. By having the MSH report be

automatically generated as an email/report to the user group that will be following up, it helps remind

the worker of the match instead of them having to manually run the report.

Expansion-Specific Recommendations

In the event additional states join the consortium and begin submitting contributory files to the NAC,

several lessons from project start-up and the pilot’s Big Bang should be applied. In addition, the increased

volume of work associated with new states taking part in the NAC calls for a level of uniformity that is not

as essential when only five states are involved.

Develop separate processes for the initial match and ongoing operations. The best practices

described above regarding automation will support effective prevention of dual participation as new

states are added. However, the addition of each new state will be accompanied by an initial match of dual

participants – a “mini-bang.” These events will impact not only the new NAC state/s, but those already

participating. The following approaches to addressing subsequent initial collisions are recommended:

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

 Establish a task force to work initial collisions. The volume of time and effort required to address

existing dual participation is considerable, and when the NAC is implemented caseworkers’ focus

should be on prevention. Both the newly-entering and existing NAC states should designate a specific

team – staffed appropriately – to handle initial collisions. This supports tracking of progress, more

consistent communication with other states, and a clearer structure for identifying the need for

establishment of overpayment claims.

 Make preparations for staff to establish overpayment claims. The first time contributory data from

a new NAC state is submitted for matching, a significant number of collisions is to be expected. And,

given that these collisions will represent possible overpayments, capacity of staff responsible for

establishing claims will be strained.

Table 21 documents the volume, types, and value of the claims established on cases in which an

individual/s was identified as a dual participant at the Big Bang.

Table 21

Claims Data on Cases Including Dual Participants Identified at the Big Bang

TOTAL

Agency Error

Inadvertent Client Error

354

544

Intentional Program Violation

133

810

475

114

1577

No claim

280

123

574

TBD/Blank

113

172

TOTAL CLAIMS CALCULATED

233

1181

484

126

163

2187

TOTAL $

$292,538

$1,908,604

$1,836,089

$193,728

$209,700

$4,440,659

AVERAGE

$1,256

$1,616

$3,794

$1,538

$1,287

$2,030

 Only send automated letters initially to high-confidence match combos. Approximately 85 percent

of matches (excluding “SSN-only) are comprised of high-confidence match code combinations–in

those instances, a contact letter to the head of household notifying them of apparent dual

participation is appropriate. However, for the other 15 percent of matches, states should conduct a

review of the information provided prior to making contact with the client

Conduct comprehensive front line staff training. Effective training of front line staff is essential to

effective project implementation. Resources should be dedicated to the delivery of training, and the

materials developed should give caseworkers real-world examples and/or hand-on experience in the

approach the state will use to operationalize the tool and communicate with other states. In addition,

training should provide background on the NAC and its potential benefits to promote buy-in from staff

that may view it as just another task to add to their many responsibilities.

Recognize and address connectivity and IP address issues as early as possible. Some pilot states

experienced problems with access to the NAC because they were blocked by the NAC’s firewall. In

November 2014, some county offices in Alabama were unable to access the NAC portal due to a network

translation change; Georgia’s efforts were hampered due to difficulty in collecting all the public IP

addresses used for system access across the state; and, Florida reported problems related to the large

number of IP addresses utilized by home-based caseworkers. While these issues were rectified, states

The value of Georgia’s claims were expected to be reduced following additional investigation.

This also applies to automated emails generated as part of the prevention process.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

preparing to join the consortium would be well-served to begin identification of all home-based staff and

take other steps to identify public IP addresses as early as possible to prevent similar barriers to NAC

access.

Implement business process standardizations. While it is important to allow states the flexibility to

develop their own procedures with respect to some aspects of the NAC, there is additional standardization

that could support effective use of the tool, particularly if additional states join the project.

 Consistent treatment of match code combinations. States have implemented similar, but not

identical, criteria for the treatment of matches depending on their strength. For consistency of actions

across the program and clear communication between states, a standard definition of the matches

considered to be valid without further investigation is recommended.

 Timely submission of contributory files. Once a month, the daily contributory files submitted by the

states include the active SNAP recipients for the next recurring month. During the pilot there was

inconsistency across states with regard to the timing of this submission. In some cases, late submission

of these files resulted in missed opportunities to prevent dual participation. The date of this recurring

file may be dependent on factors outside the control of staff responsible for NAC implementation;

however, to the extent possible, a consistent date of submission, preferably as early as possible prior

to the benefit month in question, should be established.

 Common naming conventions for email addresses used for interstate communication. (e.g.

[email protected]). If additional states are added to the NAC, each state will also be establishing a mailbox

for NAC-related communication. A common convention including the “NAC” and the state name

would simplify interstate emails.

 Standard protocols for transmission of personally identifiable information (PII). States were not

consistent in the conventions used for transmission of client information when communicating via

email about a NAC match or collision. A standard process describing the security protocols should be

established for states exchanging PII in NAC-related communications.

Opportunities for Improvement

Apart from the best practices identified that will help current and future NAC states, some areas of

improvement are apparent that, if addressed, could optimize use of the NAC.

Reconsider the business process change to require notification of closure/removal. Originally, the

common business processes included a rule designed to limit the frequency of communication between

the two states involved in a match generated by the NAC. This rule dictated that states would assume

communication delivered by the initiating state prior to the 15th of a month would result in action being

taken by the matching state in time to impact the next month’s SNAP benefits. And, because this

assumption was in place, no return email from the matching state would be needed.

However, pilot states identified examples in which the matching state did not take action timely – it was

not always safe to assume that an individual would be removed from a case or a case would be closed

prior to benefit issuance for the following month. This resulted in dual participation for the individual in

question. Following discussion, the pilot states decided that action in the initiating state would not be

taken under the assumption that the matching state had acted timely, and the business rules were

modified.

Under the new rules, an email from the matching state must be provided to the initiating state confirming

the action taken and the effective date, regardless of the date the request was made by the initiating

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

state. While this modification is understandable given states’ early experiences, the policy should be

reviewed by NAC states on a regular basis to determine if the original concept could be re-instituted.

Consider wider use of Activity Type 3 (passive matches). States receive this information when an

individual who is active in their state (the matching state) has been submitted to the NAC by another

state, presumably because an application is being submitted there. Typically, pilot states have waited for

an email from the initiating state to take action on their case/individual. However, states should explore

using the passive matches to inquire about the household situation, as it may allow action to be taken

more timely than it would be otherwise.

For example, if Alabama received an Activity Type 3 match showing that an individual active there has

been queried by Florida, this indicates that Florida had a reason to inquire about the individual’s eligibility.

Typically, Alabama would wait and for contact from Florida asking for action to be taken. However, the

receipt of the Activity Type 3 match allows Alabama to proactively inquire about the individual’s residence

there. This gives the state a better opportunity to meet the two-day turnaround standard documented

in the common business rules.

_______________________

The overarching conclusions of the National Accuracy Clearinghouse (NAC) Evaluation may be summarized

as follows:

1. The NAC has supported a decrease in dual SNAP participation;

2. When implemented using the best practices outlined above the NAC supports the prevention of

dual participation in a manner and scope not possible with the resources previously available;

and

3. Despite the relatively infrequent occurrence of dual participation, the prevention made possible

by the NAC translates into significant SNAP overpayment avoidance.

The findings suggest serious consideration should be given to expansion of the project.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Glossary

Active Dual Participation Caseload – The number of dual participants in a state during a given month.

During the pilot period, calculated by adding “Entries” and “Continuations” (see below).

Big Bang – The process of identifying and acting on all apparent instances of current dual participation

identified at the point the NAC database became available to the pilot states in June 2014. The term was

coined as a way to differentiate activities associated with the initial implementation of the NAC from the

ongoing use of the tool.

Collision – An instance in which dual participation–receipt of SNAP or D-SNAP in two or more states in the

same month - appears to have already occurred per the NAC. The term refers to both individuals identified

at the “Big Bang” and those coded as “Activity Type 4” on the Match Search History (MSH) file (see

definitions below).

Continuations – Collisions that are present in both the target month and were present in the previous

month.

Entries – Collisions that are present in the target month but were not present in the previous month.

Exits – Collisions that were present in the previous month but not the target month

LexID – The NAC leverages the LexisNexis LexIDSM technology to provide identity resolution by comparing

input information provided by participating states across billions of unique public records. These public

records, linked together using the LexID, provide context to an identity, such as how it has changed over

time (moving addresses, name changes, marriage/divorce, etc.) or where input information is missing or

incorrect. Through identity analytics, LexisNexis is able to resolve input information to a single individual

with an high degree of accuracy. The LexID is a unique, 12-digit identifier assigned after a successful

identity resolution.

Match – An instance in which a state identifies via the NAC that an individual is already receiving SNAP or

D-SNAP benefits in another state; generally in the context of dual participation prevention (prior to

approval of benefits in a second state). When states “work” a match, the following designations are used:

 Initiating State–Refers to the state in which the individual is applying and has not yet been

approved for SNAP.

 Matching State–Refers to the state in which the individual is already receiving SNAP benefits.

Match Search History (MSH) file – a daily file created for each of the five pilot states and comprised of

four “Activity Types,” documenting the previous days’ NAC activity for four types of actions:

1. Single (Activity Type 1): a record of requests (and resulting matches if applicable) for match

information made by states via NAC portal queries or via state eligibility systems connected to

the NAC by web service

2. Batch (Activity Type 2): a record of requests (and resulting matches if applicable) for match

information made via batch process

3. Passive (Activity Type 3): a notification informing a “Matching“ state that an “Initiating” state

conducted a search that generated a match

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

4. Build-Time Collisions (Activity Type 4): A record of new collisions

PARIS – Public Assistance Reporting Information System. PARIS is the data matching system administered

by the Administration for Children and Families (ACF) within the Department of Health and Human

Services (HHS). States submit data to PARIS (voluntarily) on active participants in several programs,

including SNAP, on a quarterly basis.

Spell – Refers to a period of time during which dual participation occurred. For example, for someone

who is eligible in both Alabama and Florida from January to April of 2015, a four month spell of dual

participation occurred.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Appendix

A - Common Business Rules

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

B – Methodology – Percentage of Matches that Became Collisions

For each month that is part of this analysis, two separate files were created using information found in

the daily Match Search History (MSH) files:

 Match file–All MSH Activity Type 1 matches (generated through portal and web services queries)

and Activity Type 2 matches (generated through the nightly batch process) comprising one

calendar month (the target month) were compiled.

o Duplicate matches were removed using Search SSN as the primary identifier. This was

necessary to prevent double-counting of individuals identified through both a batch

match and a portal query.

 Collision file –All Activity Type 4 collisions (generated by the nightly NAC database build) on MSH

files were compiled for a four month period including the target month, the month prior, and the

two following months (X, X-1, X+1, X+2, respectively). Additional months of collision files were

utilized because a collision could be prevented for past or future benefits months.

o The MSH files selected for each month ranged from the 2nd day of the month until the

1st day of the following month. For example, to create the collision file for March, MSH

files from 03/02/2015–04/01/2015 were utilized, because collisions do not appear until

the day following when benefits were authorized in the second approving state.

The match and collisions data were then compared to determine if an individual on the match file also

appeared as a collision in any of the four months (X, X-1, X+1, X+2). If so, a “preventable collision”

occurred. If an individual appeared in more than one of the four months, only one instance of a

preventable collision was counted.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

C – Big Bang Statistics Organized by Top 5, 6+ and SSN-only

This summary of possible dual participation at the Big Bang separates the collisions that the NAC identified on June 2, 2014 into four categories:

 Total collisions;

 SSN only–collisions in which the only common data element is the Social Security Number;

 Top 5–collisions identified through the five strongest match code combinations; and

 6+ - All other match code combinations.

Alabama

Florida

Georgia

Louisiana

Mississippi

Total

SSN

only

Top 5

Total

SSN

only

Top 5

Total

SSN

only

Top 5

Total

SSN

only

Top 5

Total

SSN

only

Top 5

601

422

115

647

431

131

211

144

601

422

115

2046

107

1639

300

208

161

194

151

647

431

131

2046

107

1639

300

227

137

147

103

208

161

227

137

202

143

211

144

194

151

147

103

202

143

TOTAL

1554

193

1057

304

3049

189

2373

487

3067

254

2310

503

732

101

501

130

754

541

154

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

D – Process for Using the NAC for Prevention of Dual Participation as of May 31, 2015

State

Non-expedited applications

Expedited applications requiring

same-day processing

Adding individual to

existing case

Alabama

 All applicants are submitted to the NAC via the Match Request File (MRF)

on the night of application registration.

 The Match Request Response (MRR) that is returned overnight is imported

to the state’s eligibility system (The Online Application and Case

Information System, or “OACIS”) to alert the caseworker conducting the

interview of the NAC hit. Caseworkers investigate and determine if email

to other state is warranted.

 In addition to emailing the other state, many caseworkers also ask the

client to provide written proof of closure

 Action on application is pended until email from other state is received or

applicant provides documentation of closure in other state.

County staff query

NAC portal

prior to approval of the

application.

Applicant’s status is verified prior

to the approval of the application

if possible; if ID has been verified,

postponed verification policy is

utilized in the rare occasion that

match appears to be accurate but

applicant is insistent that they

aren’t active in other state.

OACIS reminds worker to

query NAC if taking same-

day action to add

individual (this process

occurs in majority of

instances); if not

authorizing same day,

individual is submitted to

NAC via overnight batch.

Florida

 All applicants are submitted to the NAC via the MRF on the night of

application registration.

 Workers are notified of any NAC matches returned via the MRR via a

message (“Customer is receiving benefits in another state”) on the Work

Item Detail page the next morning before 7 AM.

 No additional information is provided to the worker in the message. When

the alert is received on the Work Item Detail page, staff must complete a

search on the NAC portal for all individuals on the application prior to

SNAP authorization for the current month.

 Staff has flexibility to either follow the common business rules (i.e. email

the matching state) AND/OR ask applicant for a closure letter.

 Applicants receive a letter notifying them of the window they have to call

for their interview (which may be with a clerk or caseworker); this letter

can include the request for proof of closure as part of the list of

information needed to determine eligibility.

There is no process in place to

check the NAC if disposition on an

application is taken prior to

processing of the overnight batch.

New individuals added to

active cases are sent

through the overnight

batch; same process as

non-expedited application

occurs.

Overnight batch process does not support dual participation prevention when state needs to take action on case the day application is registered.

At least one state (Florida) has reported an apparent increase in clients requesting closure through the online “my account,” citing the reason as applying in

another state.

Because states provide updates on active SNAP recipients to the NAC database on a nightly basis, this query provides near real-time information on the

individual’s status in the other pilot states.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

State

Non-expedited applications

Expedited applications requiring

same-day processing

Adding individual to

existing case

Georgia

 All applicants are submitted to the NAC via batch on the night of

application registration.

A report is generated from the return file of matched individuals and is worked

by a team of three claims managers. Team members send the contact to the

email address for the other state. When the response is returned, the claims

manager documents the hit information and that the match has been

addressed directly in the case record. The process is designed to work as

follows:

 The initial contact with the other state should go out within 24 hours of

receipt and the return response is due within 48 hours of notification.

 The expectation is that the documentation be in the case no later than 5

days after being notified of the hit.

With the current process in place in Georgia for applicants, cases are registered

and assigned to workers for an interview. Expedited cases are interviewed

usually no later than the 5th day following application. In this process, it is

expected that expedited cases may be missed for one month but most non-

expedited cases will be documented and the caseworker aware of the match

before the initial month’s benefits are issued. Caseworkers do not have access

to the portal. A fourth worker was added to the process at the end of

December 2014 to further ensure that caseworkers are notified and cases are

documented timely.

GA is building a new eligibility system which will include an indicator generated

by a NAC match, but implementation is at least one year away (mid to late

2016).

Caseworkers act on expedited

cases to meet timeliness rules;

same process occurs as for non-

expedited applicants who

generate matches (i.e. caseworker

typically does not become aware

of the match until after benefits

have been authorized).

Individuals added to an

active case are submitted

through the nightly batch;

same process occurs as for

non-expedited applicants

who generate matches

(i.e. caseworker typically

does not become aware of

the match until after

benefits have been

authorized).

Louisiana

 All applicants are submitted to the NAC via web service on the night of

application registration.

 Matches are loaded daily to Clearance Summary screen (outside of the

eligibility system) for review/action by caseworkers in the Parishes.

Caseworkers receive all pertinent data from the Clearance Summary

screen and do not have access to the NAC via the portal. It is the

responsibility of caseworkers to identify if the match is qualified per state

business rules.

Limited # of staff (Regional

Program Consultants) have portal

access to conduct searches when

an expedited application is to be

approved before the overnight

web service process runs.

Caseworkers forward relevant

information to their supervisor

who requests that the query be

conducted.

Limited # of staff (Regional

Program Consultants) have

portal access to conduct

searches when an

individual is being added

to an existing case.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

State

Non-expedited applications

Expedited applications requiring

same-day processing

Adding individual to

existing case

Caseworkers forward

relevant information to

their supervisor who

requests that the query be

conducted

Mississippi

 The overnight batch process was utilized from 6/1/2014 through

11/30/2014; on 12/1/2014, Web Services was implemented statewide

which provides for a real-time response (MRR) to each search request

(MRF).

 Response received from NAC is integrated into the eligibility system and is

populated on a screen that the worker is directed to if there is a match.

o System edits prevent authorization of the case if this screen has

not been handled by the worker.

 For “level one” matches (see discussion of match codes) the system

generates an email to the matching state/s where a match has been found,

requesting response within 48 hours per business rules.

o Return emails from other state are directed to a centralized unit

that in turn notifies the caseworker of necessary action.

 For “level two” matches, caseworkers evaluate the information and have

discretion regarding whether an email to the other state is needed.

 On 10/31/14, MS modified the email subject line to include the MS State

Code and Full Service Office (County) to improve email identification when

received from Matching State/s; modified the body of the email to remove

the Matching State/s' individual Client ID and DOB to prevent any possible

identity theft when emailed; modified body of email to match revised NAC

business language; modified emails to be grouped as one email for all

matching individuals by the Matching State/s Case ID rather than

individual emails each day which improves the Matching State/s' email

processing.

Same process as Non-expedited

applications with one exception:

The Match State/s will be

contacted via phone, if necessary,

to ensure timely processing of the

expedited application.

Same process as new

application.

Louisiana’s processes are designed so as not to burden workers with unproductive matches that occur when newborns who have not yet been issued a SSN

are added to a case. Not only is dual participation extremely unlikely among this population, but searches conducted without a SSN do not meet the threshold

the state has established for a valid match.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

E – Matching State Process for Responding to Request from Initiating State as of May 31, 2015

State

Process

Alabama

State office forwards the email to the applicable county. Supervisor or workers in the county sends email to other state informing them of the

action taken and the effective date in Alabama. The county is responsible for the contact to the other state. The county does not routinely cc the

state office.

Florida

A NAC email has been set up for all inquiries from other states. This email is directed to the Quality Control unit who will take action and respond to

the other state.

I. If match includes all household members of the case in FL, worker closes case effective the next possible month.

II. If match does not include all members of the case in FL, the worker will remove the member and notify the other state that the individual

has been removed with effective date.

Georgia

Emails from the Initiating State are currently received by 2 claims staff. Responses are sent back to the initiating state within the 48 hour deadline

when possible.

Louisiana

Emails from the Initiating State are received by the Inquiry Services Section; personnel in that unit check to see who is responsible for the case and

forward it to the applicable parish.

Mississippi

A centralized unit within the state office (3-4 persons) receives the emails and forwards them to the applicable county for action.

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

F – Results of State Survey for Cost/Savings Analysis

Ongoing Costs

Work Effort as Initiating State

On average, how much time (in

minutes) does a staff person

spend acting on each NAC

match?

10-15

minutes

0.21

5-7

minutes

0.1

approx.

minutes

0.33

15-30

minutes

0.38

< 1

minute

0.02

Work Effort as Matching State

On average, how much time (in

minutes) does a staff person

spend acting on each NAC

inquiry received from another

pilot state?

10-15

minutes

0.21

7-10

minutes

0.14

10-15

minutes

average

0.21

minutes +

0.5

< 5

minutes

0.08

Personnel Costs

Salary

$3,000

$2,564

$2,250

$3,400

$2,114

Benefits:

$700

$1,275

$1,350.90

$1,600

Total divided by 172 work

hours/month

$21.51

$22.31

$20.94

$29.07

$14.87

Other Ongoing Costs

Approximately how much does

your state incur in ongoing

monthly costs associated with

the NAC that are not captured

in the salaries of the staff

acting on matches?

IT Developer Salary: 40

hours@$50/hour

Data storage:

$312.50/month

IT staff salary: $1,134

None reported

"unable to quantify"

none quantified other

than hosting (L/N costs)

$8,600.00

$1,446.50

$0.00

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

Start-up Costs

State

Staff Costs

Non-staff Costs

TOTAL

Alabama

 Project Manager: 180 hours @ $40/hour = $7,200

 Business Analyst: 100 hours @ $35/hour = $3,500

 IT Developer/Systems Analyst: 160 hours @ $50/hour = $8,000

 QA/Testing: 80 hours @ $25/hour = $2,000

 Training staff: 40 hours @ $25/hour = $1,000

New software to support system

integration: $7,500

$29,200

Florida

 Project Manager: 547.5 hours @$126.79/hour = $69,417.52

 Business Analyst: 190.5 hors @$22.43/hour = $4,272.93

 Business Analyst: 21 hours @$25.08/hour = $526.68

 IT/Developer/Systems testing: 847 hours @ $84.89/hour = $71,901.83

 Training staff: 30 hours @$30/hour = $900.00

None reported

$147,019

Georgia

 $35,557 in costs billed to NAC project

$35,557

Louisiana

 State IT Staff: 315.5 hours @ $30/hour = $9,465

 State Program Staff: 134 hours @ $30/hour = $4,020

 Contractor IT: 1,342 hours @ $85/hours = $114,070

None reported

$127,555

Mississippi

 IT-Project Management: 1,333 hours @ $90/hour = $120,000

 IT-Business Analysis: 1,067 hours @ $75/hour = $80,000

 IT-Development: 1,143 hours @ $70/hour = $80,000

 Program (Policy; Training; etc.): 1,429 hours @ $35/hour = $50,000

None reported

$330,000

National Accuracy Clearinghouse (NAC) Evaluation | Final Report

October 2015

G – Percentage of Household Size Methodology

HH size

Average

monthly benefit

FY 2011

Increase from

allotment

Fraction of

allotment for last

individual added*

Fraction used when

100% of HH are

dual participants**

$153.00

$272.00

$119.00

0.438

0.500

$397.00

$125.00

0.315

0.333

$489.00

$92.00

0.188

0.250

$579.00

$90.00

0.155

0.200

$675.00

$96.00

0.142

0.167

$782.00

$107.00

0.137

0.143

*Referred to as “Methodology 2”

**Referred to as “Methodology 1”

USDA

iiillm

United States Department of Agriculture

Y 1 0

2016

The Honorable Pat Roberts

Chair

Office

the Secretary

Washington, D.C. 20250

Committee on

culture, Nutrition, and Forestry

United States Senate

328A Russell Senate Office Building

Washington, D.C. 20510

Dear Mr. Chairman:

Section 4032(c)

the Agricultural Act

of20

directed the U.S. Department

Agriculture

(USDA) Secretary to submit a report not later than 90 days after completion to the Committee on

Agriculture

the House

Representatives, and the Committee on Agriculture, Nutrition, and

Forestry

the Senate on the pilot program to test prevention

duplicate participation

the

Supplemental Nutrition Assistance Program (SNAP). The report is to assess t

feasibility,

effectiveness, and cost for the expansion

the pilot program nationwid

The enclosed

evaluation report on the National Accuracy Clearing House (NAC), prepared by Public

Consulting Group, Inc., for the State

ssissip

, fulfills this requirement.

In 2011, USDA's Food and Nutrition Service was awarded $2.5 million by the Office

Management and Budget Partnership Fund for Program Integrity Innovation, with the goal

reducing improper payments that occur due to dual participation in SNAP. This grant funded the

development

the

NAC

- a searchable database to support near real-time sharing

eligibility

information among States. Subsequently, Mississippi was awarded the funding to lead the

project on behalf

a consortium

contiguous States (also including Alabama, Florida,

Georgia, and Louisiana). The pilot operated for

months from June 2014 through May 2015.

The enclosed evaluation report assesses the technical capacity

the NAC, States' success in

utilizing the tool and implementing the accompanying business rules, and the cost savings

associated with adoption

the NAC solution.

you have any questions, please have a member

your staff contact Todd Batta, Assistant

Secretary for Congressional Relations, at (202) 720-7095. A similar letter and a copy

the

report are being sent to Ranking Member Stabenow, Chairman Conaway, and Ranking Member

Peterson.

Sincerely,

~<}

~l-

Thomas

..(;;{

Secretary

Enclosure

Equal

Opportunity

E~r

USDA

United States Department

Agriculture

MAY

1 0

2016

The Honorable Debbie Stabenow

Ranking Member

Office of the Secretary

Washington, O.C. 20250

Committee on Agriculture, Nutrition, and Forestry

United States Senate

328A Russell Senate Office Building

Washington, D.C. 20510

Dear Senator Stabenow:

Section 4032(c)

the Agricultural Act

of2014

directed the U.S. Department

Agriculture

(USDA) Secretary to submit a report not later than 90 days after completion to the Committee on

Agriculture

the House

Representatives, and the Committee on Agriculture, Nutrition, and

Forestry

the Senate on the pilot program to test prevention

duplicate participation in the

Supplemental Nutrition Assistance Program (SNAP). The report is to assess the feasibility,

effectiveness, and cost for the expansion

the pilot program nationwide. The enclosed

evaluation report on the National Accuracy Clearing House (NAC), prepared by Public

Consulting Group, Inc., for the State

Mississippi, fulfills this requirement.

In 2011,

USDA's

Food and Nutrition Service was awarded $2.5 million by the Office

Management and Budget Partnership Fund for Program Integrity Innovation, with the goal

reducing improper payments that occur due to dual participation in SNAP. This grant funded the

development

the

NAC-a

searchable database to support near real-time sharing

eligibility

information among States. Subsequently, Mississippi was awarded the funding to lead the

project on behalf

a consortium

contiguous States (also including Alabama, Florida,

Georgia, and Louisiana). The pilot operated for

months from June 2014 through May 2015.

The enclosed evaluation report assesses the technical capacity

the NAC, States' success

utilizing the tool and implementing the accompanying business rules, and the cost savings

associated with adoption

the NAC solution.

you have any questions, please have a member

your staff contact Todd Batta, Assistant

Secretary for Congressional Relations, at (202) 720-7095. A similar Jetter and a copy

the

report are being sent to Chairman Roberts, Chairman Conaway, and Ranking Member Peterson

Sincerely,

~<l

~l-...

Thomas J. VilQ •

Secretary

Enclosure

Equal Opportunity Employer

MAY

1 0

2016

The Honorable Mike Conaway

Chairman

Committee on Agriculture

U.S. House

Representatives

USDA

United States Department

Agriculture

Office of the Secretary

Washington, D.C. 20250

1301

Longworth House Office Building

Washington, D.C. 20515

Dear Mr. Chairman:

Section 4032(c)

the Agricultural Act

2014 directed the U.S. Department

Agriculture

(USDA) Secretary to submit a report not later than 90 days after completion to the Committee on

Agriculture

the House

Representatives, and the Committee on Agriculture, Nutrition, and

Forestry

the Senate on the pilot program to test prevention

duplicate participation

the

Supplemental Nutrition Assistance Program (SNAP). The report is to assess the feasibility,

effectiveness, and cost for the expansion

the pilot program nationwide. The enclosed

eva

ation report

the National Accuracy Clearing House (NAC), prepared by Public

Consulting Group, Inc., for the State

Mississippi, fulfills this requirement.

2011, USDA's Food and Nutrition Service was awarded $2.5 million by the Office

Management and Budget Partnership Fund for Program Integrity Innovation, with the goal

reducing improper payments that occur due to dual participation in SNAP. This grant funded the

development

the NAC- a searchable database to support near real-time sharing

eligibility

information among States. Subsequently, Mississippi was awarded the funding to lead the

project on behalf

a consortium

contiguous States (also including Alabama, Florida,

Georgia, and Louisiana). The pilot operated for 12 months from June 2014 through May 2015.

The enclosed evaluation report assesses the technical capacity

the NAC, States' success in

utilizing the tool and implementing the accompanying business rules, and the cost savin

associated with adoption

the NAC solution.

you have any questions, please have a member

your staff contact Todd Batta, Assistant

Secretary for Congressional Relations, at (202) 720-7095. A similar letter and a copy

the

report are being sent to Chairman Roberts, Ranking Member Stabenow, and Ranking Member

Peterson.

Sincerely,

<2w--

<l.

it&.i-.

Thomas

vi10

Secretary

Enclosure

An Equal Opportunity

MAY

1 0

2016

The Honorable Collin Peterson

Ranking Member

Committee on Agriculture

U.S. House

Representatives

USDA

United States Department

Agriculture

Office

the Secretary

Wash

gton

, D.C. 20250

1305 Longworth House Office Building

Washington, D.C. 20515

Dear Congressman Peterson:

Section 4032(c)

the Agricultural Act

of2014

directed the U.S. Department

Agriculture

(USDA) Secretary to submit a report not later than 90 days after completion to the Committee on

Agriculture

the House

Representatives, and the Committee on Agriculture, Nutrition, and

Forestry

the Senate on the pilot program to test prevention

duplicate participation in the

Supplemental Nutrition Assistance Program (SNAP). The report is to assess the feasibility,

effectiveness, and cost for the expansion

the pilot program nationwide. The enclosed

evaluation report on the National Accuracy Clearing House (NAC), prepared

Public

Consulting Group, Inc., for the State

Mississippi, fulfills this requirement.

In 2011, USDA's Food and Nutrition Service was awarded $2.5 million by the Office

Management and Budget Partnership Fund for Program Integrity Innovation, with the goal

reducing improper payments that occur due to dual participation in SNAP. This grant funded the

development

the

NAC

- a searchable database to support near real-time sharing

eligibility

information among States. Subsequently, Mississippi was awarded the funding to lead the

project on behalf

a consortium

contiguous States (also including Alabama, Florida,

Georgia, and Louisiana). The pilot operated for

months from June 2014 through May 2015.

The enclosed evaluation report assesses the technical capacity

the NAC, States' success in

utilizing the tool and implementing the accompanying business rules, and the cost savings

associated with adoption

the NAC solution.

you have any questions, please have a member

your staff contact Todd Batta, Assistant

Secretary for Congressional Relations, at (202) 720-7095. A similar letter and a copy

the

report are being sent to Chairman Conaway, Chairman Roberts, and Ranking Member Stabenow.

Sincerely,

~<}

~L.

Thomas

(;/•

Secretary

Enclosure

Equal Opportunity Employer