TREASURY INSPECTOR GENERAL FOR TAX ADMINISTRATION

 

 

The Taxpayer Assistance Center Closure Plan Was Based on Inaccurate Data

 

 

 

March 2006

 

Reference Number:  2006-40-061

 

 

This report has cleared the Treasury Inspector General for Tax Administration disclosure review process and information determined to be restricted from public release has been redacted from this document.

 

Phone Number   | 202-927-7037

Email Address   |  Bonnie.Heald@tigta.treas.gov

Web Site |  http://www.tigta.gov

 

March 22, 2006

 

 

MEMORANDUM FOR COMMISSIONER, WAGE AND INVESTMENT DIVISION

 

FROM:                            Michael R. Phillips /s/ Michael R. Phillips

                                         Deputy Inspector General for Audit

 

SUBJECT:                    Final Audit Report – The Taxpayer Assistance Center Closure Plan Was Based on Inaccurate Data (Audit # 200540025)

 

This report presents the results of our review of the Taxpayer Assistance Center (TAC) Closure Model. The overall objective of this review was to determine whether the TAC Closure Model (the Model) effectively achieved the Internal Revenue Service’s (IRS) goal of identifying which TACs to close.[1]  

Synopsis

In May 2005, the IRS announced plans to close 68 of its 400 TACs nationwide.  Closing the 68 TACs was expected to yield staffing and facilities cost savings of $45 million to $55 million.  To determine which TACs to close, the IRS and an independent contractor used an industry-standard software package and developed the Model.

After the IRS’ announcement to close 68 TACs, a law was passed delaying the closure of any TACs.[2]  The IRS is prohibited from using funds provided in the 2006 budget appropriation to reduce any taxpayer service function or program until the Treasury Inspector General for Tax Administration (TIGTA) completes a study detailing the effect of the IRS’ plans to reduce services relating to taxpayer compliance and taxpayer assistance. 

The Model, which is criteria based and data driven, ranked each of the IRS’ 400 TACs based on 5 criteria:  1) geography, 2) employee costs, 3) facilities costs, 4) workload, and 5) demographics.  The Model ranked the TACs from highest to lowest score.  The higher the score, the higher the probability a TAC was selected for closure.  To further refine the identification of TACs to be closed, the IRS applied 3 business rules, ensuring 1) a significant office presence would remain in the top 35 metropolitan areas, 2) no State would lose more than one-half of its TACs, and 3) no TACs in Alaska and Hawaii would be closed.  

One benefit of developing this Model is that, for the first time, the IRS has a single database housing key management information for each of the 400 TACs.  The retrieval and consolidation of key management information into a single database addresses a significant operational weakness reported by the TIGTA.[3]  The Model is a flexible decision‑making tool that the IRS should be able to maintain for future analyses.

To test the reliability of the data and the results of the Model, we systematically selected[4] a sample of 60 of the 400 TACs[5] and validated all data subcomponents for each of the 60 TACs.  Testing identified that, although the structure of the Model was sound, not all data used were accurate or the most current available and some of the data were based on estimates and projections instead of actual data currently available.  Data discrepancies affected the scores the Model calculated for each TAC and ultimately the ranking and overall selection of TACs for closure.  When we reran the Model for the 60 selected TACs, the overall scores for each of the 60 TACs changed.  We did not validate data for the remaining 340 TACs and therefore cannot determine if validating all data for all 400 TACS would affect the overall ranking of the TACs and/or the overall selection of which TACs are to be closed.  However, results from validating the data for 60 TACs show sufficient discrepancies and raise concerns about using the results to select which TACs to close and in determining the associated cost savings that might be achieved. 

For example, after the Model was originally run, all 400 TACs were scored and ranked, and the business rules were applied, the IRS identified 68 TACs that, if closed, would achieve the IRS’ targeted savings of $45 million to $55 million.  Each TAC selected for closure had a cumulative score higher than 154.83.  For the 60 TACs sampled, we validated the data and, using the Model, rescored them. Using 154.83 as the cutoff score, we determined the scores for 10 (17 percent) of the 60 TACs included in our sample changed to either above or below the cutoff score, thus raising the possibility that these TACs had been incorrectly identified to be closed or to remain open.

In addition, data discrepancies affected the IRS’ ability to accurately determine cost savings.  The IRS might have overselected or underselected the number of TACs that needed to be closed to reach the targeted savings of $45 million to $55 million.  The inclusion of accurate costing information for all 400 TACs could affect the number of TACs the IRS would have to select to meet its targeted cost savings.  Furthermore, the quality of the Model’s workload data and the absence of customer information diminish the effectiveness of the Model to identify which TACs to close.  In Fiscal Year 2005, we reported  that management information does not provide an adequate picture of all services provided at the TACs.[6]  As a result, neither the IRS nor we can determine the effect TAC closures might have on taxpayer compliance.

Recommendations

The Commissioner, Wage and Investment Division, should ensure data used in the Model or any decision-making tool are accurate and reliable and have been validated before using them to make decisions regarding the TAC Program.  In addition, the Model or any decision‑making tool should include data to identify customer characteristics and capture customer input to effectively measure the impact any results might have on taxpayer service and compliance.

Response

IRS management agreed that data reliability is an issue that must be addressed.  Management also agreed in principle with both of our recommendations and is taking corrective actions.  However, they expressed the concern that measuring the effect of taxpayer services on compliance is a difficult task that the IRS has been unable to accomplish reliably since the inception of taxpayer service programs in the 1940s.

The IRS will ensure data used in the Model or any decision-making tool as it relates to the TAC Program are accurate and verified.  Additionally, data and research are currently being collected for the Taxpayer Assistance Blueprint.  The Taxpayer Assistance Blueprint will be a 5-year plan that outlines what services the IRS should provide, as well as how to improve services for taxpayers by leveraging reliable data on taxpayer and partner needs and preferences.  Phase I will provide a baseline of current IRS services, taxpayer and partner needs and preferences, service industry benchmarking, and strategic service directions.  Phase II will validate the service recommendations through extensive primary research with taxpayers and will identify key operational and resource delivery requirements.  These data will be used, updated, and maintained for use in the Model (or other related decision-making tools) to assist the IRS in making informed decisions regarding all of its taxpayer services.  Management’s complete response to the draft report is included as Appendix VIII.

Copies of this report are also being sent to the IRS managers affected by the report recommendations.  Please contact me at (202) 622-6510 if you have questions or Michael E. McKenney, Assistant Inspector General for Audit (Wage and Investment Income Programs) (Designee), at (202) 622-5916.

 

 

Table of Contents

 

Background

Results of Review

The Taxpayer Assistance Center Closure Model Consolidates Key Management Information Into a Single Database

Data Discrepancies Resulted in the Inability to Correctly Select Taxpayer Assistance Centers for Closure and the Inability to Accurately Determine Cost Savings

The Quality of Taxpayer Assistance Center Workload Data and the Absence of Customer Information Diminish the Effectiveness of the Model to Identify Taxpayer Assistance Centers for Closure

Recommendations 1 and 2:

Appendices

Appendix I – Detailed Objective, Scope, and Methodology

Appendix II – Major Contributors to This Report

Appendix III – Report Distribution List

Appendix IV – Nationwide List of Taxpayer Assistance Centers

Appendix V – Taxpayer Assistance Center Closure Model Subcomponents, Rankings, and Associated Weights

Appendix VI – Taxpayer Assistance Centers Included in the Treasury Inspector General for Tax Administration Validation

Appendix VII – External Stakeholders Contacted by the Treasury Inspector General for Tax Administration

Appendix VIII – Management’s Response to the Draft Report

 

 

Background

 

The Internal Revenue Service (IRS) provides taxpayers the option of obtaining personal, face‑to‑face tax assistance at 400 Taxpayer Assistance Centers (TAC) nationwide.  IRS employees that work in the TACs assist customers by interpreting tax laws and regulations, preparing certain individual tax returns, resolving inquiries on taxpayer accounts, accepting payments, and providing various other services designed to minimize the burden on taxpayers in satisfying their tax obligations.  The IRS suggests taxpayers visit the TACs when they have complex tax issues, need to resolve tax problems relating to their tax accounts, have questions about how the tax law applies to their individual income tax returns, or feel more comfortable talking with someone in person.  See Appendix IV for a list of the 400 TACs. 

 

The IRS announced plans to close 68 TACs nationwide

In May 2005, the IRS announced plans to close 68 TACs nationwide, which was expected to yield staffing and facilities cost savings of $45 million to $55 million.  In an IRS News Release, Internal Revenue (IR)-2005-63, dated May 27, 2005, the IRS stated it:

. . . plans to close a portion of its Taxpayer Assistance Centers as part of the agency’s continuing efforts to create efficiencies, modernize operations, and reduce costs while maintaining its commitment to taxpayer service. Adjusting the number of TAC sites will allow the IRS to focus on activities that provide the most efficient services.  At the same time, taxpayers will still have access to a variety of IRS services, either by [tele]phone, through IRS.gov, and from neighboring TAC offices and IRS volunteer tax assistance programs.  As the IRS is modernizing how taxpayers receive service, the agency remains committed to improving service and meeting the needs of taxpayers . . . .

The IRS Commissioner in a released statement explained:

The President’s 2006 budget request for the IRS is crafted to continue the necessary rebuilding of our enforcement capabilities, which had dropped to unacceptable levels.  The 2006 budget request also calls for a modest amount of belt-tightening in taxpayer services.  This cut to services of 1 percent is consistent with the requests for domestic discretionary programs other than those associated with homeland security.  While we continue to rebuild our enforcement program in these difficult budgetary times, we must make some hard choices to be able to provide the best possible service at the lowest possible cost. 

In conjunction with the TAC closure announcement, the IRS stated that, out of the 2,300 employees that work in the TACs nationwide, fewer than 450 employees were located in the 68 TACs it planned to close.  Further, as the agency’s budget allowed, qualifying employees may be offered early-out retirements and buyouts.[7]  Most employees may be entitled to priority placement for other jobs within the IRS and other Department of the Treasury bureaus. 

The TAC Closure Model was built to determine which TACs to close

To determine which TACs to close with a minimal impact to the taxpaying public, the IRS and an independent contractor used an industry-standard software package and developed the TAC Closure Model (the Model).  The IRS stated the use of the Model would result in a substantial savings of taxpayers’ money while continuing to provide the same level of assistance the public has come to expect from the IRS. 

The Model, which is criteria based and data driven, ranked each of the IRS’ 400 TACs based on over 13,000 data points input to the Model.  The IRS developed five main criteria components to be used in ranking and deciding which TACs to close:  1) geography, 2) employee costs, 3) facilities costs, 4) workload, and 5) demographics.  Included in the 5 main components were 51 subcomponents.  

The IRS provided the contractor with the data used to populate the Model.  The IRS obtained these data from internal sources, as well as from the United States Census Bureau.  The IRS was also responsible for weighting each of the 5 main components and the associated 51 subcomponents.  The weighting was described as being customer centric and based on input from internal and external stakeholders.  Over two-thirds of the weighting focused on customer considerations, including demographics, geography, and workload.  The remaining one-third focused on facilities and labor costs.  Figure 1 presents the five main components and associated weighting. Appendix V details the 51 subcomponents, along with the rankings and associated weights.  

Figure 1:  TAC Closure Model – Five Main Components and Associated Weighting

Figure 1 was removed due to its size.  To see Figure 1, please go to the Adobe PDF version of the report on the TIGTA Public Web Page.

 

To assist in determining the specific weighting to be used in the Model, the IRS consulted with internal stakeholders, including the National Taxpayer Advocate, and external stakeholders, including the IRS Advisory Council.[8]  Initially, the IRS assigned equal weighting to each of the five main components. Based on the data input to populate the subcomponents and the weighting, the Model ranked the TACs from highest to lowest score.  The higher the score, the higher the probability a TAC was selected for closure. However, the results did not meet the IRS’ objective of providing a balanced program in delivering assistance to taxpayers because the TACs initially selected were the largest and most costly TACs. The IRS then revised its methodology to give slightly more weight to geography, demographics, and workload (see Figure 1).  In addition, the IRS developed and applied the following three business rules:

1)      A significant office presence would remain in the top 35 metropolitan areas based on the latest Census population information.  

2)      No State would lose more than one-half of the TACs in that State or have a TAC closed that accounted for more than 40 percent of the customers serviced in that State.  

3)      Alaska and Hawaii would not have any TACs closed, as they are remote locations away from the continental United States. 

The Model was then rerun with the new business rules; it identified 68 TACs to be closed to achieve the targeted savings. 

Congress reacts to the IRS TAC closure announcement

In response to the IRS’ announcement of its plans to close 68 TACs, Congress proposed language to be included in the Department of the Treasury Appropriations Act, 2006,[9] that would delay the closing of any TACs.  The Committee on Appropriations stated in the Senate Report:

. . . Due to the Committee’s concerns, the Committee has included an administrative provision that prohibits the use of funds provided in this Act for purposes of reducing any taxpayer service function or program until the Treasury Inspector General for Tax Administration [TIGTA] has completed a study detailing the impact of the IRS’ plans to reduce services on taxpayer compliance and taxpayer assistance.  The Committee also requests [the] TIGTA to review the accuracy of the estimated cost-savings [sic] of the reduced services.[10]  

The Report further states:

. . . The Committee is concerned about the proposed taxpayer service reductions due to the IRS’ inability to explain the potential impact of these changes on taxpayers.  Reducing taxpayer services, especially for the Nation’s most vulnerable and needy populations, is puzzling, especially given the trends in the Nation’s demographics, which indicate a growing elderly population and immigrant population.  Yet, instead of increasing and improving taxpayer services for these populations, the IRS’ budget proposes to cut services that these populations rely upon.

On July 13, 2005, the IRS announced that early-out retirements and buyouts were being placed on hold, pending Congressional action.  In IRS News Release IR-2005-77, posted July 29, 2005, the IRS announced Congress had provided additional funding with the IRS 2006 budget request and the Commissioner had decided to immediately suspend the proposed closures until any related actions required by the Fiscal Year 2006 IRS appropriation were completed. 

On November 30, 2005, a law was passed stating:

None of the funds appropriated or otherwise made available in this or any other Act or source to the Internal Revenue Service may be used to reduce taxpayer services as proposed in Fiscal Year 2006 until the Treasury Inspector General for Tax Administration completes a study detailing the impact of such proposed reductions on taxpayer compliance and taxpayer services.[11] 

The TIGTA evaluated the accuracy of data input to the TAC Closure Model

This review was performed at the IRS Wage and Investment (W&I) Division Headquarters in Atlanta, Georgia, and at the IRS Agency-Wide Shared Services (AWSS) function and Modernization and Information Technology Services (MITS) organization Headquarters in Washington, D.C., during the period May through December 2005. The audit was conducted in accordance with Government Auditing Standards.  Detailed information on our audit objective, scope, and methodology is presented in Appendix I. Major contributors to the report are listed in Appendix II. 

 

Results of Review

 

The Taxpayer Assistance Center Closure Model Consolidates Key Management Information Into a Single Database

For the first time, the IRS has a single database housing key management information for each of the 400 TACs, including information such as geographic location, staffing costs, facilities costs, workload, and demographics.  Since the IRS reorganized in October 2000 and the Field Assistance Office, which is responsible for the TAC Program, was formed, the IRS has not had key management information readily available in a format that could be used to make sound business decisions. 

The retrieval and consolidation of key management information into a single database addresses a significant operational weakness previously reported by the TIGTA in 2005.[12]  We reported  that, since the creation of the Field Assistance Office in October 2000, key management information used to make decisions and support changes in the TAC Program is either absent or based on incomplete data.  The lack of accurate and complete management information hinders the IRS’ ability to make appropriate decisions when determining the locations and services it provides to taxpayers seeking face-to-face assistance.  Our report further noted that the management information system does not track costs by TAC and does not include operating costs, such as rent, utilities, or equipment. 

The Model was developed to allow for refinements and enhancements

The Model was built using off-the-shelf spreadsheet software. The Model uses formulas based on specific criteria that allow for adjustments, including changing the weighting given to the different components and subcomponents.  Refinements and enhancements can also be easily made, including the addition and deletion of data and components or subcomponents.

For example, when the Model was initially run to identify potential TACs for closure, the IRS weighted each of the 5 main components equally (20 percent).  However, after running the Model, the IRS realized the components were not weighted to ensure balanced delivery of taxpayer assistance.  This resulted in the IRS applying different weighting to the five main components (see Figure 1 for the five main components and associated weighting).  

In addition, although the Model was initially developed to be an objective, data-driven process, it allows subjectivity to be entered into the process.  For example, to further refine the identification of those TACs to be closed, the IRS injected 3 business rules:  ensuring a significant office presence would remain in the top 35 metropolitan areas, ensuring no State would lose more than one-half of its TACs, and ensuring no TACs in Alaska and Hawaii would be closed.  This flexibility should allow the IRS to maintain the Model for future analyses and could serve as a much-needed decision‑making tool for the Field Assistance Office. 

Data Discrepancies Resulted in the Inability to Correctly Select Taxpayer Assistance Centers for Closure and the Inability to Accurately Determine Cost Savings

Although the structure of the Model was sound, not all data used to populate the subcomponents were accurate or the most current data available, and the data used in some subcomponents were based on estimates and projections instead of actual data currently available.  This affected the scores the Model calculated for each TAC and, ultimately, the ranking and overall selection of TACs for closure. 

To test the reliability of the data and the results of the Model, we systematically selected[13] 60 of the 400 TACs[14] and validated all data subcomponents for each of these TACs.  We then reran the Model using the data based on our verification for the 60 TACs. The overall scores for each of the 60 TACs changed.  We did not validate data for the remaining 340 TACs and therefore cannot determine if validating all data for all 400 TACS would affect the overall ranking of the TACs and/or the overall selection of which TACs are to be closed.  However, results from validating the data for the 60 TACs show sufficient discrepancies to raise concerns about using the results to select which TACs to close and to determine the associated cost savings that might be achieved. 

For example, after the Model was originally run, all 400 TACs were scored and ranked, and the business rules were applied, the IRS identified 68 TACs that, if closed, would achieve the IRS’ targeted savings of $45 million to $55 million.  Each TAC selected for closure had a cumulative score higher than 154.83.  For the 60 TACs sampled, we validated the data and, using the Model, rescored them. Using 154.83 as the cutoff score, we determined the scores for 10 (17 percent) of the 60 TACs included in our sample changed to either above or below the cutoff score, thus raising the possibility that these TACs had been incorrectly identified to be closed or to remain open.  Figure 2 shows that, for the 60 TACs sampled, scores for 5 TACs scheduled to be closed (Reference Numbers 24, 26, 32, 43, and 58) changed from above 154.83 to below 154.83 and scores for 5 TACs scheduled to remain open (Reference Numbers 3, 10, 36, 54, and 56) changed from below 154.83 to above 154.83.  For two TACs (Reference Numbers 4 and 51), although the IRS scores were over the cutoff, the TACs were to remain open because they met the criteria of one of the business rules.

Figure 2:  Comparison of Cumulative Scores

Figure 2 was removed due to its size.  To see Figure 2, please go to the Adobe PDF version of the report on the TIGTA Public Web Page.

 

Figure 3 illustrates how the scores of the five components and their subcomponents are calculated and how they changed based on our validation.  This example uses the information relative to the Springfield, Missouri, TAC (Reference Number 24):

Figure 3:  Example of Scoring Methodology – Springfield, Missouri, TAC[15]

Source:  TAC Closure Model and results from the TIGTA validation.

In addition, data discrepancies affected the IRS’ ability to accurately determine cost savings.  The IRS might have overselected or underselected the number of TACs that needed to be closed to reach the targeted savings of $45 million to $55 million.  For example, costing information relative to the Springfield, Missouri, TAC, which includes components Facilities Costs Impact and Employee Costs Impact, were overstated by $179,507.  Cost savings per the Model totaled $649,130, while validated data showed cost savings of $469,623.  Therefore, the inclusion of accurate costing information for all 400 TACs could affect the number of TACs the IRS would have to select to meet its targeted cost savings.  Figure 4 provides the results of the cost savings from the Model and from our validation.

Figure 4:  Comparison of Cost Savings

Figure 4 was removed due to its size.  To see Figure 4, please go to the Adobe PDF version of the report on the TIGTA Public Web Page.

 

Standards for Internal Control in the Federal Government[17] require information to be recorded and communicated to management and others within the entity who need it and in a form and within a time period that enables them to carry out their internal control responsibilities.  For an entity to run and control its operations, it must have relevant, reliable, and timely communications relating to internal as well as external events. 

The 60-day time period to create the Model hindered its development

In anticipation of a reduction in the 2006 budget for taxpayer assistance, the IRS made the decision to close a number of TACs.  For the IRS to complete the placement or reduction in force of employees by October 1, 2005, the TACs to be closed had to be identified and planning for the closures had to start as soon after April 15, 2005, as possible.  An Executive Lead Development Team was formed in mid-February 2005 and charged with developing a data-driven model with specific criteria for determining which TACs to close.  The Team was given the support of a subcontractor and 60 days to build the Model. The 60‑day time period directly related to the decisions made relative to the data used to populate the Model. 

The data needed for the Model were gathered from external sources and various functional areas within the IRS.  These functional areas were under similar time constraints.  For example, the IRS employee asked to provide potential rental cost savings was given 2 hours to provide this information.  However, the information was not readily available, and the IRS employee had only estimates for both the square footage and costs per square foot from which to calculate potential rental cost savings.  In another instance, the IRS employee responsible for providing workload information was given 1 day to gather the data.  Because of these time constraints, the data included in the Model were not validated before or after being input to the Model.  

Factors contributing to the data discrepancies include the following:

·         Use of estimates and projections.  The IRS used estimates for 13 of the 51 subcomponents, including MITS Costs, Furniture Costs, Length of Rent/Leasing Contract, and the subcomponents relating to the number and types of employees at a TAC.  For the MITS Costs, the IRS assigned an estimated cost based on the size of the TAC, not the recurring cost for information technology assigned to each employee.  TIGTA auditors worked with the responsible organizations within the IRS to obtain the budgeted per-employee recurring MITS Costs of $2,733 per person and then multiplied this per-employee cost by the number of employees at the TAC to compute the MITS Costs savings per TAC.  For Furniture Costs, the IRS attempted to estimate a recurring facility cost per employee.  However, the IRS was unable to provide the data or the computation to support the estimates.  Based on the inability to provide this information, the IRS agreed that this subcomponent should not have been included in the Model and should not be used in determining site closure selection and cost savings. 

For those categories relating to the number and types of employees, the IRS used the number and types of employees based on actual employees working at the beginning of Fiscal Year 2005 and projections for the 2005 Filing Season.[18]  To accurately determine the number and types of employees, we identified the specific employees assigned to a TAC per the Employee Service Record Report (ESRR) provided by the Field Assistance Office and then confirmed the number and types of employees via physical verification while conducting TAC site visits.  The ESRR was available for use at the time the Model was populated. 

·         Use of data that were not the most current available.  The IRS used data that were not the most current available for 3 of the 51 subcomponents, including Number of Returns Filed by Zip Code and Number of EITC [Earned Income Tax Credit] Returns Filed by Zip Code.  The IRS used Tax Year 2002 data to compute the values associated with these subcomponents.  We used Tax Year 2003 data, which were the most current information available at the time the IRS populated the Model.

The IRS used Fiscal Year 2004 information when determining the values for Modernization Efforts Applied.[19]  This resulted in the IRS not including all TACs that had been modernized during the period September 2004 to April 2005.  For example, of the 60 TACs we sampled, 6 (10 percent) had been modernized; however, the Model showed the 6 TACs had not been modernized.  The IRS estimates the costs to modernize a TAC range from $124,000 to over $360,000, depending on the size of the TAC.  For the subcomponent, Modernization Efforts Applied, the IRS assigns a low probability of selection for closure to those TACs that have been modernized. 

·         Use of averages.  The IRS used averages for 6 of the 51 subcomponents, when in 4 instances actual figures could have been obtained.  These four subcomponents included Average Salary, Space Usage in Square Feet, Total Rent/Leasing Cost, and Square Footage Costs.  When the IRS computed average salaries, locality pay was not included.  Locality pay ranges from 11.75 percent to 26.39 percent of an employee’s salary, depending on where the employee works.  To determine employees’ salaries, we researched the actual salaries of the specific employees assigned to a TAC based on the ESRR provided by the Field Assistance Office.

To compute the Space Usage in Square Feet and Square Footage Costs, the IRS used the number of employees assigned to a TAC to assign a size to a TAC – small, medium, or large.  Based on the TAC size, an average square footage and cost per square foot was applied.  To determine the actual square footage costs, TIGTA auditors reviewed monthly rental bills providing the cost per square foot and, with AWSS function employees, physically measured each of the 60 TACs included in our sample. 

In addition, the Model did not include information as to whether the space occupied by a TAC could be released and, if so, what costs may be involved in the release.  In many instances, Federal Government agencies incur a cost to revise previously occupied space to a condition that can be released back.  For example, subsequent to the TAC closure announcement, the AWSS function determined whether the space occupied by the 68 TACs scheduled to be closed could be released and whether there would be costs involved in the release.  The AWSS function determined the space occupied by 15 (22 percent) of the 68 TACs could not be released.  Reasons the space for the 15 TACs could not be released included noncancelable leases and the fact that the location of the TAC served as the main entrance to offices for other IRS functions.  Of the remaining 53 of 68 TACs, 48 would require revisions to the space prior to its being released.  The IRS estimates revision costs will total approximately $4 million, with the per-TAC revision cost ranging from $1,190 to $425,332.

For the 60 TACs TIGTA auditors visited, AWSS function representatives stated the space for 19 (32 percent) could not be released.  Reasons for nonrelease were similar to those cited above.

·         Inconsistency in the calculation approach.  Only 2 of the 14 subcomponents within the main component Demographic Impact involved inconsistent use of a geographic area in the calculation.[20]  The first subcomponent involving an inconsistent approach is the Percent Unemployed by Zip Code.  This subcomponent was actually the percentage of unemployment per the State in which the TAC is located.  IRS employees responsible for creating the Model noted that computing the percentage of unemployment by geographic area would have required much more time than taking the per-State figure.  The second subcomponent involving an inconsistent approach is the Income Level by Zip Code.  This subcomponent was actually the income level for the zip code where the TAC was located.  IRS employees responsible for creating the Model could not explain why a geographic area was not used.

·         Error in associating zip codes with the nearest TAC.  An error resulted when the IRS combined the results to associate the zip codes with a specific TAC because two different programs were used to compute latitude and longitude.  The two programs used different geographical points within each zip code. 

Associating zip codes with the nearest TAC allowed the IRS to analyze data based on a geographic area.  The IRS assumed that associating zip codes with the closest TAC would present the population of individuals most likely to seek assistance at that TAC.  Two separate programs were used to assign latitude and longitude coordinates to zip codes.  The IRS W&I Division Research function determined the latitude and longitude coordinates for the zip code in which each of the 400 TACs is located. The IRS Office of Program Evaluation Risk Analysis (OPERA) determined the latitude and longitude for the remaining zip codes nationwide.

For consistency purposes, the OPERA recomputed the latitude and longitude for all zip codes and reassociated the zip codes with the nearest TAC.  As a result, 857 zip codes were reassociated with TACs different from the ones used when the Model was first run.  When we reran the subcomponent calculations using the revised associated zip codes, the reassociation created changes to the values in those subcomponents.

The Quality of Taxpayer Assistance Center Workload Data and the Absence of Customer Information Diminish the Effectiveness of the Model to Identify Taxpayer Assistance Centers for Closure

Prior TIGTA audits have raised concerns regarding the reliability of Field Assistance Office management information as a basis for making business decisions.[21]  In 2005, we reported that, since the creation of the Field Assistance Office in October 2000, key management information used to make decisions and support changes is either absent or based on incomplete data.  Improved management information is needed to help the Field Assistance Office move toward its future goals.

In addition, although a number of studies have been conducted by the W&I Division to identify its present and future customer base, the Field Assistance Office has not recently conducted similar studies to identify the specific characteristics of customers who seek face-to-face assistance as well as the services they desire. As a result, the Model does not include demographics of IRS customers.  The majority of the Model’s demographics are based on United States Census Bureau data.  The Model includes the total number of tax returns filed on paper and electronically.  However, the IRS did not include in the Model either the filing characteristics of taxpayers that live in the geographic areas of the TACs or the demographics of taxpayers that actually sought account assistance at the TACs.

Data captured on the Field Assistance Office’s management information system are not always reliable

The IRS acknowledges that the Field Assistance Office management information system cannot be relied upon for timely, accurate workload performance information due to the manual process of recording taxpayer visits.  In addition, the Field Assistance Office current management information system does not collect the total number of services provided to taxpayers; it captures only the most significant service provided to each taxpayer.  In cases where multiple services are provided to the same taxpayer, guidelines direct TAC employees to record the service they believe was the most significant provided to the taxpayer. This is usually based on the amount of time spent assisting the taxpayer with a specific service.

IRS studies have identified that TAC employees use incorrect closing codes when documenting the type of assistance provided to taxpayers.  Testing conducted during visits to the 60 TACs sampled in this review confirmed that both inaccuracies and inconsistencies exist when TAC employees record closing codes documenting the actions taken when assisting customers.  For example:

·         Employees incorrectly recorded tax returns received and stamped in the categories of Customer w/Form Contacts and Account Work/Notices, and in one TAC each employee recorded the receipt and stamping in a different category.

·         Employees recorded the receipt of Heavy Highway Vehicle Use Tax Returns (Forms 2290) in the categories of Tax Law Question, Account Work/Notices, Return Preparation, or Other Field Assistance Contacts.  At some TACs, the receipt of a Form 2290 was not recorded at all.  Field Assistance Office procedures require this action to be recorded under Other Field Assistance Contacts.

·         Employees in one TAC recorded each IRS form handed out as a contact, while in other TACs only one contact was recorded regardless of the number of forms provided to a taxpayer.  Field Assistance Office procedures require one contact to be recorded regardless of the number of forms provided.

The Government Performance and Results Act of 1993[22] directs Federal Government agencies to focus on their missions and goals and provides guidance on how to achieve those goals and how to improve their structural organizations and business processes.  Performance measures need to be based on program-related characteristics and performance data and must be sufficiently complete, accurate, and consistent.  Performance data must be used to improve organizational processes, identify performance gaps, and set improvement goals.

The Field Assistance Office’s management information system does not provide an adequate picture of all services provided at the TACs. Specifically, not all services are captured, services are inconsistently captured, and for some services offered there is no means by which to identify the specific customer to whom the service was provided (e.g., customers that obtained tax law and tax forms assistance). In response to our prior report,[23] IRS management agreed that improving their management information system is critical to achieving improved operational performance.  The Field Assistance Office is developing a web-based Field Assistance Office Management Information System (FAMIS), which will provide management with critical program planning and control information at the local and national levels while also reducing taxpayer burden by managing wait times more efficiently. The FAMIS will be tested in Fiscal Year 2006 and deployed in all TACs during Fiscal Year 2007.

Demographics based on IRS customer information were not included in the Model

IRS research performed since 1999 has focused primarily on learning about customer expectations and ways to improve the taxpayer experience when visiting the TACs. Taxpayers who need or desire face-to-face assistance have not been involved in determining what services are offered at the TACs.  Management indicated identification of the services offered at the TACs has been based primarily on input from Field Assistance Office field management.  Other factors affecting the services provided include internal priorities, resource demands, and shifts in the IRS’ customer service perspective.

However, the IRS does have available taxpayer account and compliance data that could be used to better identify the characteristics of potential customers and those customers that have actually sought account assistance at a TAC.  For example, the Model includes demographics of individuals in a geographic area[24] based on information from the 2000 Census and IRS tax return data.  Figure 5 illustrates the different results when demographics are included based on different data.  For 10 TACs in our sample, it compares 5 key demographics using 3 different data sources labeled “Model,” “Actual,” and “Potential.”

·         “Model” includes data from the Model, including corrected information validated by the TIGTA.

·         “Actual” includes data relating to taxpayers that actually obtained assistance with their tax accounts at a particular TAC.

·         “Potential” includes data relating to taxpayers that filed their tax returns and lived in the geographic area of a TAC.

Figure 5:  Comparison of IRS Customer Key Demographics

 

Figure 5 was removed due to its size.  To see Figure 5, please go to the Adobe PDF version of the report on the TIGTA Public Web Page.

The IRS could also analyze data from electronic systems used by TAC employees to assist taxpayers with compliance issues.  We obtained and analyzed extracts from two of these systems – the Automated Collection System[25] and the Integrated Case Processing System.[26]  Results include when assistors serve taxpayers that came into the TACs voluntarily for taxpayer assistance as well as taxpayers that came into the TACs to address compliance issues.[27]  Results are shown nationwide, but the data can be analyzed on a per-TAC basis.

·         For Fiscal Year 2004, 616 TAC employees[28] accessed the Automated Collection System 20,515 times to assist 13,768 taxpayers.  Actions taken included changing a follow-up date to provide a taxpayer more time to resolve a compliance issue, adding a levy,[29] deleting a levy source, and updating a taxpayer’s financial information. 

·         For Fiscal Year 2004, 1,780 TAC employees accessed the Integrated Case Processing System 971,390 times to assist 236,675 taxpayers.  Actions taken included adding history notes to taxpayer accounts describing the nature of the contact, actions taken, and instructions given; reviewing information relative to a taxpayer’s payment history; and updating a taxpayer’s account as a result of securing a delinquent tax return(s). 

The effect of TAC closures on taxpayers

The IRS cannot determine the effect TAC closures might have on taxpayer compliance.  The IRS does not currently have the means to capture all interactions between a TAC employee and a taxpayer to determine why the taxpayer visited the TAC, what service he or she received, and, most importantly, the effect the service or action has on the taxpayer’s future compliance. In addition, as previously reported,[30] although the IRS does have management information to determine to some degree which taxpayers visit TACs, the information is not always reliable.  Because reliable information was not available, we were unable to determine the effect TAC closures might have on compliance.

The Model represents the IRS’ first step in developing a much-needed Service Delivery Assessment Tool that can be used as the basis for making informed business decisions relative to the TAC Program.  However, the effectiveness of the tool is directly impacted by the accuracy and reliability of the data used to populate the tool.  In addition, as cited in our 2005 report, a Service Delivery Plan is needed to outline the short- and long-term direction of the TAC Program based on business cases and customer input.  The IRS agrees, and the Field Assistance Office is currently exploring options to obtain customer input in the development of its Service Delivery Model.  When available, this should provide the IRS with some perspective on the effect TAC services have on its customers. 

The Senate Committee on Appropriations asked us to contact external organizations whose constituents use the services of the TACs to assess the potential impact the TAC closures might have on them.  The organizations contacted included those suggested by the Senate Committee on Appropriations, National Taxpayer Advocate, and IRS Stakeholder Partnerships, Education and Communication function.  See Appendix VII for a list of groups contacted.

·         Eight of 11 stakeholder groups believe closing the TACs may make it harder for their constituents to stay compliant with tax laws and file tax returns. 

·         Ten of 11 stakeholder groups believe their constituents would prefer not to contact the IRS toll-free telephone number. 

·         All 11 stakeholder groups believe their constituents are not currently likely to use alternative methods, such as the Internet or email, to obtain the services they need. 

The IRS has recently undertaken a large-scale study that focuses on improving taxpayer service.  IRS officials advised us the study will follow a structured approach utilizing available research to understand taxpayers’ needs and preferences.  The study will identify a strategic direction to best balance those needs with existing business constraints. Nevertheless, prior to making decisions on closing any TACs, the IRS should ensure it knows what taxpayers visit the TACs for assistance and why, so it can determine the impact on these taxpayers and ensure alternative service delivery channels are effective in meeting the needs of these taxpayers.

Recommendations

The Commissioner, W&I Division, should:

Recommendation 1:  Ensure data used in the Model or any decision-making tool are accurate and reliable and have been validated before using them to make decisions regarding the TAC Program. 

Management’s Response:  IRS management agreed with this recommendation and will ensure data used in the Model or any decision-making tool as it relates to the TAC Program are accurate and verified.

Recommendation 2: Include in the Model or any decision-making tool data to identify customer characteristics and capture customer input to effectively measure the impact any results might have on taxpayer service and compliance. 

Management’s Response:  IRS management agreed with this recommendation. Data and research are currently being collected for the Taxpayer Assistance Blueprint that will baseline customer characteristics and needs.  The Taxpayer Assistance Blueprint will be a 5-year plan that outlines what services the IRS should provide, as well as how to improve services for taxpayers by leveraging reliable data on taxpayer and partner needs and preferences. Phase I will provide a baseline of current IRS services, taxpayer and partner needs and preferences, service industry benchmarking, and strategic service directions.  Phase II will validate the service recommendations through extensive primary research with taxpayers and will identify key operational and resource delivery requirements.  These data will be used, updated, and maintained for use in the Model or other decision-making tools to assist the IRS in making informed decisions regarding all of its taxpayer services.

 

Appendix I

 

Detailed Objective, Scope, and Methodology

 

The overall objective was to determine whether the Taxpayer Assistance Center (TAC) Closure Model (the Model) effectively achieved the Internal Revenue Service’s (IRS) goal of identifying which TACs to close.[31] 

To address our overall objective, we used various electronic files from the IRS.  We did not validate these files.

• United States Census 2000 file, which provides characteristics based on the 2000 Census.  The information was originally obtained by the IRS from the United States Census Bureau.  This information was used to confirm statistics included in the Model.
• Integrated Case Processing System[32] extract identifying TAC employee accesses and actions taken on this System during Fiscal Year 2004.
• Zip code association file,[33] which provided the results of the IRS associating zip codes nationwide with the closest TAC site.  We used this information to confirm statistics included in the Model.

In addition, we used the following electronic files housed at the Treasury Inspector General for Tax Administration’s (TIGTA) Data Center Warehouse.  The files and validations performed include:

• Returns Transaction File[34] – determined if all records were received by the TIGTA and verified a sample of the fields for accuracy.
• Automated Collection System[35] – determined if all records were received by the TIGTA and verified the data values were correctly extracted from the original file and the values looked reasonable.  This was accomplished by printing out a subset of the records and reviewing the data.
• National Account Profile[36] – reviewed fields to identify any obvious problems with the data and selected a number of records to compare against the Integrated Data Retrieval System[37] to ensure the data matched.
• Treasury Integrated Management Information System[38] – determined if all files were received by the TIGTA and reviewed the data fields to ensure they were in the expected formats.

To accomplish our objective we:

I.                   Assessed the development of the Model, including the completeness and accuracy of the data. 

A.    Interviewed IRS officials to determine the goal/purpose of the Model; how it was developed; and what it was to achieve, including the cost savings the IRS wanted to achieve. 

B.     Interviewed the independent contractor to discuss the development of the Model, including background, logic, and data sources. 

C.     Determined the basis for the subcomponents used in the Model, including how the subcomponents were selected and weighted. 

D.    Determined what factors were considered when using the business rules and the basis for the rules. 

E.     Determined what validation was completed on the data used in the Model. 

F.      Met with the appropriate Congressional committees to obtain Congressional concerns. 

II.                Determined if the Model structure and methodology were sound to provide consistent, reliable results. 

A.    Consulted with a statistician to determine if the formulas and structure were sound and logical. 

B.     Reran the Model with the business rules to determine if the results matched those of the IRS. 

C.     Determined if the data used in the Model were accurate and, if not, if the inaccuracies affected the results. 

III.             Determined if all subcomponents of the Model were appropriate and, if not, eliminated them from the Model. 

A.    Determined which data subcomponents could be validated by 100 percent testing and which would be tested and validated through sampling.

B.     Consulted with a statistician to determine the appropriate sampling methodology to select those TACs to be included in our validation.  Based on the recommendation of the statistician, we used systematic sampling, which included:

1.      Selecting 30 TACs from the 68 TACs selected by the IRS for potential closure by dividing 68 by the sample size of 30 and obtaining 2.27.  We randomly selected a number between 1 and 2.27 for a starting point; the number 2 was selected.  We began at the second TAC listed and selected it, added 2.27 to the number 2 for a total of 4.27 and selected the fourth TAC on the list, and repeated the process until 30 of the 68 TACs had been selected.

2.      Selecting 30 TACs from the 332 TACs to remain open by dividing 332 by the sample size of 30 and obtaining 11.07.  We ran a random number generator until a random number between 1 and 11.07 resulted and obtained the random number 9.05.  We began at the ninth TAC on the list and selected it, added 11.07 to 9.05 for a total of 20.12 and selected the 20th TAC on the list, and repeated the process by rounding to a whole number until 30 of the 332 TACs had been selected. 

C.     Visited the 60 selected TACs[39] and met with Field Assistance Office, Agency-Wide Shared Services function, and Modernization and Information Technology Services organization personnel to obtain costs for releasing and removing or rebuilding computer and technology equipment.  

D.    Completed data analysis to validate the remaining data subcomponents for the 60 selected TACs to determine if the data in the Model appeared reasonable or accurate. 

E.     Computed employee costs for the 60 TACs by verifying the employees currently employed at the TACs at the time of our visits; by obtaining from the IRS the Employee Service Record Reports, which detail actual salaries for the employees employed at the TACs; and by working with IRS contacts to determine what was included in the average benefits figure included in the Model. 

F.      Reran the Model to determine if the scores and rankings changed. 

G.    If our results differed from IRS results, reported results to the IRS.  We discussed with the IRS the expected cost savings and determined what TACs would be closed to achieve those cost savings. 

IV.             Identified additional data subcomponents, if any, that should have been included in the Model. 

A.    Met with IRS personnel to discuss factors considered for the Model. 

B.     Identified the workload data that were not included in the Model by analyzing data from various IRS sources.  For workload data, we obtained and analyzed information from IRS’ Automated Collection System and Integrated Case Processing System.  We followed up on any anomalies. 

C.     Met with the National Taxpayer Advocate to discuss any concerns regarding TAC closures. 

D.    Interviewed appropriate IRS personnel to discuss any additional data to be considered for inclusion in the Model. 

V.                Assessed the potential effect the TAC closings may have on taxpayers.  This information was requested by the Senate Committee on Appropriations staff. 

A.    Determined the volumes of TAC customers and the services provided to them to validate the IRS management information and to determine the effect the closures might have on taxpayers. 

B.     Determined the characteristics of taxpayers in the areas near the TACs by analyzing the data on the IRS’ Returns Transaction File.  

C.     Obtained and reviewed IRS-developed research reports regarding taxpayers and the services offered by the TACs. 

D.    Met with external stakeholders to determine their concerns and what services they value or need from the TACs (see Appendix VII for a list of stakeholders contacted).

VI.             Determined costs associated with closing the selected TACs.  

 

Appendix II

 

Major Contributors to This Report

 

Michael R. Phillips, Deputy Inspector General for Audit

Michael E. McKenney, Assistant Inspector General for Audit (Wage and Investment Income Programs) (Designee)

Augusta R. Cook, Director

Scott MacFarlane, Director

Paula W. Johnson, Audit Manager

Frank W. Jones, Audit Manager

Russell P. Martin, Audit Manager

Kenneth L. Carlson, Senior Auditor

Pamela M. DeSimone, Senior Auditor

Lena M. Dietles, Senior Auditor

Deborah L. Drain, Senior Auditor

Jackie E. Forbus, Senior Auditor

Robert J. Howes, Senior Auditor

Sharon R. Shepherd, Senior Auditor

Grace M. Terranova, Senior Auditor

Jerome S. Antoine, Auditor

Robert A. Baker, Auditor

Jean M. Bell, Auditor

Jerry G. Douglas, Auditor

Roberta A. Fuller, Auditor

Andrea M. Hayes, Auditor

Patricia A. Jackson, Auditor

Mary L. Keyes, Auditor

Sylvia D. Sloan-Copeland, Auditor

Geraldine S. Vaughn, Auditor

Lindsey J. Cabral, Auditor Intern

James M. Allen, Information Technology Specialist

Kevin O’Gallagher, Information Technology Specialist

Layne D. Powell, Information Technology Specialist

Jeffrey E. Williams, Information Technology Specialist

 

Appendix III

 

Report Distribution List

 

Commissioner  C

Office of the Commissioner – Attn:  Chief of Staff  C

Deputy Commissioner for Operations Support  OS

Deputy Commissioner for Services and Enforcement  SE

Chief, Agency-Wide Shared Services  OS:A

Deputy Commissioner, Wage and Investment Division  SE:W

Director, Customer Account Services Consolidation  SE:W

Director, Customer Assistance, Relationships, and Education  SE:W:CAR

Director, Strategy and Finance, Wage and Investment Division  SE:W:S

Chief, Performance Improvement, Wage and Investment Division  SE:W:S:PI

Director, Field Assistance  SE:W:CAR:FA

Director, Media and Publications  SE:W:CAR:MP

Director, Stakeholder Partnerships, Education, and Communication SE:W:CAR:SPEC

Chief Counsel  CC

National Taxpayer Advocate  TA

Director, Office of Legislative Affairs  CL:LA

Director, Office of Program Evaluation and Risk Analysis RAS:O

Office of Management Controls  OS:CFO:AR:M

Audit Liaison:  Senior Operations Advisor, Wage and Investment Division  SE:W:S

 

Appendix IV

 

Nationwide List of Taxpayer Assistance Centers

 

The following data are from the Internal Revenue Service’s Taxpayer Assistance Center Closure Model.  Employees in the Taxpayer Assistance Centers provide face-to-face assistance to customers by interpreting tax laws and regulations, preparing certain individual tax returns, resolving inquiries on taxpayer accounts, accepting payments, and providing various other services designed to minimize the burden on taxpayers in satisfying their tax obligations.  The 68 Taxpayer Assistance Centers the Internal Revenue Service selected for closure are shaded in grey.

The chart was removed due to its size.  To see the chart, please go to the Adobe PDF version of the report on the TIGTA Public Web Page.

 

Appendix V

 

Taxpayer Assistance Center  Closure Model Subcomponents, Rankings, and Associated Weights

 

The chart was removed due to its size. To see the chart, please go to the Adobe PDF version of the report on the TIGTA Public Web Page.

 

1.       Internal Revenue Service (IRS) employees who work in the Taxpayer Assistance Centers (TAC) provide face-to-face assistance to customers by interpreting tax laws and regulations, preparing certain individual tax returns, resolving inquiries on taxpayer accounts, accepting payments, and providing various other services designed to minimize the burden on taxpayers in satisfying their tax obligations.  The IRS currently has 400 TACs.  To determine which TACs to close with a minimal impact to the taxpaying public, the IRS and an independent contractor used an industry-standard software package and developed the TAC Closure Model.

2.       Program run by the IRS through which trained community volunteers provide free tax help to individuals who qualify.

3.       A self-service, multimedia structure used to dispense tax forms and basic tax information.

4.       IRS employees who work at the TACs greeting and questioning taxpayers to determine the type of assistance needed.

5.       IRS employees who work at the TACs and are trained to provide end-to-end services to individual taxpayers.

6.       IRS employees who work at the TACs and are trained to communicate with taxpayers and to be knowledgeable of tax law and related IRS operational procedures.

 

Appendix VI

 

Taxpayer Assistance Centers Included in the Treasury Inspector General for Tax Administration Validation  

 

Employees in the Taxpayer Assistance Centers provide face-to-face assistance to customers by interpreting tax laws and regulations, preparing certain individual tax returns, resolving inquiries on taxpayer accounts, accepting payments, and providing various other services designed to minimize the burden on taxpayers in satisfying their tax obligations. 

The chart was removed due to its size.  To see the chart, please go to the Adobe PDF version of the report on the TIGTA Public Web Page.

 

Appendix VII

 

External Stakeholders Contacted by the Treasury Inspector General for Tax Administration

 

Account Ability Minnesota Mission

American Institute of Certified Public Accountants

Center for Economic Progress

Children’s Defense Fund

Federal Deposit Insurance Corporation

Iowa State University Extension

Legal Services of North Dakota

Low Income Tax Clinic/University of Missouri Graduate Tax Law Foundation

National Community Tax Coalition

University of Connecticut Law School Legal Clinic

Women’s Economic Development, Atlanta, Georgia

 

Appendix VIII

 

Management’s Response to the Draft Report

 

The response was removed due to its size.  To see the response, please go to the Adobe PDF version of the report on the TIGTA Public Web Page.