1 = Tax return/Return information

Executive Summary

Objectives and Scope

Background

Results

Conclusion

Appendix I – Detailed Objectives, Scope, and Methodology

Appendix II – Major Contributors to This Report

Appendix III – Report Distribution List

Appendix IV – Outcome Measures

Appendix V – Management’s Response to the Draft Report

Appendix VI – Abbreviations Used in This Report

Appendix VII – Memorandum #1: Risks Associated With the Development of the Integrated Submission and Remittance Processing System

Appendix VIII – Management’s Response to Memorandum #1

Appendix IX – Memorandum #2: Risks Associated With the Operation of the Integrated Submission and Remittance Processing System During the 1999 Filing Season

Appendix X – Management’s Response to Memorandum #2

Executive Summary

The Internal Revenue Service (IRS) replaced the computer systems used to annually process approximately 170 million paper tax returns and 50 million payments. The new system, referred to as the Integrated Submission and Remittance Processing (ISRP) System, is necessary because the existing computers cannot process data after
December 31, 1999. This is our third audit of the ISRP system. The first two audits evaluated the processes used by the IRS to design and develop the ISRP system. During this audit, we evaluated the IRS’ process for ensuring that problems identified during the system’s pilot were adequately corrected.

Overall, the IRS’ implementation of ISRP was successful. As of May 21, 1999, the IRS had processed over 66 million tax returns and over 11 million payments through the ISRP systems nationwide. Despite these notable accomplishments, the process of integrating ISRP with other IRS operations has resulted in additional risks and challenges affecting the IRS goal of ensuring that taxpayers are provided top quality submission processing services.

During this review, we found that document processing changes increased taxpayer burden, approval of a system enhancement placed the implementation at risk, contingency plans were incomplete and untested, and access to taxpayer information was not properly controlled.

The implementation of ISRP required changes to the IRS’ procedures for controlling documents. These changes resulted in additional taxpayer burden and created additional work in order to correct processing errors. For example, we found that:

• ISRP data entry operators occasionally failed to process tax returns (did not enter the tax return data). These unprocessed tax returns resulted because ISRP operators were no longer required to enter the unique serial number of each document processed. Based upon our review of randomly sampled tax returns, we estimate that approximately 230 tax returns processed through the ISRP systems at 3 sites were sent to warehouses without being processed and an additional 7,800 tax returns contained serial numbering problems.
• The amount of taxpayer contacts required to resolve payment processing mistakes increased. The ISRP system used to create and store computerized pictures (images) of the original payment documents (checks, money orders, payment vouchers, etc.) did not always capture usable information. Based upon our review of randomly sampled payments, we estimate that the ISRP system’s computer archives did not capture any images for approximately 2,900 payments and captured incomplete sets of images for approximately 6,000 additional payments.²
• Delays in the numbering of tax returns received with payments affected the accuracy of penalty assessments and the IRS’ ability to research payment-related processing errors. In the past, the control numbers for tax returns received with payments were assigned when the payments were processed and matched the control number assigned to the payment. Some IRS computer programs assessed penalties based upon dates reflected in the control number assigned to the tax return.

Since the ISRP system does not assign matching control numbers to tax returns and their accompanying payments, process changes caused delays in the document numbering process and resulted in discrepancies between the actual IRS received dates and the dates reflected in the control number. Had this problem not been corrected, thousands of taxpayers could have been assessed erroneous penalties and issued erroneous balance due notices. In addition, we found that the mismatched control numbers often prevented the IRS from locating the original tax return, making it more difficult for the IRS to correct processing errors related to these payments.

The IRS approved a significant enhancement to the design of the ISRP system despite delays in the project’s implementation schedule and unresolved development problems. After formally advising the vendor of concerns regarding the timely delivery of ISRP, the IRS authorized a significant system enhancement. The additional requirement increased the risk of delays in the vendor’s delivery of the system. After we advised them of the potential risks, IRS management decided to delay development of the enhancement.

On January 29, 1999, we advised the IRS of our concerns with its back-up plans for unexpected tax return and payment processing problems (i.e., contingency plans). We reported a lack of detail in the 1999 contingency plans, budget and staffing reductions that limited contingency alternatives, and the omission of the ISRP system from local recovery plans. In addition, we questioned the reliability of the IRS’ payment processing capabilities if problems with ISRP occur after the year 2000. IRS officials agreed with our recommendations and took corrective actions regarding the 1999 contingency plans, but they did not agree with our recommendations regarding contingency plans for problems in the year 2000 and beyond.

We are still concerned with the IRS’ year 2000 back-up payment processing capabilities. If detailed contingency plans are not both finalized and tested, unexpected problems with the IRS’ payment processing systems may result in untimely deposits, unprocessed payment transactions, and errors in calculating tax due balances. As of October 14, 1999, the IRS had neither completed negotiations with the vendors expected to provide the back-up services nor tested the contingency plans currently in place.

Due to the personal and sensitive nature of information on IRS computer systems and the legal requirements to safeguard the privacy of taxpayer information, the IRS established procedures to check the background of all individuals granted access to its computer systems. This includes the employees of vendors contracted to implement and maintain ISRP.

We reviewed the background investigation results for 51 vendor employees at 6 service centers. We found that 59 percent of these employees (30 of 51) did not have completed background investigations. Although we did not identify any misuse of taxpayer data, vendor employees with the computer knowledge and skills necessary to misuse this data were allowed access to ISRP before the completion of their background investigations.

The report contains specific recommendations for the IRS to improve the ISRP system’s control and accountability of all documents processed, to update submission processing contingency plans, and to improve ISRP controls designed to safeguard taxpayer information.

Management’s Response: IRS management agreed with the issues addressed in this report and stated that they have implemented corrective actions to improve the ISRP system’s control and accountability of documents, strengthen contingency plans, and improve ISRP controls designed to safeguard taxpayer information. IRS management’s complete response to the draft report is included as Appendix V.

Objectives and Scope

This is the third in a series of audit reports presenting the results of the Treasury Inspector General for Tax Administration’s (TIGTA) review of the development and implementation of the Integrated Submission and Remittance Processing (ISRP) system. We began the initial evaluation shortly after the Internal Revenue Service (IRS) selected Lockheed Martin Federal Systems (LMFS) as the system’s developer. On January 30, 1998, the IRS Inspection Service (now TIGTA) issued a report detailing significant risks to the project’s ability to meet the year 2000 implementation date, including an aggressive implementation schedule, a lack of back-up plans, and system design problems. On November 6, 1998, the IRS Inspection Service (now TIGTA) issued a second report detailing system design, project scheduling, and resource allocation risks. The IRS implemented corrective actions for each of these reports.

During this audit, we evaluated the IRS’ (1) process for ensuring that problems identified during the system’s pilot were adequately corrected, (2) preparation for the implementation of the ISRP system, and (3) operation of ISRP during the 1999-filing season. We kept IRS management advised of significant issues through various discussions, direct electronic mail messages, and the issuance of two audit memoranda
(see Appendices VII & IX).

The audit was performed in accordance with Government Auditing Standards from September 1998 through May 1999 at the IRS’ National Office and the 10 service centers. Details of our audit objectives, scope, and methodology are presented in Appendix I. Major contributors to this report are listed in Appendix II.

Background

The IRS replaced the computer systems used to process approximately 170 million paper tax returns and
50 million payments received from taxpayers each year. On August 16, 1999, the IRS began processing both paper tax returns and payment documents through ISRP at the last of 10 IRS service centers nationwide.

The ISRP system replaces the existing Distributed Input (DIS) and Remittance Processing (RPS) Systems and is necessary because neither of these systems is capable of processing data after December 31, 1999.

The IRS approved the initial development of ISRP on August 22, 1996, and on December 20, 1996, selected LMFS as the system’s developer. The initial design, development, and testing activities occurred in 1997. The Austin Service Center (AUSC) operated ISRP for the first time (piloted the system) during the 1998-filing season.

Despite numerous delays and production problems, the ISRP pilot met the AUSC’s initial tax return processing deadlines and most of its payment processing deadlines. Although ISRP met the April peak deposit program completion date (PCD), it was not able to meet the daily and monthly deposit requirements consistently. Based upon these results, the IRS modified plans for the 1999-filing season by limiting the nationwide implementation of ISRP’s payment processing components to 6 of 10 submission processing facilities (service centers). The IRS planned to implement ISRP’s payment processing component at the other four service centers before the 2000-filing season. Its plans to implement ISRP’s return processing component at all 10 service centers for the 1999-filing season did not change.

Results

Despite notable accomplishments, risks continue to affect the integration of ISRP with other IRS operations. Specifically, we identified concerns in the following areas.

• Document processing changes increased taxpayer burden.
• Approval of an enhancement increased implementation risk.
• Contingency plans were incomplete and untested.
• Access to taxpayer information was not properly controlled.

Whenever possible, we provided IRS management with detailed information on each of these findings. We also incorporated their responses and corrective actions into each issue discussed in this report.

Since ISRP operations began, IRS service centers have reported over 2,500 ISRP operating problems as "trouble tickets." As of June 2, 1999, 90 percent of the ISRP trouble tickets had been resolved, including over 98 percent of the highest priority issues. The ISRP Project Office defined high priority issues as problems capable of causing work stoppages.

The IRS’ Product Assurance Division conducted Systems Acceptability Tests (SAT) on ISRP software as it was developed. As of June 2, 1999, the SAT test group had reported over 1,900 problems during 4 phases of testing. Ninety percent of the SAT problems reported had been resolved, including 94 percent of the highest priority issues.

As noted in our prior audit reports, IRS management directed its resources to ensure that the vendor concentrated on the highest priority problems. To review, prioritize, and approve proposed system changes, the IRS formed the ISRP Configuration Control Board (CCB) consisting of IRS Executives and ISRP Project Managers. The IRS submits CCB-approved changes to LMFS as contract modifications. As of August 23, 1999, the ISRP CCB had issued over 300 Configuration Control Decisions (CCD) modifying the original system design. On January 11, 1999, the IRS began operating ISRP’s tax return processing component at the last of 10 service centers. By May 21, 1999, the IRS had processed over 66 million timely filed tax returns and met its initial tax return processing goals.

On January 25, 1999, the IRS began operating ISRP’s payment processing component at the last of the six service centers implementing the system during the 1999-filing season. By April 30, 1999, these 6 service centers had processed over 11 million payments and met their initial deposit goals.

Document Processing Changes Increased Taxpayer Burden

The implementation of ISRP required changes to the IRS’ process for controlling paper tax returns and payments. These changes increased the volumes of tax returns incorrectly sent to storage facilities (warehouses) without being processed. In addition, new research techniques were not as useful or reliable as previous techniques. As a result, the IRS increased taxpayer burden through the issuance of incorrect notices and untimely refunds.

Before ISRP creates computer records of information from tax returns, control numbers are hand stamped on each document, and the tax returns are grouped into "blocks" of up to 100 documents. The control numbers contain serial numbers unique to each tax return within the block of documents. After the tax returns are processed, the IRS sends the documents to warehouses and stores the tax return information in computerized files.

Process Changes Increased Taxpayer Burden

During the ISRP pilot, the AUSC found that data entry operators occasionally failed to process tax returns
(did not enter the tax return data). These unprocessed tax returns occurred because operators were no longer required to manually input the unique serial numbers printed on each return processed.

In order to reduce the number of keystrokes required to process tax returns, ISRP was designed to generate serial numbers automatically. The previous computer system required operators to manually input the serial number of each tax return as it was processed.

On April 30, 1998, the IRS sent LMFS a contract modification to remove the automatic serial numbering feature. However, the IRS later determined that the contract modification was too costly and did not fund the programming change.

In order to determine the significance of ISRP-related numbering problems, we reviewed blocks of tax returns processed through ISRP at three service centers. The documents reviewed were randomly sampled individual income tax returns. These tax returns were processed at the Atlanta Service Center (ATSC), the AUSC, and the Memphis Service Center (MSC) from March 15 through March 19, 1999.

Test results revealed that 23 percent (240 of 1,047) of the blocks reviewed contained ISRP-related control numbering errors. IRS employees corrected the majority of the numbering errors by manually writing the correct control numbers on the tax returns. However, seven of the blocks were sent to IRS warehouses with at least one unprocessed tax return.

Once unprocessed documents are sent to warehouses, it is unlikely that the IRS will discover them through normal work procedures. If the unprocessed documents are not found, the related numbering problems may cause untimely refunds, miscalculated tax due notices, and additional taxpayer burden in correcting these situations. During our test, we kept IRS management apprised of our findings and notified local management each time we identified unprocessed tax returns and/or incorrectly numbered documents (i.e., numbering problems).

Based on these results, we estimate that approximately 7,800 tax returns handled at the 3 service centers reviewed contained numbering problems, and approximately 230 of these tax returns were sent to warehouses without being processed. Since our tests were limited to 3 of the 10 service centers and were conducted during a 5-day non-peak processing period, these estimates understate the nationwide volume of numbering problems and unprocessed tax returns.

The volume of unprocessed and unnumbered tax returns increased significantly over the prior year.
The MSC records from January 1 through May 31, 1999, show that local operations identified and corrected 747 blocks of tax returns with numbering problems. This represents a 332 percent increase over the 173 blocks identified the prior year. These records also show numbering problems occurred in blocks of both individual and business tax returns. Further analysis of the records revealed that in 1999, 51 percent (378 of 747) of the blocks identified with numbering problems contained at least 1 unnumbered and/or unprocessed tax return.

Tax returns also went unprocessed during the 1998 ISRP pilot and caused taxpayer burden. From a limited analysis of the AUSC’s 1998 April and May duplicate tax return reports, we located 20 blocks of tax returns where numbering problems erroneously created duplicate tax return situations. These reports identify instances where more than one tax return has attempted to post to the same tax period of a taxpayer’s account.

We found the numbering problems by reviewing the reports for indications that two duplicate tax returns were processed within the same block of documents. Although taxpayers occasionally file duplicate tax returns, it is rare for the duplicate returns to be processed within the same block of documents.

Further analysis revealed unprocessed tax returns for 15 taxpayers. Four of the 15 taxpayers had requested refunds. At the time of our review, ****1****.

Eleven of the 15 taxpayers sent payments along with the unprocessed tax returns. Six of these taxpayers re-filed copies of their tax returns after receiving notices from the IRS. In addition to burdening the taxpayers to re-file, it took the IRS up to 14 months to settle the accounts.

Auditors developed a quality review tool. To determine if ISRP-related numbering problems could be readily identified, we developed a computer program to detect duplicate tax return transactions within the same block. We tested the program at the ATSC on 20 different days during the 1999-filing season and identified numbering errors in 16 blocks of documents. In each block, ****1****. Therefore, the IRS’ computer records indicated that ****1****.

The ATSC adopted the program as a local report for quality review. The report has allowed ATSC processing units to identify blocks with unprocessed tax returns and resolve numbering problems before they create taxpayer burden.

After the 1998 ISRP pilot, the AUSC prepared the "ISRP Roadmap" as a guidance document for processing and procedural changes. In addition, the IRS incorporated this information into ISRP training sessions and provided each service center copies of the roadmap. The roadmap presents a clear and comprehensive assessment of both the positive and negative aspects of ISRP processing, including its effect on other functions required to resolve problems related to the processing of tax returns and payments (downstream functions).

Although we found the roadmap valuable, it did not provide enough pilot production information to project additional workloads and staffing requirements caused by ISRP operations. For example, the roadmap indicated that ISRP operations would significantly affect the workload of the service center’s Accounting Function, but it did not provide any information regarding the volume and staffing costs of the additional workloads.

We also found that, in many instances, the ISRP Project Office had not directed the AUSC to implement any special procedures to measure the downstream effects of ISRP operations. Because downstream functions do not usually track the source of the work, normal production data was insufficient to measure the specific effects of ISRP changes.

In addition to the information in the roadmap, we found other downstream problems and analyzed production data to determine the additional workloads.

The volume of work increased because of ISRP’s method of assigning document control numbers. The ISRP system allows the IRS to process multiple payment types (i.e., individual or business) in the same block of documents. The control number generated by ISRP includes a code input by the data entry operator to specify the payment type. The code also defines the type of account to which the payment is applied (i.e., individual or business account).

The IRS designed this feature into ISRP to eliminate the need for some manual sorting of documents in the initial stages of processing. The old system could not process multiple payment types at the same time and required the sorting of payments into blocks of like documents. As long as the operators know and correctly enter specific payment codes for each document processed, the ISRP system does not require the sorting of payments.

We discovered that at the MSC the number of incorrectly coded payment transactions increased 338 percent during a 3-month period after the implementation of ISRP. At the MSC, the Reject Function renumbers incorrectly coded payment documents, and from February through April 1998, it renumbered 443 payment documents. During this same period in 1999, the function renumbered 1,941 payment documents. Although the volume of renumbered payments is often attributed to the high turnover of data entry operators, we determined that the multiple payment type processing feature of ISRP also contributed to this significant increase.

The type of work required to resolve payment processing problems changed because ISRP research tools were less effective. Unlike the DIS and the RPS, the ISRP system does not assign matching control numbers to tax returns and payment documents received together. In prior years, IRS employees relied upon the similarities of these control numbers to locate the original tax return, verify the existence of payments, and prevent erroneous notices and refunds.

To improve this situation, IRS management changed the document processing procedures. The new procedures require IRS employees to keep tax returns received with payments separated until computer records are created and coded as received with payments. Although the IRS codes the computer records, these codes do not reference the location or amount of the payment, and the service centers are not required to mark the original tax returns as received with payments.

After the 1998 ISRP pilot operations, the AUSC delayed the billing of 5,285 individual income tax liabilities because of its uncertainty that the payment code had been properly applied. In each case, the computer record of the taxpayer’s return had been coded as received with payment, but no payments had been applied to the taxpayer’s account. Without matching control numbers, the AUSC was no longer able to locate the original documents, in order to verify the receipt of the payments (as discussed above). Therefore, it delayed the notices until all suspended payment transactions had been researched and processed.

We determined that in 60 percent of the cases (3,182 of 5,285) no additional payments were found, and the tax due notices were subsequently issued because the taxpayers had not satisfied their tax liabilities. In some cases, the notices were delayed for up to four months after the balance due situation was originally identified. The AUSC located payments for the remaining 2,103 taxpayers and either stopped the tax due notice or corrected the balance due before the notice was issued.

To prevent this delay during the 1999-filing season, the AUSC implemented a requirement to stamp the tax returns received with payments. The stamp provided downstream functions an indication that payments were received with the tax returns.

In a similar situation, the Ogden Service Center (OSC) developed computer programs to create files containing taxpayer payments not credited to taxpayer accounts. Before the OSC released balance due notices for tax accounts coded as received with payments, tax examiners reviewed these files to determine if any related tax payments had been suspended.

A new ISRP research tool used to resolve problems with payments was not always reliable. The ISRP system creates and stores (archives) computerized pictures (images) of the original payment documents (checks, money orders, payment vouchers, etc.). A set of images consists of pictures of the front of the payment, the back of the payment, and sometimes the front of the payment voucher.

Our review of payments processed at 6 service centers showed that in approximately 1 percent of the payments reviewed the archive system did not capture usable images. We reviewed a random sample of payment transactions processed from January 11 through February 19, 1999. During this period, the 6 service centers processed over 900,000 payments through ISRP.

Thirty-one of the 3,167 payments reviewed had either no image (10 cases) or an incomplete set of images (21 cases) on the archive database. Incomplete image sets contained pictures of either the front or the back of the payment, but not both sides. Instead, the set contained an extra copy of either the payment voucher or one side of the payment document. The front of the payment often contains information regarding the taxpayer and the exact amount paid, and the IRS prints processing information on the back of the payment. Therefore, documentation of the payment transaction is incomplete without images of both sides of the payment, and IRS employees may have to contact taxpayers in order to correct processing problems with these payments.

Based on these results, we estimate that during the processing period tested, ISRP’s image archive database did not capture any image for approximately 2,900 payments and captured incomplete documentation for approximately 6,000 additional payments. Since we conducted the tests during a non-peak processing period, the test results may understate the total volume of imaging errors created during the 1999-filing season.

In a prior report, the IRS Inspection Service (now TIGTA) reported that the ISRP image archive system was incomplete and unreliable. When employees cannot locate information to correct problems with payments, taxpayer contact may be necessary. The IRS agreed with the finding and established the reliability of ISRP’s image archive system as 1 of 10 critical issues requiring resolution before the 1999-filing season.

On March 30, 1999, the IRS closed this critical issue as resolved. We have not tested the system since that date, but documents dated April 2 and May 14, 1999, from two service centers report that images of some payments processed during April were not stored on the archive database.

Computer programs may assess incorrect penalties. Thousands of taxpayers could have been assessed erroneous penalties had we not notified the IRS that the received dates coded into the control numbers were unreliable. In addition to identifying the document, control numbers also contain codes indicating the date the number was assigned. In some circumstances, computer programs use these control numbers to approximate the received date of the tax return when calculating estimated tax (ES) penalties.

For example, IRS regulations provide that taxpayers with income from fishing and/or farming activities are not required to make periodic ES payments if they file their tax returns and pay all taxes due by March 1. The prior system numbered these tax returns when the payments were processed. Since the IRS normally processes payments within 24 hours of the date received, the control number closely approximates the tax return’s received date, allowing the computer programs to calculate ES penalties correctly. After the payments have been removed, these tax returns are usually shelved and input after the higher priority refund returns are processed.

However, ISRP does not assign control numbers to the tax returns at the time the payments are processed. In addition, the IRS decided to wait and assign the control numbers to these tax returns when they were removed from the shelves for input. In some instances, this would have been after March 1, and the computer programs would have incorrectly assessed penalties.

To determine if the IRS had made appropriate changes to prevent erroneous ES penalty notices, we interviewed local management at four of the six service centers processing payments through ISRP. At two of the four service centers, management was not aware of the effect delaying the assignment of the control numbers would have on IRS computer programs.

In reaction to our discussions with local management, the IRS implemented national actions to prevent erroneous penalty notices from generating on taxpayers with fishing and/or farming income by reprogramming the computer’s penalty assessment criteria. These actions prevented the IRS from potentially miscalculating ES penalties and issuing erroneous notices to approximately 1.9 million taxpayers per year, nationwide, filing a Profit or Loss From Farming (Schedule F) on their Individual Income Tax Return (Form 1040).

The IRS should improve the ISRP system’s control and accountability over documents processed by:

1. Changing the method used to assign serial numbers.
2. Implementing additional quality controls, including a check to detect duplicate tax returns within the same block before the documents are sent to storage facilities.
3. Ensuring that the most significant effects of ISRP operations on downstream functions and computer programs are identified, measured, and properly addressed, including the areas identified in the "ISRP Roadmap" and this report. The measurements should include the volume and staffing costs of the additional work caused by changes to the existing control numbering processes.

Management’s Response: IRS management agreed with our findings and has implemented configuration changes which require ISRP operators to manually enter the serial number for each tax return rather than the computer automatically generating the serial numbers. Each site also received additional staff hours to perform this manual task. The staff hours will be automatically included in subsequent years’ resource allotments.

IRS management also implemented a formal quality review process that reviews a random sample of the ISRP operator’s work after the tax return numbering and input processes.

IRS management does not plan to take actions to further identify or measure the effects of IRSP implementation on all impacted operational areas and computer programs. Management indicated that this process was completed and issues were addressed during each stage of the nationwide rollout. In addition, resources needed to analyze and assess the impact resulting from a system change that has been in place for a full season would have to be taken from other necessary programs and projects.

 Approval of an Enhancement Increased Implementation Risk

 

The IRS approved a significant enhancement to the design of ISRP despite delays in the project’s implementation schedule and many unresolved development problems. After formally advising LMFS of concerns regarding the timely delivery of ISRP, the IRS authorized a significant enhancement to the method for transferring data from service centers to data processing centers at remote locations.

In correspondence dated October 29, 1998 (PTD-99-022), the IRS provided LMFS with a statement of work requesting a technical analysis for migrating the current ISRP architecture to an electronic data transfer environment. The letter authorized work to begin immediately and subjected the vendor to a $100,000 price ceiling. The purpose of the request was to thoroughly explore file transfer protocol alternatives to the current magnetic media (magnetic tape) data transfer processes. The ISRP system was originally developed to produce magnetic tapes that could be carried to other computer systems in the same general location. The proposed enhancement would have directly connected ISRP to these computers and allowed the direct transfer of this information.

However, in the same month this task was added to the contract, the IRS formally advised LMFS of concerns about its ability to meet the original contract requirements. To support these concerns, the IRS cited untimely delivery of proposals to previous change requests, unacceptable delays in the delivery of program corrections, and a lack of vendor support staff with in-depth knowledge of the system.

In an Audit Memorandum dated November 25, 1998 (see Appendix VII), the IRS Inspection Service (now TIGTA) advised the IRS that the development of this enhancement may increase the risk of untimely delays in the vendor’s delivery of a fully functional and operationally ready ISRP system.

Recommendation

The IRS should:

4. Rescind the additional contract requirement.

Management’s Response: In a response dated December 13, 1998 (see Appendix VIII), IRS management agreed with the findings and recommendations of our Audit Memorandum and outlined their actions to delay development of the ISRP enhancement until after the 1999-filing season. However, after further review of the tasks and risks associated with the development for the enhancement, IRS management decided to terminate the proposed contract modification to develop the enhancement. IRS management sent a Notification of Termination to the contractor on September 22, 1999.

Contingency Plans Were Incomplete and Untested

In preparation for the 1999-filing season, each service center prepared plans estimating the volume of tax returns it expected to receive and the staffing necessary to process those tax returns within certain time frames. They also developed alternative plans (contingency or back-up plans) to be implemented in case of unexpected problems.

On January 29, 1999, we issued a memorandum to the IRS (see Appendix IX) expressing our concerns with the:

• Lack of detailed contingency plans.
• Limitations budget and staffing reductions may have on contingency alternatives.
• Omission of the ISRP system from local contingency plans.
• Reliability of back-up plans for problems with ISRP that may occur after the year 2000.

In August 1998, the IRS National Office requested that each service center evaluate its ISRP Contingency Plan for the 1999-filing season. Eight of nine service centers responded with contingency plans. One service center had not developed a plan, and the AUSC was not queried since it had already operated ISRP during the 1998-filing season.

The local contingency plans called for re-installing legacy equipment as needed, but none of the plans contained detailed scenarios on how this would be accomplished. Four of the eight centers commented that capabilities to re-install legacy equipment were dependent upon the services of other IRS personnel and/or vendor technicians.

Without the support of detailed scenarios, the contingency plans did not provide assurances that coordination among on-site functions had been accomplished. In addition, the ISRP implementation schedule did not allow time to test the contingency plans. Testing contingency plans for business resumption and disaster recovery evaluates their likelihood for success.

Service centers undergo a filing season readiness process to ensure they can meet the program completion dates (PCDs) for processing timely filed individual tax returns. These dates are established as deadlines for processing tax returns. As part of this process, management prepares contingency plans that include the transfer of excess tax returns to other service centers for processing (transshipment) as one of the primary methods of ensuring that all taxpayer submissions are processed before PCDs.

In a prior report, the IRS Inspection Service (now TIGTA) questioned Fiscal Year (FY) 1999 budget savings based upon projected ISRP productivity gains. The operating budgets of each service center were affected by a 10 percent returns processing productivity rate gain established in the original ISRP business case. Their FY 1999 budgets sustained a $9 million reduction, nationwide, because of the projected gain. Consequently, each service center was required to plan 10 percent improvements in the processing of every type of tax return.

During our review, all service centers expressed confidence in the readiness of ISRP. However, all expressed concern about their ability to meet the 10 percent productivity gain. Several service centers expressed concern that if the savings were not realized, local recruiting and training needs may not be fully funded. In general, the service centers were concerned with the lack of validation of the projected ISRP processing cost savings. In particular, the Fresno Service Center (FSC) indicated that these cuts could have a negative effect on its ability to process transshipped tax returns during the April peak processing period.

To be prepared for unexpected problems, the service centers are required to prepare local business resumption plans. These plans define back-up systems and establish disaster recovery procedures designed to ensure that IRS operations continue despite unexpected problems.

Under a Memorandum of Understanding among IRS functions, the Service Center Directors share responsibility with the Executive Officer for Service Center Operations (EOSCO) to incorporate ISRP contingency plans into the local business resumption plans.

Our review of ISRP contingency plans prior to the 1999-filing season found that only 2 of the 10 service centers had included ISRP in their local business resumption plans. Although other service centers were aware of the need to include ISRP, they had planned to do so later. At the time of our review, all 10 service centers were operating the ISRP system’s tax return processing components and 6 service centers were operating the ISRP system’s payment processing component. Until the local contingency plans are updated and properly tested at all 10 service centers, the IRS’ risk of processing problems remains high.

None of the local ISRP contingency plans reviewed contained scenarios beyond the 1999-filing season. The General Accounting Office (GAO) noted this condition in its report on the IRS’ 1998 Filing Season and recommended that the IRS develop a contingency plan for ISRP that provides for the possibility of a system-wide failure of the remittance processing function past 1999. The IRS disagreed with this recommendation and responded that (1) normal disaster recovery procedures would be in place in case extended downtime occurred with the remittance processing equipment and (2) it would have a system in place to direct payments to private banks (Lockbox Banks) as needed.

In an Audit Memorandum dated January 29, 1999 (see Appendix IX), we recommended that the IRS maintain an existing piece of equipment (RPS-II) because of its availability at all 10 service centers and the ability to upgrade it for year 2000 processing. Although IRS management disagreed with this recommendation, one service center used this equipment to process payments for a short period of time during the 1999-filing season when the ISRP system was not fully functional.

1. The IRS should improve ISRP 1999 contingency plans to ensure taxpayers are provided quality service in case of unexpected production problems.
2. The IRS should finalize and test contingency plans for year 2000 payment processing.

Management’s Response: On March 11, 1999, the IRS provided a detailed response to our January 29, 1999, Audit Memorandum (see Appendix X). In response to our concerns regarding the 1999 contingency plans, the IRS agreed with our findings and took the following actions:

• Developed a blanket ISRP business resumption template.
• Worked with each service center to ensure that the local business resumption plans were properly updated.
• Had each service center develop detailed contingency plans to revert to the old processing systems.
• Tested these plans before the peak processing period of the 1999-filing season.
• Instructed each service center to hire and train sufficient staffing to meet PCDs regardless of the established work plans.

Office of Audit Comment: We are still concerned with the IRS’ year 2000 back-up payment processing capabilities. If detailed contingency plans are not both finalized and tested, unexpected problems with the IRS’ payment processing systems may result in untimely deposits, unprocessed payment transactions, and errors in calculating tax due balances. As of October 14, 1999, the IRS had neither completed negotiations with the Lockbox Banks nor tested the contingency plans currently in place.

Management’s Response: IRS management amended the FY 2000 Memorandum of Understanding between the IRS and the Financial Management Service to reflect that Lockbox banks will serve as backup/alternate for remittance processing and deposit activities in the event that ISRP is not able to process taxpayer payments. Management also indicated that "this arrangement was tested as part of the Y2K contingency plan".

Access to Taxpayer Information Was Not Properly Controlled

Due to the personal and sensitive nature of taxpayer information on IRS computer systems and the legal requirements to safeguard taxpayer privacy, all individuals accessing IRS computers must have successfully completed a personal background investigation. This includes employees of vendors contracted to implement and maintain ISRP. In addition, IRS computer systems containing taxpayer information are required to maintain audit trails (records) of accesses to the system. The audit trail is used to evaluate the appropriateness of system access and use.

We reviewed the background investigation results for 51 vendor employees at the Andover, Atlanta, Brookhaven, Memphis, Kansas City and Ogden Service Centers. We found that 59 percent of these employees (30 of 51) did not have completed background investigations.

This problem was compounded by the IRS requirement that the vendor augment on-site support for ISRP during the filing season. Consequently, the vendor hired 40 additional specialists who required background investigations before accessing the system.

According to the ISRP Project Office, the Minimum Background Investigation (MBI) packages were funded and transferred to the National Background Investigations Center (NBIC) as soon as the requests were received from the service centers. However, according to the ISRP Project Office, delays in the IRS accounting process and NBIC procedures prevented NBIC’s timely initiation of these requests.

In addition, discussions with local system security officers at two service centers revealed that they had not received instructions or training on how to review the audit trail information. According to the ISRP Project Office, data security responsibilities were assigned to ISRP system security administrators. However, subsequent audit tests designed to review the content of ISRP audit trail information found that the audit trail records at one of these two service centers had not been properly maintained and were unreadable.

Although we did not identify any misuse of taxpayer data, vendor employees with the computer knowledge and skills necessary to misuse this data were allowed access to ISRP before the completion of their background investigations.

The IRS should improve ISRP controls designed to safeguard taxpayer information by:

1. Completing background investigations on vendor employees before granting them access to ISRP.
2. Reviewing the coordination and assignment of ISRP data security responsibilities.

Management’s Response: IRS management agreed that background investigations should be completed on vendor employees before granting them access to ISRP. However, the process of obtaining a MBI is time consuming due to the accounting procedures and processing constraints of the NBIC. Therefore, the Contracting Officer Technical Representative’s Acquisition Office, in conjunction with NBIC, established a streamlined policy that allowed a 5-day review to look for adverse findings while awaiting the results of the MBI. In addition, each service center was kept informed of the status of the pending MBI while following local procedures regarding that individual’s access to taxpayer data.

IRS management also indicated that they "funded the training for 6 or more local technicians in each submission processing center to be NT 4.0 and NT 3.51 System Administrators. The NT administrative coursework was extensive and included the function specified in this recommendation. The Project Office has recommended that the System and Security logs be checked periodically by a Systems Administrator using Event Viewer, which is a widely recognized tool used for IRS log file review. This procedure is documented in the ISRP Systems Management Guide for Systems Administrators."

Conclusion

The IRS implemented portions of the ISRP system nationwide and completed the 1999-filing season. Despite these accomplishments, significant risks continued to affect the integration of ISRP with other IRS operations. The IRS needs to ensure that ISRP changes do not create taxpayer burden and that additional work requirements are adequately identified and measured. The IRS also needs to ensure that access to the ISRP system is properly controlled and that adequate plans to recover from unexpected processing problems are in place before the 2000-filing season.

Appendix I

The overall objectives of this audit were to determine whether (1) effective management processes were in place to ensure that problems identified during the Integrated Submission and Remittance Processing (ISRP) System’s pilot were adequately corrected, (2) the Internal Revenue Service (IRS) adequately prepared for the implementation of the system, and (3) necessary actions were taken to ensure a successful filing season in 1999. We also followed up on the findings and recommendations of prior audit reports issued in January and November 1998. To accomplish this objective, we:

1. Evaluated information supporting the IRS’ critical rollout decision.
1. Surveyed the Directors and Processing Division Chiefs at each of the 10 service centers [Andover (ANSC), Atlanta (ATSC), Austin (AUSC), Brookhaven (BSC), Cincinnati (CSC), Fresno (FSC), Kansas City (KCSC), Memphis (MSC), Ogden (OSC), and Philadelphia (PSC) Service Centers] to identify their concerns and to determine local involvement with the ISRP implementation.
2. Attended both local and IRS National Office project meetings to determine if local concerns were addressed at the national level.
3. Evaluated the impact of critical system development decisions by reviewing the minutes of the ISRP Configuration Control Board (CCB) meetings and evaluating the prioritization and supporting documentation of critical Configuration Change Decisions (CCD).
4. Determined if all concerns were resolved or if alternative plans were developed by reviewing documentation supporting certification of the ISRP system’s operational readiness at all service centers.
2. Evaluated the effectiveness of problem tracking and resolution controls.
1. Determined whether controls assured the resolution of problems by reviewing management’s processes, procedures, and preparations to record, prioritize, and resolve both local and national problems.
1. Compared problem reporting documentation to criteria established for prioritizing and resolving problems.
2. Reviewed the status of ISRP-related problems reported as of
   October 30, 1998, during three phases of Systems Acceptability Testing (SAT) activities and one phase of pilot operations at the AUSC.
2. Determined if ISRP’s interfaces to other computer systems were adequately tested.
1. Identified major systems and projects interfacing with ISRP (e.g., the Generalized Mainline Framework (GMF) and the Service Center Mainframe Consolidation project) and reviewed documentation to support tests related to these systems.
2. Reviewed documentation regarding system security and capacity tests.
3. Reviewed listings of tax forms certified by SAT as ready for processing through the ISRP system.
3. Assessed the effect of the 1998 AUSC ISRP pilot operations and quantified ISRP’s effect on other service center functions required to resolve tax return and payment processing problems (i.e., downstream functions).
1. Identified and reviewed the IRS’ efforts to measure the system’s effect on downstream functions.
2. Interviewed local managers, technicians, and analysts to determine how the AUSC tracked the downstream impact of ISRP.
3. Quantified the effect pilot operations had on AUSC downstream functions.
1. Reviewed 1997 and 1998 Work Performance and Cost (WP&C) reports for the periods ending June 30 and September 30.
2. Reviewed 1997 and 1998 Program Analysis Reports for the periods ending June 30 and September 30.
3. Reviewed 1997 and 1998 audit trail reports for the 5-month periods ending September 30.
4. Reviewed Unidentified Remittance Reports prepared by an AUSC program analyst.
5. Analyzed the June and December 1998 data on the AUSC Unidentified Remittance File.
6. Reviewed procedures developed to assist employees in resolving and securing source documents for payment tracer cases.
7. Reviewed local reports from AUSC managers in the Automated Underreporter Function to substantiate additional hours incurred due to ISRP changes.
8. Compared volumes of remittances processed at the AUSC during identical periods in 1997 and 1998.
4. Evaluated the effect of the ISRP pilot operations on existing controls designed to ensure the timely processing of records transmitted from ISRP to GMF programs.
1. Identified and discussed work stoppages with the AUSC Submission Processing Division Chief, Information Systems Division Chief, Quality Assurance and Management Support Division Chief, Computer Operations Branch Chief, the Manager of the Revenue Accounting Control System Unit, and the ISRP System Support Staff.
2. Reviewed the Internal Revenue Manual for accounting and balancing control procedures over processed documents.
3. Reviewed AUSC Accounting Branch GMF run-to-run control reports.
4. Discussed balancing control procedures with managers and employees in the AUSC Data Control and Balancing Units.
5. Reviewed the December 1998 and January 1999 deposit tickets and reports.
4. Assessed the adequacy and accuracy of the ISRP system functionality implemented for the 1999-filing season at each of the 10 service centers.
1. Interviewed various levels of service center management (e.g., Directors, Division Chiefs, Processing Division Branch Chiefs, and various Processing Division Unit Managers) to gather information on concerns or problems related to ISRP operations.
2. Conducted tests at each service center implementing the ISRP Distributed Input System functionality during the 1999-filing season.
1. At each service center, randomly sampled tax returns processed through the ISRP system and determined if they accurately posted to the taxpayers’ accounts.
2. At various service centers, evaluated ISRP system downtime, the use of legacy systems to supplement ISRP processing, and their ability to process tax returns transshipped from other service centers.
3. Conducted the following tests at each service center implementing the ISRP Remittance Processing System functionality during the 1999-filing season to determine the integrity of the remittance transaction images stored on the image archive system.
1. Defined the population as 902,849 remittance transactions processed through ISRP from January 11, 1999, through February 19, 1999, at the ANSC, AUSC, BSC, KCSC, MSC, and OSC.
2. Reviewed a sample of ISRP remittance transactions.
1. Selected a total random sample of 3,167 remittance transactions. The sample selection was distributed as follows:
1. Randomly selected and reviewed 527 transactions from a population of 92,305 transactions at the ANSC.
2. Randomly selected and reviewed 528 transactions from a population of 170,424 transactions at the AUSC.
3. Randomly selected and reviewed 528 transactions from a population of 142,456 transactions at the BSC.
4. Randomly selected and reviewed 528 transactions from a population of 170,884 transactions at the KCSC.
5. Randomly selected and reviewed 528 transactions from a population of 153,249 transactions at the MSC.
6. Randomly selected and reviewed 528 transactions from a population of 173,531 transactions at the OSC.
2. Reviewed the sample of remittance transactions for any digital pictures of the source documents on the ISRP image archive system. In total, the final test results revealed that 10 (0.32 percent) of the 3,167 transactions reviewed did not contain digital pictures of the source documents on the ISRP image archive system (missing archive images). Based upon these results, we are 95 percent confident that approximately 2,889 (plus or minus less than 0.01 percent) of the 902,849 transactions within the population were missing archive images. The test results were distributed as follows:
1. At the ANSC, the final test results revealed that 1 (0.19 percent) of the 527 transactions reviewed were missing archive images.
2. At the AUSC, the final test results revealed that 4 (0.76 percent) of the 528 transactions reviewed were missing archive images.
3. At the BSC, the final test results revealed that 4 (0.76 percent) of the 528 transactions reviewed were missing archive images.
4. At the KCSC, the final test results revealed that none of the
   528 transactions reviewed were missing archive images.
5. At the MSC, the final test results revealed that none of the
   528 transactions reviewed were missing archive images.
6. At the OSC, the final test results revealed that 1 (0.19 percent) of the 528 transactions reviewed were missing archive images.
3. Reviewed the sample of remittance transactions for a complete set of corresponding digital pictures of the source documents (sets of remittance images). In total, the final test results revealed that 21 (0.66 percent) of the 3,167 transactions reviewed contained incomplete sets of remittance images. Based upon these results, we are 95 percent confident that approximately 5,959 (plus or minus less than 0.01 percent) of the
   902,849 transactions within the population contain incomplete sets of remittance images. The test results were distributed as follows:
1. At the ANSC, the final test results revealed that none of the
   527 transactions reviewed contained incomplete sets of
   remittance images.
2. At the AUSC, the final test results revealed that 9 (1.70 percent) of the 528 transactions reviewed contained incomplete sets of remittance images.
3. At the BSC, the final test results revealed that 10 (1.89 percent) of the
   528 transactions reviewed contained incomplete sets of
   remittance images.
4. At the KCSC, the final test results revealed that none of the
   528 transactions reviewed contained incomplete sets of
   remittance images.
5. At the MSC, the final test results revealed that 2 (0.38 percent) of the 528 transactions reviewed contained incomplete sets of
   remittance images.
6. At the OSC, the final test results revealed that none of the
   528 transactions reviewed contained incomplete sets of
   remittance images.
4. Identified and monitored at various service centers specific issues affecting ISRP operations during the 1999-filing season.
1. Compared current volumes of unprocessed and unnumbered tax returns at the MSC to prior year data.
2. Interviewed local management at the AUSC, KCSC, ANSC, and MSC to determine if ISRP processing changes would result in the assessment of erroneous penalties.
3. Reviewed the results of background investigations on ISRP vendor employees at the ANSC, ATSC, BSC, KCSC, MSC, and OSC to determine if access to taxpayer information through ISRP was adequately controlled.
5. Determined if the ISRP automatic serial numbering feature contributed to
   ISRP-related document control numbering problems and/or unprocessed tax returns sent to storage facilities.
1. Defined the population as 34,211 blocks of paper Individual Income Tax Returns (Forms 1040, 1040A, 1040EZ, and 1040PC) processed through ISRP at the ATSC, AUSC, and MSC, for which from March 15 to March 19, 1999, at least 1 of the tax return transactions within the block attempted to post to the Masterfile.
2. Reviewed a sample of blocks of ISRP tax return transactions.
1. Selected a total random sample of 1,047 blocks of tax returns. The sample selection was distributed as follows:
1. Randomly selected and reviewed 347 blocks of tax returns from an ATSC population of 13,125 blocks.
2. Randomly selected and reviewed 350 blocks of tax returns from an AUSC population of 10,226 blocks.
3. Randomly selected and reviewed 350 blocks of tax returns from an MSC population of 10,860 blocks.
2. Reviewed the sample of blocks of ISRP-processed tax returns to determine if all of the tax returns had been processed before the block was sent to storage facilities. In total, the final test results revealed that
   7 (0.67 percent) of the 1,047 blocks reviewed were sent to storage facilities with at least 1 unprocessed tax return. Based upon these results, we are 95 percent confident that approximately 229 (plus or minus
   0.61 percent) of the 34,211 blocks of tax returns within the population were sent to storage facilities with at least 1 unprocessed tax return. The test results were distributed as follows:
1. At the ATSC, the final test results revealed that 4 (1.15 percent) of the 347 blocks reviewed were sent to storage facilities with at least
   1 unprocessed tax return.
2. At the AUSC, the final test results revealed that 2 (0.57 percent) of the 350 blocks reviewed were sent to storage facilities with at least
   1 unprocessed tax return.
3. At the MSC, the final test results revealed that 1 (0.29 percent) of the 350 blocks reviewed were sent to storage facilities with at least
   1 unprocessed tax return.
3. Reviewed the sample of blocks of ISRP-processed tax returns to identify tax returns with ISRP-related numbering problems. In total, the final test results revealed that 240 (22.92 percent) of the 1,047 blocks reviewed contained at least 1 tax return with an ISRP-related numbering problem. Based upon these results, we are 95 percent confident that approximately 7,841 (plus or minus 2.52 percent) of the 34,211 blocks of tax returns within the population contain numbering problems. The test results were distributed as follows:
1. At the ATSC, the final test results revealed that 96 of the 347
   (27.67 percent) blocks reviewed contained at least 1 tax return with an ISRP-related numbering problem.
2. At the AUSC, the final test results revealed that 49 of the 350
   (14.00 percent) blocks reviewed contained at least 1 tax return with an ISRP-related numbering problem.
3. At the MSC, the final test results revealed that 95 of the 350
   (27.14 percent) blocks reviewed contained at least 1 tax return with an ISRP-related numbering problem.
5. Followed up on prior audit findings.
1. Furnished management with data analysis to quantify the cost benefit of corrective actions necessary to improve remittance processing controls.
2. Determined the impact that ISRP productivity gains had on each service center’s 1999-filing season budgets and work plans.
3. Evaluated the IRS’ readiness to execute ISRP contingency plans.
1. Identified and applied the appropriate criteria for the development and implementation of contingency plans by reviewing the sufficiency and detail of national and local contingency plans.
2. Assessed the coordination of local and national contingency activities.
3. Evaluated recovery procedures for system failures during production.
4. Determined if ISRP contingency plans were incorporated into the service center disaster recovery plans.

Appendix II

Walter E. Arrison, Associate Inspector General for Audit (Wage and Investment Income Programs)

M. Susan Boehmer, Director

Terry W. Black, Audit Manager

Robert A. Olson, Senior Auditor

Paul R. Baker, Auditor

Nelva U. Blassingame, Auditor

Michelle D. Brasfield, Auditor

Adrienne D. Byers, Auditor

Thomas W. Dalton, Auditor

Charlene L. Elliston, Auditor

Roy J. Evans, Auditor

Jackie E. Forbus, Auditor

Steven W. Gibson, Auditor

Perrin T. Gleaton, Auditor

John M. Hannon, Auditor

Areta G. Heard, Auditor

Kathy D. Henderson, Auditor

Robert J. Howes, Auditor

Frank W. Jones, Auditor

Richard A. Lambrecht, Auditor

Donald J. Martineau, Auditor

Bill H. Richards, Auditor

Lynn A. Rudolph, Auditor

Anthony A. Saranchak, Auditor

Sharon R. Shepherd, Auditor

Wallace C. Sims, Auditor

Lawrence R. Smith, Auditor

Ronald C. Swallow, Auditor

Roy E. Thompson, Auditor

Larry J. Wyrick, Auditor

Joseph C. Butler, Computer Specialist

Appendix III

Deputy Commissioner Modernization C:DM

Deputy Commissioner Operations C:DO

Chief Information Officer IS

Chief Management and Finance M

Chief Operations Officer OP

Chief Financial Officer M:CFO

Deputy Chief Information Officer, Operations IS

Deputy Chief Information Officer for Information Resources Management IS:IR

Assistant Commissioner (Customer Service) OP:C

Assistant Commissioner (Forms and Submission Processing) OP:FS

Assistant Commissioner (Procurement) M:P

Assistant Commissioner (Service Center Operations) IS:SC

Assistant Commissioner (Systems Development) IS:S

Director, Program Evaluation and Risk Analysis M:O

Executive Officer for Service Center Operations OP:SC

Director, Century Date Change Project Office IS:CD

Director, Submission Processing Division IS:S:SP

National Director for Budget M:CFO:B

National Director, Submission Processing OP:FS:S

Director, Andover Service Center D

Director, Atlanta Service Center D

Director, Austin Service Center D

Director, Brookhaven Service Center D

Director, Cincinnati Service Center D

Director, Fresno Service Center D

Director, Kansas City Service Center D

Director, Memphis Service Center D

Director, Ogden Service Center D

Director, Philadelphia Service Center D

Project Manager, Integrated Submission and Remittance Processing System IS:S:SP:I

Project Manager, Integrated Submission and Remittance Processing System OP:FS:S

National Director for Legislative Affairs CL:LA

Office of Management Controls M:CFO:A:M

Office of the Chief Counsel CC

 Audit Liaisons:

Appendix IV

This appendix presents detailed information on the measurable impact that our recommended corrective actions have on tax administration. These benefits will be incorporated into our Semiannual Report to the Congress.

The implementation of the Integrated Submission and Remittance Processing (ISRP) System required changes to the Internal Revenue Service’s (IRS) system of processing tax returns and payments. Tax returns and payments are normally processed in "blocks" of up to 100 documents. Control numbers are assigned to each tax return processed and contain unique serial numbers identifying each document within the block.

During the 1998 ISRP pilot operations, the Austin Service Center (AUSC) found that data entry operators occasionally failed to process tax returns (did not enter the tax return data). These unprocessed tax returns resulted because operators were no longer required to manually input the unique serial number of each return processed. In addition, Memphis Service Center (MSC) records showed that control numbering problems had significantly increased since the implementation of ISRP. Subsequent audit tests confirmed that ISRP process changes did result in control numbering errors and unprocessed tax returns during the 1999-filing season (see page 5).

In this report, we recommend that IRS management change the method used to assign serial numbers and implement additional quality controls. These recommendations would allow service centers to identify and correct numbering problems before taxpayer burden occurs (see page 14).

Reliability of Information (Actual)

Taxpayer Burden/Taxpayer Rights and Entitlements (Actual)

If these recommendations had been implemented at the 3 service centers during the 1999-filing season, the IRS could have prevented over 230 tax returns from being sent to warehouses without being processed and over 7,800 document control numbering problems resulting in additional work to correctly process the documents.

Since many of the IRS’ production reports used to plan and monitor submission processing activities are based upon counts of the control numbers assigned to the transactions processed, document control numbering problems undermine the reliability of this information. If these recommendations are not implemented nationwide, we judgmentally project that in 1 year of nationwide production, the ISRP system could potentially cause approximately 390,000 document control numbering problems and result in the IRS sending approximately 11,000 tax returns to warehouses without being processed.

In order to determine the significance of ISRP-related numbering problems, we reviewed blocks of tax returns processed through ISRP at 3 of the 10 service centers during a 5-day non-peak processing period. The documents reviewed were randomly sampled individual income tax returns processed at the Atlanta Service Center (ATSC), AUSC, and MSC from March 15 through March 19, 1999.

Test results revealed that 22.92 percent of the blocks we reviewed (240 of 1,047 blocks) contained tax returns that were incorrectly numbered. Although the IRS detected and corrected the majority of the numbering problems, seven of the blocks were sent to IRS warehouses with at least one unprocessed tax return. During our review, we kept IRS management apprised of our findings and notified local management each time we identified unprocessed tax returns and/or incorrectly numbered documents (i.e., numbering problems).

Based on these results, we estimate that from March 15 through March 19, 1999, approximately 7,800 tax returns handled at the 3 service centers reviewed contained numbering problems, and approximately 230 of these tax returns were sent to warehouses without being processed. Since our tests were limited to 3 of the 10 service centers and conducted during a 5-day non-peak processing period, these estimates understate the nationwide volume of numbering problems and unprocessed tax returns.

By applying the results of our sample to the expected yearly production presented in the ISRP business case, we can judgmentally estimate the potential nationwide volumes of numbering problems per year. The ISRP business case expects the system to process approximately 170 million paper tax returns per year. Since the IRS processes tax returns in blocks not to exceed 100 documents, we based our estimate on the minimum amount of blocks (1.7 million) required to process the expected volume of tax returns (170 million documents divided by 100 documents per block).

As a result, we judgmentally estimate that if the occurrence rate of sampled blocks remained consistent for 1 full year of nationwide production, ISRP could potentially cause approximately 390,000 document control numbering problems and result in the IRS sending approximately 11,000 tax returns to warehouses without being processed. Since the sample selection was limited to 3 of the 10 service centers and conducted during a 5-day non-peak processing period, the nationwide annual estimate is purely judgmental because its statistical confidence and precision cannot be determined.

IRS regulations provide that taxpayers with income from fishing and/or farming are not required to make periodic estimated tax (ES) payments if they file their tax returns and pay all taxes due by March 1. Since the IRS normally processes payments within 24 hours of the date received, the control number closely approximates the received date, allowing the IRS’ computer programs to calculate ES penalties correctly. After the payments have been removed, these returns are usually shelved and input after the highest priority refund returns are processed. However, ISRP does not assign control numbers to the tax returns at the time the payments are processed. In addition, the IRS decided to wait and assign the control numbers to these tax returns when they were removed from the shelves for input. In some instances, this would have been after March 1, and the computer programs would have incorrectly assessed penalties. Taxpayers would have been assessed erroneous penalties had we not notified the IRS that the received dates coded into the control numbers were unreliable (see page 13).

In reaction to our discussions with local service center management, the IRS initiated national actions to prevent erroneous penalty notices from generating on taxpayers with fishing and/or farming income by reprogramming the computer’s penalty assessment criteria (see page 14).

Taxpayer Burden (Potential)

Implementation of this recommendation prevented the IRS from potentially miscalculating ES penalties and issuing erroneous notices to approximately 1.9 million taxpayers per year, nationwide, filing a Profit or Loss From Farming (Schedule F) on their Individual Income Tax Return (Form 1040).

Since taxpayers with fishing and/or farming income are required to file Forms 1040 Schedule F to report their income or loss, we based the value of the benefit on the annual average number of Forms 1040 Schedules F filed, nationwide, over the past four years. We used the following data from the IRS’ 1999 taxpayer usage study in our calculation:

Based upon this data, the IRS has processed an annual average of 1.9 million Forms 1040 Schedule F from 1995 through 1998.

On November 25, 1998, we issued a memorandum to the IRS advising it of the potential risks of implementing a significant enhancement to the design of the ISRP system. The IRS approved the enhancement despite delays in the project’s implementation schedule, unresolved development problems, and after formally advising the vendor of concerns regarding the timely delivery of ISRP. The additional requirement increased the risk of delays in the vendor’s delivery of the system. In a memorandum dated December 13, 1998, IRS executives advised us they had decided to delay development of the enhancement. However, after further review of the tasks and risks associated with the development for the proposed enhancement, IRS management decided to terminate the contract to develop the enhancement. A Notification of Termination of the proposed contract modification was sent to the contractor on September 22, 1999. (See page 16.)

Cost Savings from Funds Put to Better Use (Actual)

The IRS avoided the expenditure of $100,000 in costs to modify the ISRP contract (TIRNO-94-D-00028).

In correspondence dated October 29, 1998 (PTD-99-022), the IRS provided Lockheed Martin Federal System (LMFS) with a statement of work requesting a technical analysis for migrating the current ISRP architecture to an electronic data transfer environment. The letter authorized work to begin immediately and subjected the vendor to a $100,000 price ceiling. The purpose of the request was to thoroughly explore file transfer protocol alternatives to the current magnetic media (magnetic tape) data transfer processes.

On January 29, 1999, we issued a memorandum to the IRS expressing our concerns with the following (see page 17):

• Lack of detailed contingency plans.
• Limitations budget and staffing reductions may have on contingency alternatives.
• Omission of the ISRP system from local contingency plans.
• Reliability of back-up plans for problems with ISRP that may occur after the year 2000.

On March 11, 1999, the IRS provided a detailed response to our January 29, 1999, Audit Memorandum. In response to our concerns regarding the 1999 contingency plans, the IRS agreed with our findings and took the following actions (see page 21):

• Developed a blanket ISRP business resumption template.
• Worked with each service center to ensure that the local business resumption plans were properly updated.
• Had each service center develop detailed contingency plans to revert to the old processing systems.
• Tested these plans before the peak processing period of the 1999-filing season.
• Instructed each service center to hire and train sufficient staffing to meet program completion dates (PCDs) regardless of the established work plans.

Taxpayer Burden/Taxpayer Rights and Entitlements (Potential)

Reliability of Information (Actual)

If the IRS had not developed and tested ISRP contingency plans, the accurate and timely processing of approximately 170 million paper tax returns and 50 million payments per year, as presented in the ISRP business case, would have been at risk of unexpected processing problems. As a result, the IRS actions affected over 100 million taxpayers expected to file paper tax returns with the IRS.

If the IRS’ had not transferred approximately 290 staff years (approximately $9 million) from other programs, its submission processing operations would have been under funded and may not have met critical PCDs. The under funding occurred because unsubstantiated ISRP productivity gains were factored into the IRS’ Fiscal Year
(FY) 1999 submission processing budgets.

Per the ISRP project development business case, the system is replacing computers used to process approximately 170 million paper tax returns and 50 million payments received each year. Our estimate of over 100 million taxpayers affected is based upon the 1998 calendar year volume of paper Forms 1040, 1040A, 1040EZ, and 1040PC tax returns received and processed at the 10 service centers. This estimate assumes a strong correlation between the number of individual taxpayer accounts and the number of current year Form 1040 series tax returns filed per year. In a memorandum dated June 11, 1998, the Commissioner was presented estimated cost savings related to the effects of productivity gains from ISRP on the Submission Processing Division budgets. The estimate, prepared by the Chief Financial Officer’s Financial Analysis Division, projected that ISRP productivity gains could reduce FY 1999 Submission Processing Division budgets by as much as $9 million (290 staff years). This estimate was based upon the 10 percent productivity gain documented in the ISRP Project’s business case.

Due to the personal and sensitive nature of information on IRS computer systems and legal requirements to safeguard the privacy of taxpayer information, the IRS has established procedures to check the background of all individuals granted access to its computer systems. This includes the employees of vendors contracted to implement and maintain ISRP.

We reviewed the background investigation results for 51 vendor employees at 5 service centers. We found that 59 percent of these employees (30 of 51) did not have completed background investigations. Although we did not identify any misuse of taxpayer data, vendor employees with the computer knowledge and skills necessary to misuse this data were allowed access to ISRP before the completion of their background investigations (see page 22).

Protection of Resources/Taxpayer Privacy and Security (Potential)

The IRS’ actions to ensure proper background investigations will reduce the risk of inappropriate access and potential misuse of taxpayer data for approximately 170 million paper tax returns and 50 million payments processed per year through ISRP. The IRS’ actions to improve the privacy and security of taxpayer information processed through the ISRP system will directly affect over 100 million taxpayers expected to file paper tax returns with the IRS.

Per the ISRP project development business case, the system is replacing computers used to process approximately 170 million paper tax returns and 50 million payments received each year. Our estimate of over 100 million taxpayers affected is based upon the 1998 calendar year volume of paper Forms 1040, 1040A, 1040EZ, and 1040PC tax returns received and processed at the 10 service centers. This estimate assumes a strong correlation between the number of individual taxpayer accounts and the number of current year Form 1040 series tax returns filed per year.

Appendix V

Response has been removed due to its size. To see the complete Response, please go to the Adobe PDF version of this report.

 Appendix VI

ANSC – Andover Service Center

ATSC – Atlanta Service Center

AUSC – Austin Service Center

BSC – Brookhaven Service Center

CCB – Configuration Control Board

CCD – Configuration Control Decision

CSC – Cincinnati Service Center

DIS – Distributed Input System

ERB – Executive Review Board

ESC – Executive Steering Committee

EOSCO – Executive Officer for Service Center Operations

ES – Estimated Tax

FSC – Fresno Service Center

FY – Fiscal Year

GAO – General Accounting Office

GMF – Generalized Mainline Framework

IRS – Internal Revenue Service

ISRP – Integrated Submission and Remittance Processing System

KCSC – Kansas City Service Center

LMFS – Lockheed Martin Federal Systems

MSC – Memphis Service Center

NBIC – National Background Investigations Center

OSC – Ogden Service Center

RPS – Remittance Processing System

PCD – Program Completion Date

PSC – Philadelphia Service Center

SAT – System Acceptability Testing

TIGTA – Treasury Inspector General for Tax Administration

WP&C – Work Performance and Cost

Appendix VII

Memorandum has been removed due to its size. To see the complete memorandum, please go to the Adobe PDF version of this report.

Appendix VIII

Response has been removed due to its size. To see the complete Response, please go to the Adobe PDF version of this report.

Appendix IX

Memorandum has been removed due to its size. To see the complete memorandum, please go to the Adobe PDF version of this report.

Appendix X

Response has been removed due to its size. To see the complete Response, please go to the Adobe PDF version of this report.