Objective and Scope

Our objective was to provide, if possible, the Assistant Commissioner (Criminal Investigation) with information which would enable him to answer two congressional questions about returns that the Internal Revenue Service (IRS) rejected from electronic filing (e-file).

We developed the requested information at the Andover Service Center and the National Office. After receiving and installing the necessary computer hardware, successfully locating and obtaining usable data from around the country in late 1998, and familiarizing auditors with software capable of handling large data files, we performed the major parts of the analysis in February and March 1999.

IRS e-file computer systems are not designed to provide the type of management information necessary to answer the congressional questions. To accomplish our objective, we:

• Obtained and analyzed relevant portions of over 237 million records of individual taxpayers who filed returns (e-file or paper) for Tax Years (TY) 1996 and 1997.
• Obtained, combined, and analyzed separate computer files containing a total of over 49 million taxpayer e-file records for TYs 1996 and 1997 from the Andover, Austin, Cincinnati, Memphis, and Ogden Service Centers.
• Developed customized programs and analyses to manipulate and match data and tabulate the results.

Appendix I provides more information on our detailed objective, scope, and methodology. Appendix II lists the major contributors to this report.

Background

Before accepting e-file returns for processing, the IRS subjects them to computerized checks designed to keep out e-file returns having certain questionable characteristics. These characteristics include items that might cause the federal tax system to issue inflated, fraudulent, or erroneous refunds. E-file returns not accepted by the IRS in this way are termed "rejected."

During a May 8, 1997, House Ways and Means Committee hearing on Revenue Protection, the IRS Deputy Commissioner and the Assistant Commissioner (Criminal Investigation) were questioned by Chairman Bill Archer as to the percentage of e-file returns the IRS had rejected and whether these rejected e-file returns had been later filed as paper returns.

The Assistant Commissioner (Criminal Investigation) testified that the percentage of e-file taxpayer returns that the IRS rejects could not be determined because IRS e-file computer systems do not keep track of "rejects" by taxpayer. Instead, the IRS counts "rejects" by occurrence, no matter which taxpayer was involved. For example, e-file returns for the same taxpayer may be resubmitted electronically many times (and rejected many times), at several or all of the five e-file service centers. IRS e-file computer systems count each of the rejected occurrences, even though the same taxpayer’s return is being resubmitted.

The Assistant Commissioner (Criminal Investigation) also advised Chairman Archer that because IRS computer systems do not track rejected e-file returns for later IRS activity, they could not determine if rejected e-file returns were subsequently filed as paper returns.

As a result of the IRS’ inability to respond to the above detailed congressional questions, the Assistant Commissioner (Criminal Investigation) asked the IRS Inspection Service (now Treasury Inspector General for Tax Administration (TIGTA)), for assistance. This request was endorsed by the Chief Operations Officer, who considered the project important enough to offer to fund the purchase of a special computer for the IRS Inspection Service (now TIGTA) personnel to use for the project.

Results

The answers to the two questions asked of IRS executives relating to rejected e-file returns are summarized in the following table:

Additional Information About Rejected Electronic Returns

While the primary thrust of this report is answering the two congressional questions posed to the IRS, we also point out that the methodology and capabilities used to answer these questions could be developed and used by the IRS to address other Revenue Protection Program issues.

For example, the following table provides further Revenue Protection information revealed during this project relating to rejected e-file returns:

The data supporting the above information have been shared with Criminal Investigation personnel.

Management’s Response: The National Director (Tax Refund Fraud) responded that the information provided addressed the two congressional questions posed to IRS executives.

Conclusion

Our review provided quantifiable answers to the congressional questions about the IRS’ Revenue Protection Program. The IRS rejected 9.5 percent of the e-file returns for TY 1996, and 10.2 percent for 1997. Additionally, we found that 290,737 taxpayers whose e-file returns were rejected for 1996 subsequently filed a paper return. The number filed on paper increased to 300,955 for TY 1997. We also noted that the methodology used to develop this information could be used to answer other Revenue Protection questions and problems.

Detailed Objective, Scope, and Methodology

Our objective was to provide, if possible, the Assistant Commissioner (Criminal Investigation) with information which would enable him to answer two congressional questions about returns that the Internal Revenue Service (IRS) had rejected from electronic filing (e-file).

In order to accomplish our objective, we:

1. Designed specific criteria for accessing and compiling data from these records.
2. Coordinated with personnel from the Treasury Inspector General for Tax Administration’s (TIGTA) Data Analysis Section to perfect the computer extract format.

1. Obtained, combined, and analyzed computer files containing over 49 million taxpayer e-file records for TYs 1996 and 1997 from the Andover, Austin, Cincinnati, Memphis, and Ogden Service Centers. To do this we:
1. Contacted personnel at the Office of Electronic Tax Administration and determined that sufficient computerized e-file information was maintained by each of the five service centers that accept e-file returns for TYs 1996 and 1997.
2. Coordinated with Information Systems and Electronic Filing staffs at each of the five service centers that accept e-file returns to obtain copies of computerized files containing e-file records.
3. Performed extensive validity checks to ensure the integrity and accuracy of data included in each set of data. For example, we compared the counts of the computerized files containing e-file records sent by the five Service Centers to the counts of the information received and downloaded onto the TIGTA computer.
4. Maintained a log of all cartridges containing the extracted information sent from the Martinsburg Computing Center to ensure all cartridges were received at the Andover Service Center tape library.
2. Validated the data in the Masterfile extract and in the e-file records against the data in the primary IRS on-line computer system.
1. For the Masterfile, out of the universe of 117 million, we sampled 44 individual tax account records, selected by interval and random sampling, and checked selected fields of the Masterfile records against IRS records.
2. For the e-file database, out of the universe of 24 million, we sampled 5 e-file tax account records, selected by interval sampling, and checked selected fields of the e-file records against IRS records.
3. Developed and executed customized programs and computer instructions, in an advanced database management language, Microsoft Sequential Query Language (SQL), to manipulate, match, and analyze information contained in both sets of data. To do this we:
1. Trained auditors how to design and use the advanced database management language (Microsoft SQL) to devise customized computer instructions.
2. Executed the customized (SQL) computer instructions to identify and then count or compute:
1. The number of unique taxpayers filing electronically (thus avoiding counting duplicate attempts to file electronically, either in the same location or anywhere in the country).
2. The number of unique taxpayers whose e-file returns the IRS had rejected (again, avoiding counting duplicate attempts to file electronically either in the same location or anywhere in the country).
3. The percentage of rejected e-file returns (by unique taxpayers) versus total e-file transmitted returns (by unique taxpayers).
4. The number of rejected e-file returns (by unique taxpayers) that were resubmitted electronically and accepted.
5. The number of rejected e-file returns (by unique taxpayers) successfully resubmitted and accepted as paper tax returns.
6. The number of rejected e-file returns (by unique taxpayers) not resubmitted either electronically or on paper.
4. Tested the intermediate (COBOL) programs to ensure total record counts were complete, and the output matched and accounted for all records received from the e-file service centers. We verified:
1. Total records received.
2. Identification of reject codes detailed on file.
3. Reject code occurrences.
4. Records input.
5. Total unique Social Security Numbers (SSN).
6. Total records containing a reject code of "A," "D," "R" or "blank."
7. Initial transmission resulting in acceptance, rejection, duplication, or other.
8. Total records containing an invalid primary SSN.
9. Total reject codes on file.
10. Total rejected returns where the reject code field is blank.
11. Total primary SSNs whereby the rejected return contained in excess of 15 reject codes.
12. Total reject codes in excess of 15 or primary SSNs where the rejected return contained in excess of 15 reject codes.
5. Tested the auditor-written SQL statements by having two independent auditors not assigned to the project perform the following validations. To do this we:
1. Traced and verified information contained in SQL methodologies.
2. Reviewed the logic of the queries performed.
3. Re-ran the queries to verify the results matched those shown in the report.

Major Contributors to This Report

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

Kerry R. Kilpatrick, Director

Mark A. Nathan, Acting Deputy Director

Donald J. Butler, Audit Manager

Russell P. Martin, Senior Auditor

Lisa Stoy, Senior Auditor

Pamela DeSimone, Auditor

Grace Terranova, Auditor

Eric Guthrie, Computer Program Analyst

Joseph Rosa, Computer Program Analyst

Report Distribution List

Deputy Commissioner Operations C:DO

Chief Compliance Officer OP

Chief Information Officer IS

Assistant Commissioner (Criminal Investigation) OP:CI

Assistant Commissioner (Electronic Tax Administration) OP:ETA

Assistant Commissioner (Program Evaluation and Analysis) M:OP

National Director, Tax Refund Fraud OP:CI:ORF

Director, Andover Service Center

National Director for Legislative Affairs CL:LA

Office of the Chief Counsel CC

Office of Management Controls M:CFO:A:M

Management's Response to the Draft Report