TIGTA Cites Flaws in IRS Correspondence Audit Review System

IMPACT ON TAXPAYERS

An audit is one of the primary enforcement tools the IRS uses to address noncompliance with the tax laws. Because problems with correspondence audits are not always recognized and reported, the IRS may be missing opportunities to reduce the noncompliance that contributes to the Tax Gap and promote tax system fairness among the tax-paying public.

WHY TIGTA DID THE AUDIT

This audit was initiated to determine the accuracy of the results from the National Quality Review System (NQRS) and how management uses the feedback to enhance the quality of correspondence audits. The review is part of our Fiscal Year 2013 Annual Audit Plan and addresses the major management challenge of Tax Compliance Initiatives.

WHAT TIGTA FOUND

The NQRS is designed to provide IRS managers at all levels with estimates of audit quality from a sample of audits to use in identifying areas in which corrective actions are needed. However, TIGTA identified areas that could be strengthened to increase the accuracy of NQRS review results, enhance the ability of managers to identify and address quality problems with correspondence audits, and ensure that the NQRS sample is selected at random.

The auditing standards and NQRS quality measures need to be better aligned. The auditing standards, including the consideration given to significant issues, contain key requirements not evaluated under the NQRS. This can create inconsistencies in how examiners conduct audits and in how the NQRS evaluates the quality of those audits to identify errors.

For example, TIGTA evaluated a statistical sample of 127 of 2,913 correspondence audits that had been reviewed by the NQRS during an 18-month period and found errors with penalty determinations in 65 of the audits (51 percent) that had not been detected and reported by NQRS quality reviewers.

IRS executives and stakeholders should be provided with a more comprehensive snapshot of audit quality so that needed corrective actions can be timely recognized and taken. Only one overall measure of audit quality is currently reported quarterly by the NQRS to IRS executives and other key stakeholders even though as many as 71 items are reviewed.

Finally, the random selection of audits for NQRS review could not be verified. As such, TIGTA was not able to confirm the statistical validity of the NQRS results.

WHAT TIGTA RECOMMENDED

TIGTA recommended that the IRS ensure that 1) the auditing standards align with the NQRS quality measures, 2) a more complete picture of correspondence audit quality is provided to NQRS customers on a regular basis, and 3) audits are selected randomly for NQRS review.

In their response to the report, IRS management agreed with the first two recommendations and disagreed with the third, indicating that they do not have a cost-effective way to allow the randomness of the NQRS case selection process to be verified. Because the IRS's conclusion was reached after the draft report was issued, the underlying details supporting the conclusion were not evaluated. If the sample selection process cannot be verified, the IRS cannot be assured of the statistical validity of NQRS results.

Management's complete response to the draft report is included as Appendix VII. Copies of this report are also being sent to Internal Revenue Service managers affected by the report.

If you have any questions, please contact me or Nancy Nakamura, Assistant Inspector General for Audit (Compliance and Enforcement Operations).

                          Table of Contents

In Fiscal Years 2008 through 2012, IRS statistics show that it conducted almost 5.7 million correspondence audits and, in the process, recommended approximately $40.4 billion in additional taxes. This represents about 77 percent of all audits the IRS conducted of individual income tax returns and about 56 percent of the estimated $72.4 billion in recommended additional taxes resulting from those audits. The responsibility for conducting correspondence audits rests largely with the IRS's Small Business/Self-Employed (SB/SE) Division, which handles complex individual tax returns, and its Wage and Investment Division, which handles simple tax returns filed by individuals reporting wages, interest, dividends, and other investment income.

In contrast to the more detailed and lengthy face-to-face audit at an IRS office or in the field at a taxpayer's place of business, the correspondence audit process is less intrusive, more automated, and conducted by examiners who are trained to deal with less complex tax issues. Because of its automated features and less complex tax issues, the correspondence audit process enables the IRS to reach more taxpayers at a lower cost. The IRS currently conducts correspondence audits in approximately 37 program areas.

Regardless of the program, the audit process typically begins with the IRS mailing a computer-generated letter from one of its campuses to a taxpayer. The letter outlines the examination process, identifies one or more items on the tax return being questioned, and requests supporting information to resolve the questionable items. Once the requested information is returned, examiners review it to determine whether it resolves the questions. If the questions can be sufficiently answered by the information provided, the audit is generally closed without any changes to the tax; if not, the taxpayer is sent a letter requesting more information or indicating a recommended change to the tax.

The taxpayer at this point can do one of the following:

• Agree with the examiner.

• Provide the examiner with clarifying information.

• Appeal the decision to the IRS's Office of Appeals.

Prior Treasury Inspector General for Tax Administration (TIGTA) audits

In June 2010,² TIGTA reported that penalties were not considered and assessed in accordance with IRS policies and procedures for 211 (92 percent) of 229 sample audits closed in Fiscal Year 2008.³ As a result, opportunities were missed to promote the preparation and submission of complete and correct information on tax returns, impose an economic cost on those who chose not to voluntarily comply with Federal tax laws, and increase revenue for the Department of the Treasury by an estimated $17.5 million.

In February 2010, TIGTA reported that tests of a statistical sample of 298 correspondence audits involving sole proprietors closed in Fiscal Year 2007 identified that significant issues were not addressed in 129 (43 percent) of the audits.⁴ As a result, it was estimated that taxpayers may have avoided tax and interest assessments totaling approximately $83 million.⁵

This review was performed at the IRS National Headquarters in Washington, D.C., the SB/SE Division Headquarters in Lanham, Maryland, and the IRS Campus in Ogden, Utah, during the period September 2011 through April 2013. This performance audit was conducted in accordance with generally accepted government auditing standards. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives. Detailed information on our audit objectives, scope, and methodology is presented in Appendix I. Major contributors to the report are listed in Appendix II.

To ensure that correspondence audits are conducted in a quality manner, the IRS uses a comprehensive quality review system. The system includes a statistical sampling of correspondence audits. The IRS has also integrated various internal controls into its quality review system as outlined in the Standards for Internal Control in the Federal Government.⁶ These help the IRS determine the effectiveness and efficiency of its correspondence audits and whether examiners are ensuring that taxpayers are complying with laws and regulations.

Establishment of standards and performance measures

The IRS has established seven auditing quality standards. Each standard has key elements that elaborate on and further define the overall standard. For example, one of the key elements for Adequate Consideration of Significant Issues instructs examiners to consider and/or pursue audits of the prior and/or subsequent year returns when they contain the same issues as in the year examined. Appendix IV contains additional details on the key elements associated with each quality standard. Figure 1 explains each standard and provides an overview of the standard.

    Figure 1: Quality Auditing Standards for Correspondence Audits

The Internal Revenue Manual, quality standards, and examiner training courses emphasize the need for examiners to document their decisions and the actions taken during each audit in the audit case file. For example, the standard for considering penalties requires that examiners document their decisions to assess or not assess all applicable penalties, along with the basis for that decision. Such documentation is an important control activity because it helps ensure that examiners have considered all aspects of an audit and provides the information needed in subsequent reviews that evaluate whether examiners made the correct decisions.

Management reviews at the activity level

First-line managers review the documentation for a sample of audit case files to identify and correct quality problems in conjunction with evaluating the performance of the examiners they supervise. Managers rate examiner actions against established attributes based on the type of audit being reviewed. Attributes are collected into specific groups that follow the progression of a case and are mapped to the critical job elements for examiner performance evaluations. In addition to rating whether the examiner action met the attribute, managers may also enter narrative comments.

Management reviews at the functional level

While first-line manager reviews serve as the initial mechanism for identifying and correcting quality problems, each of the five IRS campus sites that conduct correspondence audits also perform quality review audits as part of the IRS-wide National Quality Review System (NQRS).⁷ Unlike first-line manager reviews that are generally used to evaluate a specific examiner's performance, NQRS reviews focus on providing IRS managers at all levels with statistically reliable estimates of audit quality by evaluating a small sample of closed audits.

Steps Can Be Taken to Strengthen the National Quality Review System

The NQRS data related to correspondence audits is not accurate. Although three of the four quality attributes we tested on a statistical sample of correspondence audits identified no significant concerns, tests of the accuracy-related penalty determinations identified a significant number of errors that had not been detected and reported by NQRS quality reviewers. The errors were caused, in large part, by inconsistencies between the NQRS attributes and the auditing standards.

Our audit tests included a review of statistical samples of 66 discretionary⁸ correspondence audits and 61 Earned Income Tax Credit correspondence audits (127 total audits) from a population of 2,913 audits reviewed by the NQRS between October 1, 2009, and March 31, 2011. Auditors' case reviews focused on four technical quality attributes that determined whether:

Except for a few instances, TIGTA did not identify any concerns with how NQRS reviewers rated the quality of the large, unusual, and questionable issues that examiners were assigned to audit. In addition, nearly all the closed audit case files reviewed contained the required tax account information from IRS automated systems that examiners could use to help address the large, unusual, and questionable issues they were assigned to audit. However, we did find a concern with penalty determinations.

Penalty determination errors were undetected

Sixty-five (51 percent) of the 127 total correspondence audits in our two statistical samples contained errors in accuracy-related penalty determinations that NQRS quality reviewers had not detected and reported. As such, the reported NQRS quality measure of 95 percent appears to be significantly overstated.

During audits, examiners are responsible for considering accuracy-related penalties when recommending adjustments to tax liabilities. These penalties, which typically include negligence and substantial understatement penalties in correspondence examinations, are designed to promote the preparation and submission of complete and correct information on tax returns as well as impose an economic cost on taxpayers who choose not to comply with the tax law.

There are inconsistencies between the auditing standards and the NQRS attributes and job aids used to evaluate correspondence audits

The auditing standards laid out in the IRS's Internal Revenue Manual and the attributes used for NQRS rating are not all aligned. Two inconsistencies identified are penalty determinations and multiyear examinations.

Auditing standards require that accuracy-related penalties, as well as all other applicable penalties, be considered in all audits and assessed when warranted. The standards also specify that penalties are to be accurately computed if they are recommended for assessment.

However, the NQRS job aids call the penalty quality measure "Penalty Computation" and instruct quality reviewers to rate the measure any time a penalty is calculated or determined by the examiner. Four of the seven quality reviewers we interviewed advised us that their interpretation of this NQRS quality measure meant limiting the review to checking if penalties were accurately computed, not whether a penalty should have been considered. Three quality reviewers' responses were in line with audit standards because they evaluated if applicable penalties were considered and computed accurately when recommended for assessment.

There is also inconsistency and confusion over when, or if, the scope of single-year audits should be expanded to include the prior and/or subsequent year returns. As of January 1, 2013, the auditing standards clearly indicate that the scope of single-year audits are to be expanded to the prior and/or subsequent year returns when they contain the same or similar large, unusual, or questionable issues as in the year examined. In contrast, the NQRS quality attribute that deals with audit scopes and large, unusual, and questionable issues does not discuss considering and pursuing these issues in prior/subsequent year returns. However, it does instruct quality reviewers to assess if all related tax years were properly considered in conjunction with addressing all large, unusual, and questionable issues. This slightly different wording led several quality reviewers to review cases differently than what is required in the auditing standards.

During interviews with NQRS quality reviewers in December 2012, TIGTA auditors received a variety of answers when asked how the reviewers rate the attribute "Are all related tax years considered?" One reviewer was unaware of the requirement or that any other quality attribute required examiners to evaluate related tax returns. Some reviewers stated that looking at related tax returns depends on the type of issues that were identified for audit. Other reviewers stated that prior and/or subsequent year tax returns are never considered in the context with single-year audits.

IRS officials stated that correspondence examiners, unlike examiners who conduct face-to-face audits, are not required to perform Required Filing Checks of other tax years during correspondence audits. Instead, the IRS relies on its audit sources and return selection process for correspondence audits to determine if a prior and/or subsequent year return should be audited. Required filing checks are used during face-to-face audits, in part, to determine whether the same pattern of noncompliance identified on the audited tax return is present on the prior and/or subsequent year tax returns and if those tax returns also warrant auditing. Therefore, the NQRS attribute that asks if related tax years were considered is not valid for correspondence audits.

If the NQRS attributes do not measure quality according to the auditing requirements, the NQRS results will be incorrect and could be misleading. The NQRS attributes need to be adjusted to align with the actual requirements for correspondence examinations.

A more complete picture of audit quality could be provided to NQRS customers

The ability of IRS executive managers and other NQRS customers to identify and correct quality concerns with correspondence audits could be facilitated if the NQRS reporting mechanisms provided not just information on overall audit quality but also specific information on the rating of the various quality attributes evaluated by NQRS reviewers during their case reviews. Considering the potential effect that poor audit quality can have on addressing underreporting noncompliance, such information could be very useful for isolating problem areas so that corrective actions can be taken. It would also help ensure that quality problems in key areas are not masked by reporting one overall quality score.

The various IRS functional areas summarize high-level performance information on a quarterly basis through the IRS's Business Performance Review reporting process. The process is designed to provide the opportunity for IRS executive managers to share accomplishments and compliance concerns with the IRS Commissioner, Deputy Commissioner, and IRS Oversight Board. However, because the Business Performance Review reports are intended to present high-level summary information, just one overall quality score is typically included in the quarterly review even though as many as 71 items are reviewed.⁹

Although not specifically required by the Business Performance Review reporting process, the Large Business and International Division quality reviewers also supplement the reports each quarter by providing executive managers with a detailed analysis of the attributes within each auditing standard for which improvements can be made. For example, the analysis for the first quarter of Fiscal Year 2013 showed that although the Large Business and International Division examiners achieved an overall quality rating of 90 percent, there were at least two attributes that were rated below 60 percent. When shared with examiners and their managers, such analyses can serve as a means to communicate and reinforce the importance of adhering to IRS audit quality guidelines and directives along with focusing attention on areas in the audit process that need improvement. This practice should also be considered for the SB/SE Division executives.

The random selection of audits for NQRS review cannot be verified

The NQRS is designed to provide statistically reliable estimates of audit quality nationwide by evaluating a relatively small sample of audits. The results, in turn, provide IRS management with data that can be used to identify training and educational needs and improve audit quality. To help ensure that the sample audits reviewed by the NQRS provide measures of audit quality that are representative of all correspondence audits and can be relied upon to identify areas that need improvement, IRS statisticians design and provide sampling plans for each campus. Every three months, the statisticians reassess the sampling plans and provide the quality reviewers in each campus with a memorandum that specifies the number of closed audits that are to be randomly selected at specific intervals for review.

Audit tests of the SB/SE Division campuses quality review data showed only minor differences between the number of reviews IRS statisticians designated for the SB/SE Division sampling plans and the number of reviews completed by quality reviewers at all SB/SE Division campuses. Figure 2 shows the results of our analysis comparing the number of NQRS quality reviews planned to the actual number of quality reviews completed from October 2009 through March 2011.

            Figure 2:  Comparison of the Number of Quality

Although the requisite number of audits called for in the sampling plans was reviewed, we were not able to verify or replicate the sampling process to determine if the audits were randomly selected for review at the specified intervals because such documentation was not maintained. Therefore, we were not able to confirm the statistical validity of the NQRS results.

According to published statistical sampling procedures from both within and outside of the Federal Government, random selection is critical to ensure that samples are representative of the population from which they were selected and to eliminate personal bias or subjective considerations from the selection process. Moreover, the Office of Management and Budget guidance on statistical surveys specifies that sample documentation should include information necessary to allow a third party to replicate and evaluate statistical sampling results.

Recommendations

The Commissioner, SB/SE Division, should ensure that:

Recommendation 1: The auditing standards are better aligned with the NQRS attributes and quality measures.

² TIGTA, Ref. No. 2010-30-059, Accuracy-Related Penalties Are Seldom Considered Properly During Correspondence Audits (Jun. 2010).

³ The projection was made using a confidence level of 95 percent, precision rate of ± 5 percent, and expected occurrence rate (error rate) of 79 percent.

⁴ TIGTA, Ref. No. 2010-30-024, Significant Tax Issues Are Often Not Addressed During Correspondence Audits of Sole Proprietors (Feb. 2010).

⁵ The projection was made using a confidence level of 95 percent, precision rate of ± 5.32 percent, and expected occurrence rate (error rate) of 43.29 percent.

⁶ Government Accountability Office (formerly known as the General Accounting Office), GAO/AIMD-00-21.3.1, Standards for Internal Control in the Federal Government (Nov. 1999).

⁷ Appendix IV provides an overview of the quality attributes considered by NQRS reviewers.

⁸ Discretionary audits focus on a variety of tax issues other than Earned Income Tax Credits and are identified from a number of sources, including the Discriminant Index Function (an automated system for scoring individual tax returns according to their audit potential), studies/research projects, third-party document matching, and Federal and State Government agency referrals.

⁹ Appendix V contains the correspondence audit portion of the IRS's campuses Business Performance Review report for the first quarter of Fiscal Year 2013.

To accomplish these objectives, we:

I. Evaluated the controls and procedures for the NQRS program when considering correspondence audits.

Internal controls methodology

Internal controls relate to management's plans, methods, and procedures used to meet their mission, goals, and objectives. Internal controls include the processes and procedures for planning, organizing, directing, and controlling program operations. They include the systems for measuring, reporting, and monitoring program performance. We determined the following internal controls were relevant to our audit objectives: IRS policies, procedures, and practices for determining whether examiners are meeting certain NQRS attributes during correspondence audits. We evaluated these controls by reviewing source materials, interviewing management, and reviewing a sample of 127 closed correspondence audits that had been previously reviewed by the NQRS.

Augusta R. Cook, Acting Assistant Inspector General for Audit (Compliance and Enforcement Operations)

Carl Aley, Acting Assistant Inspector General for Audit (Compliance and Enforcement Operations)

Frank Dunleavy, Director

Michelle Philpott, Acting Director

Robert Jenness, Audit Manager

Debra Mason, Lead Auditor

Donna Saranchak, Senior Auditor

William Tran, Senior Auditor

Ali Vaezazizi, Auditor

Office of the Commissioner -- Attn: Chief of Staff C

Office of the Deputy Commissioner for Services and Enforcement SE

Deputy Commissioner, SB/SE Division SE:S

Director, Campus Compliance Services, SB/SE Division SE:S:CCS

Director, Communications, Liaison, and Disclosure, SB/SE Division SE:S:CSO

Director, Examination, SB/SE Division SE:S:E

Director, Campus Reporting Compliance, SB/SE Division SE:S:CCS:CRC

Director, Exam Planning and Delivery, SB/SE Division SE:S:E:EPD

Director, Exam Policy, SB/SE Division SE:S:E:EP

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 Internal Control OS:CFO:CPIC:IC

Audit Liaison: Commissioner, SB/SE Division SE:S

Audit Information Management System -- A computer system used to control returns, input assessments/adjustments to the Integrated Data Retrieval System, and provide management reports.

Campus -- The data processing arm of the IRS. The campuses process paper and electronic submissions, correct errors, and forward data to the Computing Centers for analysis and posting to taxpayer accounts.

First-Line Manager -- A group manager in the Examination function responsible for supervision of IRS examiners.

Fiscal Year -- A 12-consecutive-month period ending on the last day of any month. The Federal Government's fiscal year begins on October 1 and ends on September 30.

Individual Master File -- The IRS database that maintains transactions or records of individual tax accounts.

Integrated Data Retrieval System -- The IRS computer system capable of retrieving or updating stored information; it works in conjunction with a taxpayer's account records.

National Quality Review System -- Allows national reviewers to evaluate closed audit files to determine whether examiners complied with quality attributes established by the IRS.

Skip Interval -- The number of elements in the population divided by the number of sampling units in the sample.

Tax Gap -- The difference between taxes that are legally owed and taxes that are paid on time.

The auditing standards outline the criteria determined to produce a quality examination and are guidelines to assist examiners in the completion of their cases. They are used in conjunction with applicable program procedures for evaluating case quality. In the campus environment, due in part to the type of examination conducted and in part to the nature of our processes, some auditing standards will be less applicable. Our program procedures and quality measures define how they align with our work. Campus examinations are generally narrowly focused. They are limited to a particular schedule or issue, or focus on stopping erroneous refunds. We will determine if discrepancies exist between the auditing standards and the quality attribute coding and adjust the process or guidance as needed.

Our campus organizations manage a wide variety of programs. Because of the number and variety, we limit our highest level reporting, the Business Performance Review (BPR), to an overall quality score. However, we measure and evaluate each of these programs on a number of attributes.

We regularly share more specific information about the program performance with executives and all levels of management who use it to monitor, evaluate and take action to improve the programs. We will determine if providing additional quality data with more frequency will be helpful.

We do not agree with your third recommendation. Audit cases selected for NQRS are determined based upon a statistically valid methodology designed by Statistics of Income (SOI) as described in the sample plan. We ensure the randomness of a sample by selecting the "Nth" case using a skip interval based on the number of required reviews and the population of work. Upon closure of paper cases they are placed in a designated area ready for potential review. The clerk pulls the cases by applying the skip interval as determined in the sample plan. Once the sample has been selected the unused cases are removed and the area is stocked with the next day's cases. While the process is valid, it is not reproducible because the cases are removed daily. We have reviewed our current system to determine alternatives for creating a reproducible process. Given the volume of open cases is over 300,000, the resources necessary to implement a reproducible manual process would not be cost effective. In the alternative, a systemic report could be created, but this would take funding and years to implement. Based upon labor and cost restrictions, we are unable to implement a systemic method as per your recommendation.

Attached is a detailed response outlining our corrective actions.

If you have any questions, please contact me, or a member of your staff may contact Denice Vaughan, Director, Campus Compliance Services, at (404) 338-9116.

Attachment

Ensure that the auditing standards are better aligned with the NQRS attributes and quality measures.

CORRECTIVE ACTION:

We will determine if discrepancies exist between the auditing standards and the quality attribute coding and adjust the process as needed.

IMPLEMENTATION DATE:

October 1, 2014

RESPONSIBLE OFFICIAL(S):

Director, Campus Reporting Compliance, Small Business/Self-Employed

CORRECTIVE ACTION MONITORING PLAN:

IRS will monitor this corrective action as part of its internal management system of controls.

RECOMMENDATION 2:

Ensure that a more complete picture of correspondence audit quality is provided to NQRS customers on a regular basis.

CORRECTIVE ACTION:

We will determine if providing our executives and program managers additional quality data with greater frequency would be helpful and feasible, and make any appropriate changes.

IMPLEMENTATION DATE:

October 1, 2013

RESPONSIBLE OFFICIAL(S):

Director, Campus Reporting Compliance, Small Business/Self-Employed

CORRECTIVE ACTION MONITORING PLAN:

The IRS will monitor this corrective action as part of its internal management system of controls.

RECOMMENDATION 3:

Ensure that the audits selected for NQRS review are randomly selected and documentation is maintained that will allow IRS management and third parties to verify that the sampling plan was properly implemented.

CORRECTIVE ACTION TAKEN:

We do not agree with this recommendation. We have explored the feasibility of using a reproducible listing process to select cases for review. We do not have the capability to pull a daily automated listing using our current systems. This automation change would take funding and years to implement. The manual gathering of this information would be extremely labor intensive and subject to error; therefore, it would not be feasible to implement your recommendation at this time.

IMPLEMENTATION DATE:

N/A

RESPONSIBLE OFFICIAL:

N/A

CORRECTIVE ACTION MONITORING PLAN:

N/A