National Taxpayer Advocate Releases 2005 Report to Congress

The 21 problems described in the report are as follows:

1. Trends in Taxpayer Service. As the IRS proposes to allocate more resources to collection, examination, and criminal investigation functions and fewer resources to taxpayer service functions, the IRS is also increasing efforts to "migrate" taxpayers toward electronic services and away from face-to-face contact. Before altering the mix of service and enforcement, the National Taxpayer Advocate believes the IRS should spend more time studying what types of services different taxpayer segments need and how best to deliver these services to help taxpayers remain compliant. The National Taxpayer Advocate recommends that the IRS undertake a research-driven needs-assessment, from the taxpayers' perspective, to help identify what services taxpayers need and want and how best to deliver them. An assessment of needs will also help identify groups of taxpayers that may be resistant to, or unable to access, certain services. Once the IRS conducts a detailed assessment and understands how any proposed changes to taxpayer service may affect compliance, the IRS should develop a detailed strategy for migrating taxpayers from the current to the proposed model of delivering taxpayer service.

2. Criminal Investigation Refund Freezes. The IRS Criminal Investigation function (CI), through its Questionable Refund Program (QRP), places a "freeze" on hundreds of thousands of refund claims each year that it believes may contain indicia of fraud. CI personnel currently review the refund claims and "determine" whether they are fraudulent -- without notifying taxpayers that their claims are under review and without giving taxpayers an opportunity to present documentation supporting their positions. Last year, the Taxpayer Advocate Service (TAS) received more than 28,000 requests for assistance from taxpayers whose refunds had been frozen. TAS studied a randomly selected sample of nearly 500 cases to determine the ultimate disposition of these cases. When TAS assisted the taxpayers, CI ultimately agreed to issue the full amount of the refund claimed (or more) in 66 percent of the decided cases and to issue a partial refund in an additional 14 percent of the decided cases. Thus, taxpayers received a full or partial refund in 80 percent of frozen-refund cases brought to TAS. The median Adjusted Gross Income (AGI) of these taxpayers was $13,330, and the median refund was $3,519. Thus, the refund constituted, on average, more than 26 percent of the claimant's AGI for the year, and the taxpayers were required to wait, on average, more than 8-1/2 months to receive their refunds. The National Taxpayer Advocate believes that the QRP is an important program to protect against tax fraud, but the IRS must implement procedures to notify taxpayers that their refunds have been frozen, provide taxpayers with an opportunity to submit documentation, and bring cases to a quicker resolution.

3. The Cash Economy. Underreported income (and related self-employment tax) from the so-called "cash economy" is probably the single largest component of the "tax gap." It may exceed $100 billion per year. Because income from the cash economy is not subject to information reporting, many of the IRS's traditional means of enforcement are unlikely to be effective in addressing it. The IRS has a number of initiatives that could be effective if coordinated and pursued more aggressively. However, no single function coordinates research, outreach, and compliance initiatives aimed at improving reporting compliance among cash economy participants. Nor does the IRS give these initiatives the same level of attention as other initiatives, such as those addressing tax shelters or the Earned Income Tax Credit (EITC). The IRS must develop a comprehensive strategy for addressing the cash economy if it is to significantly reduce the tax gap.

4. Training of Private Debt Collection Employees. In or around July of 2006, the IRS's Private Debt Collection (PDC) initiative will go into operation. The PDC initiative marks a dramatic departure from traditional federal tax collection practice in that private collection firms will now be seeking payment from taxpayers with delinquent accounts, arranging installment agreements, and obtaining taxpayers' financial information over the phone. The National Taxpayer Advocate is concerned about taxpayers interacting with private collection agents who have no understanding of important tax laws and will be trained by the contractors instead of the IRS. With the abundance of tax experts employed at the IRS, it only makes sense to use the IRS's expertise to directly train private collectors at a level that approximates the training received by new IRS collection employees.

5. EITC Exam Issues. In the past few years, the IRS has made significant progress in improving the administration of the EITC. Despite this progress, problems still exist for taxpayers subject to EITC examinations and recertification. These problems include delays in exam and recertification procedures, low taxpayer response rates, a lack of standardized treatment in recertification cases, and problems with documentation. The National Taxpayer Advocate recommends that the IRS continue to identify and adopt measures designed to ease the burden on taxpayers and improve the current EITC examination and recertification processes.

6. Levies on Social Security Payments. In general, recipients of Social Security benefits are elderly or disabled workers, or the surviving dependents of deceased workers. The IRS continues to process levies on Social Security payments without sufficient managerial review, causing undue burden on a vulnerable population of taxpayers. The National Taxpayer Advocate urges the IRS to implement safeguards that would prevent levies from being imposed on Social Security payments to low income and other at-risk taxpayers.

7. Appeals Campus Centralization. The IRS Office of Appeals ("Appeals") has centralized its work on certain types of cases at six IRS campuses; these cases previously were resolved in field offices closer to taxpayers. Appeals campus centralization is aimed at decreasing the time it takes for a taxpayer to resolve a case. While this aim is appropriate and admirable, the National Taxpayer Advocate is concerned that centralizing Appeals case resolution may actually increase taxpayer burden. Centralization may reduce opportunities for taxpayers to have their cases resolved at the local level, diminish working relationships between taxpayers and Appeals employees, increase emphasis on processing at the expense of independent judgment, and diminish service for low income and unrepresented taxpayers. The National Taxpayer Advocate urges Appeals to alleviate these problems and protect taxpayers' appeal rights by developing specific training that will help employees carry out Appeals' independent mission in the campus environment. We recommend that Appeals monitor campus activity to ensure that taxpayers -- particularly low income taxpayers -- receive full and fair consideration of their cases and are adequately notified of their right to request a case transfer to a local Appeals office.

8. Refund Anticipation Loans: Oversight of the Industry, Cross-Collection Techniques, and Payment Alternatives. Given that a significant percentage of RAL customers are EITC recipients, the IRS has a compelling reason to consider improved oversight of the industry as well as seriously consider alternative refund delivery methods. The IRS contributes to the demand for Refund Anticipation Loans (RALs) by: (1) failing to deliver refunds in the quickest manner possible and (2) failing to provide RAL alternatives for the "unbanked." It is also unclear if RAL customers fully understand the ramifications of cross-collection provisions in standardized RAL contracts, or if customers would even purchase the products if they were adequately informed of the cross-collection practices. The National Taxpayer Advocate recommends that the IRS review the effect of its Revenue Protection Strategy (RPS) on refund turnaround times and work with the Department of Treasury to develop alternative means of delivering refunds to taxpayers.

9. Identity Theft. Identity theft occurs when someone uses another individual's personal information without permission to commit fraud or other crimes. Although the most common type of identity theft involves consumer fraud, an increasing number of identity theft victims find they need to contact the IRS to untangle their tax accounts. The National Taxpayer Advocate urges the IRS to resolve the tax issues faced by victims of identity theft more quickly and efficiently, which may require coordination between the IRS and other federal agencies.

10. Complexity of the Employment Tax Deposit System. The Internal Revenue Code places significant responsibilities on employers for depositing, reporting, filing, and paying employment taxes. Recent data shows that the IRS assesses failure to deposit (FTD) penalties on one out of every 16 employment tax returns, yet eventually abates more than 60 percent of the FTD penalty amounts it originally assessed. This suggests that the rules and regulations governing federal employment tax deposits are overly complex, presenting significant compliance problems for employers and administrative challenges for the IRS.

11. Automated Collection System Levy Releases. Collection efforts through the IRS's Automated Collection System (ACS) can result in levies of bank accounts, wages, or other income such as Social Security. In response to these levies, taxpayers will contact the IRS seeking to enter into a collection alternative, such as an installment agreement. The IRS can also designate an account as "currently not collectible" if the taxpayer can demonstrate a financial hardship. The Internal Revenue Code and Treasury Regulations require that the IRS promptly release levies when taxpayers enter into installment agreements or when they demonstrate the existence of a hardship. Some taxpayers encounter problems with the levy release, including clerical errors, delays that result in additional levies on taxpayer assets or income, or the IRS not returning levy proceeds when a delayed levy release results in additional levies. Delays are also due in part to taxpayers' failure to request expedited levy releases so that the release can be faxed to the third party levy source. The IRS has improved the levy release process for currently-not-collectible accounts and has agreed to consider similar system changes for installment agreements, and require additional training for employees. We encourage the IRS to consider revising its procedures so that all levy releases are assumed to require expedited procedures.

12. Regulation of Electronic Return Originators. Electronic Return Originators (EROs), along with return preparers, are taxpayers' entry point into the tax system and have unprecedented access to taxpayers' financial data and social security numbers. As such, EROs should be closely monitored to protect taxpayers, but the IRS's current regulation of EROs is minimal. The National Taxpayer Advocate recommends the IRS increase its efforts, in part by making more visits to EROs, tracking EROs that are not filing any tax returns, and requiring those non-active EROs to recertify. The National Taxpayer Advocate further recommends that the IRS consolidate end-to-end responsibility for EROs, from approval of applications to monitoring ERO activities, in one IRS organization and develop a long-term strategic plan for the e-file program and EROs to ensure adequate and effective oversight.

13. Limited Scope of Backup Withholding Program. Underreporting of individual income tax is the single largest source of the tax gap, accounting for over half of the gross tax gap. Backup withholding is one of the tools available to the IRS as it attempts to narrow the gap. The IRS has the authority to require payors to deduct and withhold tax under certain conditions, such as when a recipient fails to furnish a valid Taxpayer Identification Number. However, the IRS has not implemented the backup withholding program effectively and has failed to provide noncompliant taxpayers with sufficient incentive to fulfill their reporting obligations.

14. Accessibility of e-Services for Tax Practitioners. The IRS's web-based e-Services suite is available to tax practitioners who are active participants in the IRS e-file program and electronically file five or more accepted individual income tax returns in a filing season. The IRS uses e-Services "as a reward and incentive for e-filing" rather than addressing it from a customer service-oriented perspective. This policy of limiting access to e- Services deprives many highly trained and experienced tax practitioners of an extremely useful and efficient tool, and has the unintended result of making the Electronic Account Resolution program available to preparers without regard to their professional qualifications. The National Taxpayer Advocate recommends that the IRS provide access to e-Services to all practitioners qualified to practice before the IRS pursuant to Circular 230. She also recommends that the IRS expand access to the Transcript Delivery System in e-Services to all taxpayers, while taking the necessary security measures to safeguard confidential tax data on the Internet.

15. Mandatory Briefings for IRS Employees on TAS. Internal Revenue Code § 7803(c)(2)(C)(ii) requires the National Taxpayer Advocate to develop guidance for all IRS officers and employees, outlining the criteria for referral of cases to the Taxpayer Advocate Service (TAS). The IRS has denied the request of the National Taxpayer Advocate to include TAS training among its annual mandatory briefings for all employees. Instead, the IRS has agreed to provide TAS training to contact employees in the Wage and Investment and Small Business/Self-Employed operating divisions on a one-time basis, and urges TAS to train the rest of the employees through methods such as the IRS intranet, inserts in employee Earnings & Leave statements, and "wallet cards." The National Taxpayer Advocate urges the IRS to rethink its position and grant her request to make TAS training mandatory for all employees.

16. Allowable Expense Standards for Collection Decisions. Each year the IRS publishes schedules of national and local expense allowance standards. These standards reduce the subjectivity involved when IRS employees consider the collection alternative to pursue when a taxpayer is having difficulty paying the IRS (e.g., an offer-in-compromise, installment agreement, or suspension of collection). However, for any given taxpayer, the expense amounts provided by the standards will not necessarily cover his or her reasonable basic living expenses. The IRS relies on the subjective judgment of its employees to allow more than the standard amounts when appropriate. Many practitioners report that the IRS often fails to allow such additional amounts and uses the standards as an excuse to reject reasonable collection alternatives. Our report discusses a number of reasons that the standards are often inadequate and provides recommendations for improving them, encourages IRS employees to allow additional amounts when appropriate, and addresses practitioner concerns.

17. Inadequate Taxpayer Service to Exempt Organizations, Resulting in Unnecessary Penalties. Most tax exempt organizations are very small entities with meager resources and modest budgets that rely largely on the services of volunteers. Tax filing requirements for these organizations are complicated and time consuming. The IRS automatically assesses penalties on these organizations when they do not comply precisely with these complex requirements. Over the last 13 fiscal years, however, the IRS has abated almost 75 percent of the dollars assessed for filing-error penalties on exempt organizations because the organizations later corrected their mistakes. This high abatement rate is an unnecessary waste of IRS resources and indicates that exempt organizations would avoid filing errors if better informed of their filing obligations. Currently, however, the IRS substantially underfunds customer service to exempt organizations; for example, only 60 percent of organizations that call the IRS Tax Exempt and Government Entities Division toll-free help line receive service. The National Taxpayer Advocate urges the IRS to reconsider its funding decisions with respect to exempt organization customer service and better utilize the resources available.

18. Direct Deposit of Income Tax Refunds. Under present law, there are no procedures for the IRS, the government's Financial Management Service, and financial institutions to address inadvertent errors by taxpayers relating to direct deposits of tax refund checks. The taxpayer and the financial institution must resolve any dispute over the accuracy of a direct deposit refund, with little assistance from the IRS. The National Taxpayer Advocate recommends that the IRS consider all of the recommendations put forth by the Direct Deposit Task Force as well as other procedural improvements that can eliminate the potential for a misdirected direct deposit refund. The National Taxpayer Advocate also encourages the IRS to continue to work with financial institutions in an effort to recover misdirected funds.

19. Innocent Spouse Claims. One spouse (called the "innocent spouse") may apply to the IRS for relief from joint liability for deficiencies or underpayments attributable to the other spouse. The IRS has difficulty communicating with taxpayers regarding the requirements for relief and how to fill out related forms. As a result, the IRS is able to grant fewer than three in ten requests for relief, and despite recent cuts in processing time, the IRS collectively takes more than two years to process requests that are appealed. Various IRS computer systems contain inconsistent information about innocent spouse claims processed by the Appeals function. These inconsistencies make it more difficult for the IRS to identify the source of any processing delays.

20. Limitations of Collection Account Databases. The lack of access to full taxpayer account histories in one place makes it difficult for IRS contact employees to respond to taxpayers' questions or provide proper guidance on potential case resolutions. This inaccessibility often leads taxpayers to seek the assistance of the Taxpayer Advocate Service. The Desktop Integration (DI) system provides IRS employees with greater access to the information needed, but not all systems interface with DI and not all IRS employees are required to use DI. The National Taxpayer Advocate urges the IRS to expand the number of systems interfacing with DI as well as requiring all IRS contact employees to use DI.

21. Reasonable Cause Assistant (RCA). The IRS does not utilize the Reasonable Cause Assistant (RCA) program -- a computer-based decision support tool -- for all determinations of penalty abatements for reasonable cause. In addition, the high rate of RCA abatement conclusions indicates that the IRS needs to eliminate unnecessary penalty assessments. Because existing IRS data does not sufficiently differentiate RCA cases from non-RCA cases, the IRS cannot effectively analyze the penalty systems in place. The low rate at which RCA users abort the program's conclusions may indicate that RCA does not encourage users to override an RCA decision even when appropriate. The National Taxpayer Advocate urges the IRS to review RCA and all related training and guidance materials to determine whether the application actually gives the user the flexibility to fully consider unique facts and circumstances.

KEY LEGISLATIVE RECOMMENDATIONS

Tax Reform Core Principles: A Taxpayer-Centric Approach. Two months ago, the President's Advisory Panel on Federal Tax Reform submitted a report to the Secretary of the Treasury proposing significant revisions to the Internal Revenue Code, and it appears that Congress may give serious consideration to fundamental tax reform in the next year or two. We recommend that Congress give priority emphasis to six core principles as it considers various tax-reform proposals: (1) the tax system should not "entrap" taxpayers; (2) the tax laws should be simple enough so that taxpayers can prepare their own returns without professional help, simple enough so that taxpayers can compute their tax liabilities on a single form, and simple enough so that IRS telephone assistors can fully and accurately answer taxpayers' questions; (3) the tax laws should anticipate the largest areas of noncompliance and minimize the opportunities for such noncompliance; (4) the tax laws should provide some choices, but not too many choices; (5) the tax laws should not necessarily avoid refundable credits but, if it includes them, should design them in a way that is administrable; and (6) the tax system should incorporate a periodic review of the tax code -- in short, a sanity check.

Measures to Reduce Noncompliance in the Cash Economy. The IRS estimates that the annual federal tax gap for 2001 was between $257 billion and $298 billion. The IRS receives about 130 million income tax returns each year. Thus, every taxpayer is forced to pay an average $2,000 "surtax" each year to subsidize noncompliance. IRS data show that the highest rate of noncompliance by far is attributable to transactions that are not reported to the IRS on a Form W-2, Form 1099, Schedule K-1, or similar form. These unreported transactions occur largely in the so-called "cash economy." To reduce the tax burden on compliant taxpayers, we recommend that Congress (1) create a three-pronged reporting and payment system that encourages compliance in certain cash economy transactions by (a) instituting backup withholding on payments to taxpayers who have demonstrated "substantial noncompliance"; (b) releasing backup withholding on payments to "substantially noncompliant" taxpayers who have demonstrated "substantial compliance" and agree to schedule and make future estimated tax payments through the IRS Electronic Funds Transfer Payment System (EFTPS); and (c) providing that payors will not be required to institute backup withholding on payments to independent contractors that present payors with a valid IRS "compliance certificate"; (2) require the IRS to promote the making of estimated tax payments through EFTPS; (3) authorize voluntary withholding agreements between independent contractors and service recipients; and (4) require third-party information reporting for certain payments to corporations with 50 or fewer shareholders.

Tax Reform for Families: A Common Sense Approach. The Internal Revenue Code contains six provisions related to a taxpayer's family status: the Earned Income Tax Credit (EITC), the Child Tax Credit (CTC), the Child and Dependent Care Credit, personal and dependency exemptions, the head-of-household filing status, and the "separated spouse" rules of IRC § 7703(b). Each of these six provisions directly or indirectly confers a tax benefit on taxpayers who meet the various eligibility requirements, and at least one of these six provisions impacts every U.S. individual taxpayer. To build upon the recently enacted Uniform Definition of Child and to further simplify the family status provisions, we recommend that Congress (1) combine the exemptions, CTC, and part of the EITC and head of household filing status into a refundable Family Credit comprising two components -- one for the taxpayer (and his or her spouse) and one for whomever is the "main carer" of a child or children based on a per-child amount; (2) separate the Child and Dependent Care Credits into two credits; (3) eliminate head-of-household filing status; (4) modify the EITC so that it provides a refundable credit to low income workers based solely on the taxpayer's earned income and is available to workers age 18 and over, regardless of the existence of children in the household; (5) permit married taxpayers who have a legal and binding separation agreement and who live separate and apart as of the last day of the calendar year to be considered "not married" for purposes of filing status; and (6) provide a separate credit for noncustodial parents of Qualifying Children who pay all child support obligations due for that calendar year.

Another Marriage Penalty: Taxing the Wrong Spouse. The federal income tax liabilities of married persons are often imposed on or collected from a spouse who did not earn the income subject to tax, i.e., the "wrong" spouse. Current law provides some relief to a spouse held liable for tax on the other spouse's income, at least in cases where the first spouse did not know about the income and did not significantly benefit from it. However, the relief rules are sometimes overly narrow, complex, costly for the IRS to administer, and burdensome for taxpayers. Even if relief rules apply so that one spouse is not liable for his or her spouse's tax, the IRS may be able to collect the liability from the non-liable spouse in community property states. We recommend that Congress amend the law to tax the "right" spouse in the first instance and to prevent the IRS from undermining this rule through its collection efforts. Our recommendation would better align each person's tax with his or her individual ability to pay, significantly reduce complexity, and minimize the impact of state property and collection laws that subject taxpayers to different amounts of federal income tax solely because they reside in different states.

Requiring Brokers to Track and Report Cost Basis for Stocks and Mutual Funds. Many financial institutions through which investors own stocks and mutual funds ("brokers") do not currently keep track of an investor's basis in the stocks or mutual funds, and no brokers report basis information to both taxpayers and the IRS on a Form 1099-B. The absence of information reporting creates serious problems for many taxpayers and the government alike. For taxpayers, tracking basis can be extraordinarily complex and many taxpayers seeking to comply with the law find that they simply cannot do so with accuracy, leaving them exposed if audited. From the government's perspective, the absence of information reporting enables underreporting by taxpayers who deliberately overstate their basis (thereby reducing their gain or even generating a loss), because they know the IRS generally cannot detect errors in basis reporting in the absence of an audit. One recent estimate puts the revenue loss to the government from such underreporting at $250 billion over the next 10 years. We recommend that brokers be required to keep track of an investor's basis, transfer basis information to a successor broker if the investor transfers the stock or mutual fund holding, and report basis information to the taxpayer and the IRS (along with the proceeds generated by a sale) on Form 1099-B. To offset the cost of implementing a tracking system, we note that Congress could provide a one-time tax credit for brokers.

Tracking Cost Basis as a Result of Estate Tax Repeal. Under the current estate tax regime, persons acquiring property from a decedent are able to use a "stepped-up" basis equal to the fair market value of the property at the date of the decedent's death (or, if they so elect, on the date six months after the decedent's death). Once the estate tax is repealed in 2010, these taxpayers must use the modified carryover basis, which may require extremely complex calculations to determine the property's adjusted basis in the hands of the decedent just prior to death. Reconstructing adjusted basis in property is difficult enough while taxpayers holding such assets are alive; after death, it can become impossible. Congress should explore ways to lessen this compliance burden.

Restructuring and Reform of Collection Due Process Provisions. Collection Due Process (CDP) hearings afford taxpayers the opportunity to obtain meaningful review of IRS collection actions by an impartial IRS Appeals Officer and the courts, either after the initial filing of a Notice of Federal Tax Lien or before an initial levy on a taxpayer's assets. The current statutory CDP rights are both under-inclusive and over-inclusive, denying judicial review of some lien and levy actions while encouraging counterproductive behavior on the part of some taxpayers and the IRS. To enhance taxpayer protections in the tax collection process while ensuring that the IRS's ability to collect the correct amount of tax is not unreasonably impaired, we recommend that Congress (1) require the IRS to issue a separate CDP Right to Hearing notice at the time it undertakes the first levy action with respect to a tax, describing with specificity the levy source and date such levy will occur and providing the taxpayer with the name and contact information of an IRS employee to call about the levy action; (2) consolidate judicial review of CDP hearings in the United States Tax Court, clarify the role and scope of Tax Court oversight of Appeals' continuing jurisdiction over CDP cases, and address the Tax Court's standard of review for the underlying liability in CDP cases; and (3) codify both the IRS Collection Appeals Program and the IRS Audit Reconsideration Process and specifically include Audit Reconsideration as an alternative to be considered at CDP hearings.

ADDITIONAL LEGISLATIVE RECOMMENDATIONS

Direct Deposit of Income Tax Refunds. Under present law, there are no procedures in place for the IRS, the government's Financial Management Service, and financial institutions to address inadvertent errors by taxpayers relating to direct deposits of tax refund checks. Disputes over the accuracy of a direct deposit refund due to taxpayer or preparer error must currently be resolved between the taxpayer and the financial institution itself, with little assistance from the IRS. The National Taxpayer Advocate recommends that Congress amend the Internal Revenue Code to create a process through which the IRS and financial institutions work together to identify the incorrect recipient of a direct deposit refund and require the return of the improperly deposited funds.

Social Security Levies. Current law exempts from IRS levy certain pension and annuity payments (including payments under the Railroad Retirement Act), but it does not exempt from levy retirement, survivors, and disability insurance payments made under the Social Security Act. Levies by the IRS on Social Security benefits can cause particularly severe hardships for low income taxpayers who rely on these payments as their primary or sole source of income. The National Taxpayer Advocate recommends that Congress exempt Social Security payments altogether from IRS levy. In the alternative, the National Taxpayer Advocate recommends that Congress extend the exemption amount applicable to manual levies to automated levies under the Federal Payment Levy Program.

Debt Collection Techniques on EITC Benefits by the Refund Anticipation Loan Industry. Refund anticipation loan (RAL) customers may not completely understand the ramifications of the debt offset collection provisions included in standardized RAL contracts. The provisions give the contracting financial institution or bank the authority to offset RAL proceeds to satisfy outstanding delinquencies owed on RALs previously issued by either the contracting bank or a third-party bank. The practice allows banks to effectively seize EITC benefits and transfer the funds to themselves or third-party banks to satisfy these prior delinquencies. The National Taxpayer Advocate recommends that Congress amend IRC § 32 to prohibit banks from exercising their right to set off on EITC benefits, a protection that currently exists for Social Security benefits. At the very least, the law should prohibit banks from transferring any portion of a federal tax refund representing the EITC to a third-party bank.

Appeals from Collection Due Process Hearings under IRC §§ 6320 and 6330. Collection Due Process (CDP) hearings provide taxpayers an opportunity for independent review of the first lien filed by the IRS or the first proposed levy action with respect to a tax liability. As in 2003 and 2004, CDP was the most frequently litigated tax issue in the federal courts. We reviewed 209 decisions during the current period. Our analysis shows that collection alternatives were litigated more than any other such issue in these cases (40 out of 209 cases). Taxpayers prevailed in five cases. Taxpayers also attempted to litigate the underlying liability in 39 cases; however, in 26 cases the courts ruled that taxpayers were unable to litigate that issue because they had another opportunity to do so earlier in the process.

Gross Income under IRC § 61 and Related Sections. The issue of what constitutes gross income for purposes of Internal Revenue Code (IRC) § 61 has been among the Most Litigated Issues since 1998, the first year that the National Taxpayer Advocate was required to report on litigated issues. Gross income is again the second most litigated issue this year. The cases generally fell into one of four categories: (1) awards or settlements, (2) disability and Social Security income, (3) constructive dividends, and (4) unreported income. Taxpayers prevailed, in whole or in part, in 25 out of 108 cases. While no clear patterns are evident, taxpayers appeared to have the most difficulty in cases where other sections of the Internal Revenue Code exclude income items that would otherwise be taxable.

Failure to File Penalty under IRC § 6651(a)(1). The number of cases involving the failure to file penalty under IRC § 6651(a)(1) increased by 60 percent from last year to make this the third most litigated issue. We reviewed 75 cases involving the failure to file penalty in federal courts. Taxpayers seek relief from the penalty by asserting "reasonable cause" for the failure to file a timely return. Taxpayers argued that varying excuses constitute reasonable cause, including illness of the taxpayer or family member, reliance on a tax professional, reliance on a spouse, or ignorance of the law. Taxpayers prevailed in five of 75 cases (7 percent).

Trade or Business Expenses under IRC § 162 and Related Code Sections. The deductibility of trade or business expenses is perennially one of the ten most litigated tax issues in the federal courts. We reviewed 67 cases that included trade or business expense issues. The courts affirmed the IRS position in nearly 75 percent of the cases, while taxpayers prevailed just seven percent of the time; the remaining cases resulted in split decisions. Substantiation of trade or business expenses was the primary sub-issue litigated by taxpayers, who often failed to provide sufficient documentation of expenses they incurred, causing them to lose otherwise permissible deductions. The IRS can assist these taxpayers, and minimize litigation, by continuing to provide clear guidance on the deductibility of trade or business expenses.

Frivolous Issues Penalty under IRC § 6673. We reviewed 67 cases in which taxpayers litigated the IRC § 6673 penalty for advancing frivolous arguments or conducting litigation solely for delay. This represents a 90 percent increase over last year's 35 cases. In these cases, the Tax Court considers whether taxpayers' arguments or actions warrant application of the penalty. IRS counsel can move for imposition of the penalty or the court can do so on its own initiative. Courts consistently held that penalties were warranted where the same type of "boiler-plate" arguments had been held by other courts to be frivolous, having no basis in law or fact. The increase in these cases suggests that the IRS is more willing to seek imposition of the penalty and courts are more willing to sustain it.

Negligence Penalty under IRC § 6662(b)(1). We reviewed 57 cases involving the accuracy-related penalty under IRC § 6662(b)(1). This penalty is assessed against taxpayers for underpayment of tax due to the taxpayer's negligence or disregard of tax rules or regulations and was generally decided in conjunction with other issues. The IRS prevailed in 39 cases (or 68 percent), while taxpayers prevailed in 13 cases (23 percent), and five cases ended in split decisions. A common reason for rulings against taxpayers was the lack of any evidence that there was reasonable cause and the taxpayers acted in good faith. However, the taxpayer victories raise the question of whether the IRS always exercises proper oversight on the imposition of the negligence penalty. The IRS abates a large percentage of other types of penalties. This fact and the relatively significant number of taxpayers who prevailed on reasonable cause indicate that the IRS should study whether the accuracy-related penalty is achieving its original goal of promoting voluntary compliance by encouraging accurate tax returns.

Family Status Issues under IRC § 2, 21, 24, 32, and 151. The Earned Income Tax Credit (EITC), dependency exemption, head of household filing status, child tax credit, and child and dependent care credit frequently arise in the same cases and involve similar factual determinations. We reviewed 45 cases concerning these issues, two-thirds of them dealing with multiple credits and issues where the determination of one issue often affected others. Our analysis shows that taxpayers who wish to claim the family status credits and deductions often do not understand the qualification requirements or how to properly satisfy them. Taxpayers were represented by counsel in only six of the 45 cases, even though many were highly fact-specific and involved a complicated web of statutory provisions.

Relief from Joint and Several Liability under IRC § 6015. Spouses filing joint federal income tax returns are jointly and severally liable for any deficiency or tax due. IRC § 6015 provides three avenues for relief from joint and several liability. We reviewed 45 federal court opinions involving relief under IRC § 6015, with 33 cases decided in favor of the IRS, 11 in favor of the taxpayer, and one split decision. While the courts considered many factors in determining the appropriateness of relief, the most significant was whether the requesting taxpayer had actual or constructive knowledge of the tax deficiency. Taxpayers' success or failure often turns on their willingness and ability to provide documentation to support their claims.

Summons Enforcement under IRC § 7604. The IRS has the authority to summon the production of books, records, or testimony from witnesses when investigating either a civil or criminal tax liability. The court ruled in favor of the IRS in 42 of the 44 cases reviewed, while a taxpayer prevailed in one case and another ended in a split decision. In at least nine of the cases, the summons in question was issued to third parties who the IRS believed were marketing unlawful tax shelters to taxpayers, reflecting the IRS's increased attention to shelter activity. As the IRS becomes more aggressive in its enforcement initiatives, it will likely increase its use of the summons enforcement tool, and the courts will see a growing number of these cases.

Trust Fund Recovery Penalty under IRC § 6672. The Trust Fund Recovery Penalty is a means by which the government holds certain persons responsible for willfully failing to withhold or remit the trust fund portion of payroll taxes. We reviewed 34 opinions in which the TFRP was an issue. Taxpayers prevailed in whole or in part in 13 of the 34 cases, though in three cases the court was denying the IRS's motion for summary judgment, thereby requiring the parties to go to trial. The cases often involved officers of small businesses who had some role in determining expenditures. As one court noted, these cases often reflect the difficult choices for corporate officers who are faced with the dilemma in a struggling business between losing one's job or violating the nation's tax laws. Still, these choices do not excuse the responsibility of paying payroll taxes.

I respectfully submit for your review the National Taxpayer Advocate's 2005 Annual Report to Congress. This is the fifth such report I have submitted as the National Taxpayer Advocate. More than anything else, this year's report is a call for tax reform. Our tax code has grown so complex it creates opportunities for taxpayers to make inadvertent mistakes as well as to game the system. As the Code gets more complex -- especially for individual taxpayers and small businesses -- it becomes more difficult for the IRS to provide comprehensive, quality taxpayer service.

As taxpayers become confused and make mistakes, or deliberately "push the envelope", the IRS understandably responds with increased enforcement actions. The exploitation of "loopholes" leads to calls for new legislation to crack down on abuses, which in turn makes the tax law more complex. Thus begins an endless cycle -- complexity drives inadvertent error and fraud, which drive increased enforcement or new legislation, which drives additional complexity. In short, complexity begets more complexity. This cycle can only be broken by true tax simplification, followed by ongoing legislative and administrative discipline to avoid subsequent "complexity creep".

Our own contribution to the tax reform debate, though modest, is taxpayer-centric. Among the legislative recommendations included in this report, we first set out some general principles for tax reform, from the perspective of taxpayers. We then submit nine further legislative proposals -- some substantive, some procedural, but all involving issues that impose significant burden on taxpayers. Our proposals run the gamut from reform of the Code's six "Family Status" provisions and repeal of joint and several liability for taxpayers who file returns under married-filing-jointly status, to proposals to protect Social Security benefits from IRS levy action. However diverse these proposals may be, we advocate in each for simplification and reduced compliance burdens.

Our report also highlights some significant contrasts. These contrasts are illustrated by the first three issues we identify as Most Serious Problems for taxpayers -- trends in taxpayer service, criminal investigation refund freezes, and the cash economy. I am concerned that the IRS is decreasing the level of resources it dedicates to taxpayer service, including lowering the level of service on its toll-free lines and restricting the types and amount of service provided in its walk-in sites. The IRS justifies this reduction in service funding by the need to dedicate more resources to IRS enforcement functions. Yet the IRS is doing little to address the largest component of the tax gap -- the cash economy.¹ Meanwhile, the IRS is expending significant resources on a Criminal Investigation program that probably freezes over 300,000 refunds each year (the IRS doesn't keep track of the number), classifies taxpayers as "criminals" without providing them an opportunity to produce exculpatory evidence, continues to automatically freeze those taxpayers' refund claims for years into the future, and causes financial hardship for tens of thousands of taxpayers whose claims are legitimate but are nonetheless forced to wait 8 1/2 months or longer to receive their refunds. While the first two issues raise questions of resource allocations and organizational priorities, the last issue raises significant taxpayer rights and due process concerns.

Even where there are real successes, we see contrasts. Take the Earned Income Tax Credit (EITC), for example. Although we identify the EITC as a Most Serious Problem, the EITC Program Office and Examination staff have, in fact, been working hard to improve the program by analyzing and redesigning its procedures, learning about the characteristics and limitations of the EITC's target population, and applying that learning to its processes. EITC personnel are open to stakeholder and Taxpayer Advocate Service suggestions and engage them in dialogue. Thus, we identify many significant improvements even as we recommend additional steps.

It is important that the EITC program is making significant improvements, because 48 percent of IRS individual examinations involve the EITC despite the fact that only 7 percent of individual tax returns claim the EITC. This disproportionate attention is driven, in part, by the mandates of the Improper Payments Improvement Act. Here is another contrast: the IRS is vigorously attacking improper payments to taxpayers in the EITC but not pursuing equally vigorously the improper nonpayment of tax by taxpayers in the cash economy. The IRS is overly focusing on $9 billion of the annual tax gap attributable to low income taxpayers, while inadequately focusing on the estimated $ 00 billion-plus of the annual tax gap attributable to the cash economy.

We have tried, in our report, to point out alternative approaches to problems, both administrative and legislative, that are taxpayer-friendly and cost-effective, particularly in the long run. Over the next year, the Office of the Taxpayer Advocate will advocate for these positions and thereby seek to help the IRS fulfill its mission of serving taxpayers.

But to return to my initial premise, there comes a time when the only effective systemic change is tax reform grounded in simplification. There is a big constituency for tax simplification, after all -- the taxpayers themselves. They are speaking up for reform daily, through the phone calls they make and the letters they send to the IRS, through the number of errors they make on tax returns, through the reluctant decision of a majority of them to pay professionals to prepare their tax returns for them, through their growing awareness of the Alternative Minimum Tax. We encourage Congress to listen to these folks and take action. I hope this report will assist in that endeavor.

The Taxpayer Advocate Service (TAS) conducted a study, in collaboration with Criminal Investigation (CI), of TAS cases involving tax refunds frozen by Cl. The study period included fiscal year (FY) 2004 and the first half of FY 2005. This study serves as Volume II of the National Taxpayer Advocate's 2005 Annual Report to Congress (ARC) and documents findings that are referenced in the Most Serious Problem, Criminal Investigation Refund Freezes, in Volume I of the ARC.

BACKGROUND

From fiscal years 2002-2005, the number of cases referred to TAS as a result of CI actions to deter fraud grew by over 400 percent and became the largest single component of TAS inventory. TAS has raised the concern that the high percentage of outcomes where TAS obtains the relief sought by the taxpayer suggests problems with the CI case selection process. A review of TAS case closures for FY 2005 through March 31, 2005, reflects that 53 percent of the cases were closed with the taxpayers receiving full relief.¹

In contrast, CI statistics show a much higher rate of fraud among cases referred to TAS. To resolve this discrepancy, CI and TAS agreed to collaborate in a study of a representative sample of recently closed TAS cases involving refunds frozen by CI. In addition to reconcling differences between the TAS and CI reported outcomes, TAS also recognized that the study presented an opportunity to examine CI fraud detection and prevention procedures and to possibly identify improvements.

METHODOLOGY

TAS Research developed a sample of CI refund freeze cases that were closed by TAS during fiscal 2004 and the first half of FY 2005 (period ending March 31, 2005). An approximately equal number of cases was selected for each fiscal year. The sample cases were randomly selected using the Taxpayer Advocate Management Information System (TAMIS) Quality Review random selection application.²

TAS Research chose a sample size of 500 to ensure that a sufficient number of cases would be available to allow study findings to be projected to the population of TAS CI refund freeze cases closed during the study period with a margin of error no greater than +/- five percent at the 95 percent confidence level.

TAS Research used Master File data to conduct the vast majority of tile analyses contained in this report, since reliance on Master File data minimizes data transcription errors inherent in the manual data collection process.³ The analyses use transaction code data indicating the issuance or reversal of refunds to determine whether a case contained indicia of fraud. By comparing the refund actually issued to the refund claimed on the return, the analyses clearly show whether the claimed refund was received. The definitions for determining whether indicia of fraud existed follow:

• No Fraud. Fraud is considered to benotpresent, if the taxpayer received the full refund after the tax module had been placed in freeze status. In all instances, the refund issued must have equaled or exceeded the amount claimed on the original return.

• Fraud. Fraud is considered present if the taxpayer's refund was reduced or reversed while the module was in freeze status. Since CI sometimes identifies fraud after the refund is released, we also considered fraud to be present if the refund was released and a freeze was later put on the module during the same calendar year.Note that this is a very broad definition. Although some or all of the refund was disallowed, we did not determine the reason for disallowance.

• Undecided. TAS Research found no indication that the refund claimed on the return was reduced or reversed, but no refund had been issued. These cases appeared not to have been worked to completion.

FINDINGS

Principal study findings included:

• Most of the returns in the sample were in the lower income strata, and on the average the frozen refund represented over 25 percent of the taxpayers' yearly incomes. About 75 percent of the taxpayers claimed the Earned Income Tax Credit (EITC).

• We found that nearly two-thirds (66 percent) of the decided cases in our sample were not fraudulent, and that over 80 percent received at least a partial refund.

• Taxpayers had to wait about nine months, on average, to receive their refunds. The delay is longer for some, as one out of every five taxpayers in this group did not receive their refund for over a year.

• Over one half of the taxpayers in our sample had an adjusted gross income (AGI) below $13,000 and many ultimately received the EITC.⁴ As stated above, on average the frozen refund represented over 25 percent of the taxpayers' yearly income (and much more for those with lower incomes). These findings suggest that the extended delays often impose a severe economic hardship on taxpayers who ultimately received refunds.

• In almost 18 percent of the sample cases, CI prohibited TAS from contacting the taxpayer in response to the taxpayer's inquiry about the frozen refund. In over 20 percent of these cases, the taxpayer ultimately received a full refund. It is possible that more taxpayers would have received a full refund if TAS had been given permission to contact them.

• Only six percent of taxpayers in the sample were found to have repeat indicia of fraudulent actions.⁵ In over 85 percent of these cases, the CI identification of fraud was too late to stop issuance of the refund in the initial year. This suggests that CI efforts to prevent future fraud are effective if the compliance action materially affects the taxpayer. This finding also suggests that CI's current policy of freezing future refunds for a certain number of years may be overly restrictive.

• At most, after considering interest paid on delayed refunds, CI protected less than 20 percent of the refund amounts originally frozen in the sample.⁶ The amount of interest paid to taxpayers who ultimately received their refunds was more than 15 percent of the amount protected. Depending on the outcome of one large undecided case, the amount of revenue protected may actually be negative.

Review marginal Revenue

TAS recommends that CI and the Business Operating Divisions determine the marginal revenue gained by working different categories of returns identified as fraudulent. The marginal revenue gained by working these cases can then be compared to the yield gained in Examination by working non-fraudulent cases to see if the expenditure of resources on some of these cases is merited.

Consider alternative treatments

CI should consider alternative treatments for addressing noncompliance among taxpayers suspected of fraud. These alternative treatments may encompass methods similar to those used currently by Examination. For example, a "soft" letter may be an effective tool for addressing noncompliance among taxpayers suspected of fraud, where the dollar amount of the alleged noncompliance is relatively small or there is evidence of a "facilitator" (e.g., a preparer).

Notify taxpayers of frozen refunds

CI should notify taxpayers soon after their refunds are frozen that their refund claims are being reviewed and will not be released until a determination 'is made. Notification should include complete coverage of taxpayer rights (e.g., their right to file suit in Tax Court to claim their refunds) and a list of resources available to taxpayers for assistance (e.g., TAS and the Low Income Taxpayer Clinics).

Rank suspected fraudulent refund returns

CI should consider using its scoring algorithm to rank cases suspected of fraud. This would enable CI to freeze only those cases that have the highest scores and greatest risk of fraud, and for which the IRS can ensure timely processing.

Review practice of freezing refunds on returns filed for tax years subsequent to the fraudulent return

CI should review its policy of freezing refunds for a certain number of subsequent tax years after fraud is initially detected. Study findings show that the likelihood of taxpayers repeating fraudulent behavior is very small. Further, the resources devoted to these subsequent periods could be better utilized in working new inventory more expeditiously. Subsequent returns will also be reviewed automatically by CI's ongoing fraud detection procedures. TAS is concerned that this practice could potentially harm taxpayers who qualify for refunds by discouraging them from filing (e.g., taxpayers who qualify for the EITC).

Review the 180 day period during which other IRS organizations cannot assist the taxpayer

CI should consider shortening this period to 60 or 90 days from the date the refund freeze begins. This would help alleviate taxpayer burden by enabling TAS to assist taxpayers much sooner. TAS also recommends that CI develop expedited procedures for certain circumstances, including economic hardship, EITC, and identity theft cases.

The Taxpayer Advocate Service (TAS) routinely reviews its inventory to detect issues affecting taxpayers that are indicative of IRS systemic problems. In this regard, TAS has noted that the volume of Criminal Investigation (CI) cases among TAS receipts has increased dramatically in the last four years. These cases are now the number one source of taxpayer cases coming to TAS. Figure 1 shows the growth in these cases. These cases predominantly involve taxpayers whose refunds have been frozen by CI, although taxpayers may not be aware of CI's involvement in the case.⁷

 

FIGURE 1: TAS RECEIPTS OF CI CASES⁸

 

 

 

BACKGROUND

CI has ten Fraud Detection Centers (FDCs) charged with identifying fraudulent claims for income tax refunds and preventing their issuance. When the FDCs identify cases as having certain indicia of fraud, they freeze the taxpayer accounts causing extensive delays in the issuance of refunds.

CI manages several programs to detect tax fraud, including the Questionable Refund Program (QRP) and the Return Preparer Program (RPP). Fraud Detection Centers implement these programs to freeze accounts that meet certain criteria, thereby stopping all processing related to the account, including the issuance of refunds.⁹ CI also freezes other taxpayer accounts during its investigation of return preparers.

CI initiates most refund freezes under the QRP Program. QRP uses a scoring algorithm to analyze the tax returns submitted by taxpayers seeking refunds to identify potentially fraudulent returns. Although most QRP schemes are detected during the scoring process, some are identified through communications from electronic return originators, financial institutions, return preparers, and concerned citizens.¹⁰

A contractor developed the QRP scoring algorithm by using data mining software¹¹ to analyze fraudulent schemes detected in prior year returns. When the score assigned to a return exceeds a predetermined threshold, the return is selected for further review.

To prevent questionable refunds from being automatically issued, the selected returns are resequenced. Resequencing removes returns from processing until the next refund processing cycle. Since each refund processing cycle is one week, the resequencing step gives CI one week to analyze the refund claim further to determine whether a freeze should be placed on the account. In processing year 2004, CI resequenced over one million refund claims.¹²

During this one week period, Cl manually reviews returns to determine which should be further verified. The verification process is used to address the most prevalent illegal refund schemes, including:

• Falsifying Form W-2 withholding to obtain refunds.

• Fraudulently misstating income to obtain a refund from the EITC.

After verification, if fraud was not detected, the IRS releases the refund to the taxpayer. If CI believes the refund claim is fraudulent, it puts the account in Z freeze status.¹³ This prevents a refund from being issued unless CI subsequently reverses or overrides the Z freeze. All cases with a Z freeze that involve refundable credits like the EITC are then referred to Examination for final resolution. If Examination determines that the taxpayer is due a refund, CI must approve this determination and reverse or override the Z freeze so that the refund can be released to the taxpayer. Once a Z freeze has been placed on a taxpayer's account, returns for subsequent tax years will automatically go directly into Z freeze status and must be reviewed by CI.

TAS has raised the concern that the high percentage of outcomes where TAS obtains the relief sought by the taxpayer suggests problems with the Cl case selection process. A review of TAS case closures for FY 2005 through March 31, 2005, reflects that 53 percent of the cases are closed with the taxpayers receiving full relief.¹⁴

In contrast, CI statistics show a much higher rate of fraud among cases referred to TAS. To resolve this discrepancy, CI and TAS agreed to collaborate in a study of a representative sample of recently closed TAS CI refund freeze cases. TAS also recognized that the study presented an opportunity to examine CI fraud detection and prevention procedures and to possibly identify improvements.

METHODOLOGY

Sample Selection Process

This study was conducted on a sample of TAS cases involving tax refunds frozen by Cl and closed from the beginning of FY 2004 through the first half of FY 2005. Slightly more than half of the sample was selected from FY 2004, and the remaining cases were selected from FY 2005.¹⁵ The sample cases were randomly selected using the Taxpayer Advocate Management Information System (TAMIS) Quality Review random selection application.¹⁶

TAS Research chose a sample size of 500 to ensure that a sufficient number of cases would be available to allow study findings to be projected to the population of TAS Cl refund freeze cases closed during the study period with a margin of error no greater than +/- 5 percent at the 95 percent confidence level. Ultimately, 27 cases were removed from the sample because they did not fall within the sample parameters. For example, some of the cases did not involve frozen refunds, and others were removed because insufficient information was available to analyze them. After dropping these cases, the final sample consisted of 473 cases.

DATA COLLECTION AND ANALYSIS

TAS, with input from CI, developed mo, comprehensive Data Collection Instruments (DCIs) to capture the relevant data.¹⁷ The TAS and CI DCIs focused on different characteristics of the sample cases. The DCIs were designed to capture information from the sample case files that could be used-to supplement the data available from Master File.

There was significant overlap of the data elements contained on the CI and TAS DCIs to allow for cross-checks on critical issues. Both DCIs contained information on the case outcome, the presence or absence of indicia of fraud, and time frames for case resolution. In addition, the CI DCI contained information on interaction with TAS, such as Operations Assistance Request (OAR) receipts.¹⁸ The TAS DCI contained information on the presence or absence of the EITC, taxpayer AGI, and TAMIS database information on OAR issuance.

TAS collected data at the IRS's Austin Campus, relying on a combination of IDRS and TAMIS data to complete the DCIs. CI sent its DCIs to the FDCs for completion. Each sample case was reviewed by the FDC responsible for working the case originally.

Coincident with the manual case reviews conducted by TAS and CI, TAS Research obtained data on the sample cases from the Compliance Research Information Tracking System (CRITS). CRITS provides access to requested data elements available on the online master file for the list of Taxpayer Identification Numbers (TINs) supplied with the CRITS request. TAS Research used CRITS data to verify, to the extent possible, the data collected and analyses conducted during the manual review. TAS shared the CRITS data and all analyses conducted on both the DCI information and the CRITS data with Cl for verification.

TAS Research used Master File data to conduct the vast majority of the analyses contained in this report, since reliance on Master File data minimizes data transcription errors inherent in the manual data collection process.¹⁹ The analyses use transaction code data indicating the issuance or reversal of refunds to determine whether a case contained indicia of fraud. By comparing the refund actually issued to the refund claimed on the return, the analysis clearly shows whether the claimed refund was received. A separate set of transaction codes was utilized to determine when CI involvement in the case began. The definitions for determining whether indicia of fraud existed follow:

• No Fraud:Fraud is considered to benotpresent if the taxpayer received the full refund after the tax module had been placed in freeze status. In all instances, the refund issued must have equaled or exceeded the amount claimed on the original return.

• Fraud.Fraud is considered present if the taxpayer's refund was reduced or reversed while the module was in freeze status. Since CI sometimes identities fraud after the refund is released, we also considered fraud to be present if the refund was released and a freeze was later put on the module during the same calendar year.Note that this is a very broad definition. Although some or all of the refund was disallowed, we did not determine the reason for disallowance.

• Undecided: TAS Research found no indication that the refund claimed on the return was reduced or reversed, but no refund had been issued. These cases appeared not to have been worked to completion.

Limitations

The sample data for this project is only representative of taxpayers with refunds frozen by CI who seek assistance from TAS. While the sample results may have implications for the population of refunds frozen by Cl, the results cannot be explicitly attributed to the population of refunds frozen by CI.

Additionally, the definitions TAS Research used to determine the presence of fraud are very broad. In this study, any reduction in the refund originally claimed by the taxpayer is considered as an indication of fraud. It is likely that in some cases the adjustments to the refund amount resulted from honest taxpayer error. See Appendix I for a discussion of the legal basis for fraud determinations.

Another limitation is that the refund amount was based solely on the amount actually received by the taxpayer. No data was obtained to determine those instances in which the originally claimed refund was awarded but the amount actually received by the taxpayer was reduced because of an offset of part or all of the refund to another tax or obligation collected by the IRS. It is likely that some taxpayers in the sample who were classified as fraudulent actually had their refunds reduced by offset and were not fraudulent.

OBJECTIVES

The principal objectives of the study were to:

I. Describe the characteristics of taxpayers coming to TAS because their refunds were frozen by CI.

II. Determine the number of taxpayer refunds held where the CI refund freeze was effective (in cases of fraud) or ineffective (in cases where fraud was not substantiated and the refund was subsequently released).

III. Determine the length of the CI freeze that holds taxpayer refunds.

IV. Determine the likelihood that taxpayers will become repeat filers of fraudulent returns.

V. Compute the amount of revenue actually protected and compare to interest that is paid on subsequently released refunds.

VI. Determine if TAMIS is correctly coding CI issue cases.

FINDINGS

The findings are ordered by objective, and a restatement of each objective precedes a brief summary of the key findings relevant to this objective (see italicized text). The research issues pertinent to the objective are then listed, followed by a more detailed review of the findings.

Objective I: Describe the characteristics of taxpayers coming to TAS because their refunds were frozen by CI.

When considering all 473 current year returns in the sample, half of all taxpayers had adjusted gross incomes below $12,850, with 80 percent of all sample returns having an adjusted gross income less than $23,000. Nearly three quarters (74 percent) of the sample returns claimed the Earned Income Tax Credit (EITC), and a majority of these taxpayers ultimately received the EITC.

All of the 473 sample cases were analyzed to determine the following characteristics²⁰ regarding these returns:

• Referral of the case to TAS.

• Return characteristics including:

• • The income of taxpayers included in the sample.

  • The number claiming EITC.

  • The average size of the refund frozen by CI.

  • The filing status claimed on the return.

Eighty-nine percent of the sample cases entered TAS inventory through an Accounts Management (AM) referral. Figure 2 shows that no other source accounts for a significant volume of these cases in the TAS inventory.

 

FIGURE 2: RECEIPT OF CI CASES IN TAS

 

 

 

An examination of the income of the sample cases indicates that most taxpayers who seek TAS assistance to resolve refunds frozen by CI are in the lower strata of income. The distribution of AGI for all 473 taxpayers in our sample is shown in Figure 3 as the line labeled "All Cases."

 

FIGURE 3: ADJUSTED GROSS INCOME BROKEN OUT BY ALL CASES VERSUS CASES

 

WITH EITC ONLY

 

 

 

Nearly 75 percent of the taxpayers who claimed the EITC were allowed the credit after IRS processing. The average amount of the EITC initially claimed was $2,853.²¹

Most of the refunds frozen by CI were sizeable, as half of all taxpayers in our sample received a refund of $3,500 or more.²² As illustrated in Figure 4, the refund amounted to over 25 percent of the AGI for half of all taxpayers.

 

FIGURE 4: MEDIAN REFUND RETURNED AFTER IRS PROCESSING VERSUS MEDIAN

 

AGI

 

 

 

The study addressed the following questions pertinent to this objective:

• What is the current status of the sample cases in regard to the presence or absence of fraud?

• What is the current status of the EITC sample cases in regard to the presence or absence of fraud?

• How many of the sample taxpayers had refunds frozen for at least four consecutive years and received a full refund each year?

• How many taxpayers have refunds which remain frozen?

• How many cases contained a CI prohibition that TAS could not respond to the taxpayer's refund inquiry?

• How many sample cases had their refunds frozen because of stolen identity issues?

The majority of sample returns were found to have no indicia of fraud, as shown in Figure 5. The outcome of approximately 16 percent of the sample cases remained undecided. We also found that almost 75 per cent of the sample returns had claimed the EITC.

 

FIGURE 5: PRESENCE OF FRAUD²⁵

 

 

 

Multiple Year Frozen Refunds

While most of the analyses for this study focused on the sample tax year, analysis was also completed on the four consecutive tax years from 2001 through 2004. TAS Research conducted this multiple year analysis to explore the effectiveness of CI's policy to freeze refunds for years subsequent to the first incidence of refund fraud, as well as to examine the associated burden on taxpayers who continue to experience these refund freezes.²⁷ All four of the tax years with available data (2001 through 2004) were analyzed. This analysis showed that 29 percent of the taxpayers in the sample had their refunds frozen for four consecutive years, but ultimately received full refunds in all of these years.

Additional Modules With Refunds Still Frozen

As previously mentioned, TAS Research analyzed the tax returns filed by taxpayers in the sample for Tax Years 2001 through 2004. In addition to the 75 sample tax periods where a decision on the outcome of the refund had not yet been made, 23 percent of the taxpayers in the sample (108 different taxpayers) had at least one other return with its refund still frozen.²⁸

TAS Prohibited from Contacting Taxpayer

In almost 18 percent of the sample cases, CI prohibited TAS from contacting the taxpayer in response to the taxpayer's inquiry about the frozen refund.²⁹ In over 20 percent of these cases, the taxpayer ultimately received a full refund. It is possible that more taxpayers would have received a full refund if TAS had been given permission to contact the taxpayer.

Objective III: Determine the length of the CI freeze that holds taxpayer refunds.

Over two-thirds (67 percent) of all sample cases resulted in the issuance of at least a partial refund to taxpayers. When considering only the decided cases, over 80 percent of the sample taxpayers received some or all of their originally claimed refunds. The median wait for these refunds was nearly nine months (over 37 weeks). Over 70 percent of these taxpayers were real also entitled to EITC.

The specific issues addressed in conjunction with this objective included the following:

• The elapsed time to resolve all frozen refunds.

• The elapsed time to resolve frozen refunds that include EITC.

• The elapsed time to resolve identity theft cases.

• Elapsed Time from Return Filing to Receipt of Refund

The time from return filing to refund issuance was used to determine the length of time CI froze refunds.³⁰ Figure 6 depicts the average elapsed time (in weeks) from return filing to refund issuance for all sample case refund recipients and for those sample refund recipients entitled to EITC.

 

FIGURE 6: AVERAGE LENGTH OF TIME CI FREEZES REFUNDS

 

 

 

In a small subset of the sample cases, the taxpayers were the victims of identity theft.³¹ These cases comprised about six percent (27 cases) of the sample. Although not a significant portion of the sample, these cases are of particular concern because the taxpayers whose identities were stolen experienced frozen refunds as a result of someone else's illegal actions. After examining the sample cases, we determined that 25 of the 27 taxpayers were issued a refund.³² At 38 weeks, the median time necessary to issue refunds on identity theft cases was similar to that for all refund cases in the sample. During our analysis of the four tax years from 2001 through 2004, we determined that 30 refunds³³ were frozen in tax years after the original freeze. The median time for issuance of refunds on these subsequent tax years was six months.³⁴

Objective IV: Determine the likelihood that taxpayers will become repeat filers of fraudulent returns.

Only six percent (29 cases) of the sample cases had indicia of repeat fraud. Moreover, in 25 of the 29 cases, the initial identification of possible fraud was not in time to freeze the refund. Therefore, less than one percent of all taxpayers in the sample were repeat offenders if they had their refunds frozen in a prior year.

The following component questions of this objective are addressed below:

• How many taxpayers with prior indicia of fraud file a subsequent refund return with indicia of fraud?

• How many of the sample cases showed no indication of repeat fraud?

• What is the effect of the timing of the CI refund freeze?

Figure 7 displays the probability that a taxpayer will repeatedly submit refund returns indicative of fraud.³⁵ It illustrates that about 75 percent of the sample cases show no indication of repeat fraud, while only about six percent of the sample cases have indicia of repeat fraud. The presence or absence of repeat fraud in 19 percent of the cases remained undecided, because either a fraud determination had not been made in the sample year or in a year subsequent to the sample year.

 

FIGURE 7: INCIDENCE OF PRIOR YEAR FRAUD

 

 

 

The vast majority of the cases with an indication of repeat fraud (25 out of 29 cases) were those where the refund freeze was applied too late to stop the refund from being issued when fraud was first suspected. Accordingly, less than one percent of the sample cases are representative of instances where the taxpayer has demonstrated a willingness to repeat non-competent behavior after being previously impacted by CI compliance actions.

Objective V: Compute the amount of revenue actually protected and compare to interest that is paid on subsequently released refunds.

When considering all sample cases, more than 8 0 percent of The refund amounts frozen by CI are ultimately returned to the taxpayer often with interest.

The following questions will be addressed in this section:

• What was the dollar amount of refunds frozen by CI?

• What was the final amount of refund after CI intervention?

• What was the amount of revenue protected by the refund freeze?

• How much interest was paid by the IRS on frozen refunds ultimately released to the taxpayer?

• What was the outcome of cases input into the Scheme Tracking and Referral System (STARS)?

TAS Research determined the amount of revenue protected by subtracting the refund actually issued to the taxpayer (based on the net Master File transaction codes for refund issuance and reversal) from the refund amount claimed on the original return (IRS computed amount after return processing).³⁶

Figure 8 addresses the first three components of this objective by depicting the aggregate amount of refunds claimed by the taxpayer (IRS computed amounts after processing), the aggregate amount of refunds actually issued and the aggregate net revenue protected for all sample cases. Revenue protected is computed as the difference between the refund claimed on the return (IRS computed amount after processing) and the amount of refund actually issued by the IRS.

          FIGURE 8: NET REVENUE PROTECTED (ALL SAMPLE CASES)

Figure 8 includes all 473 sample cases. The table also displays two "outlier" cases separately: one undecided case in which a large refund (over $1 million) was claimed, and one decided case in which the actual refund exceeded the refund claimed on the original return (IRS computed amount after processing) by over $1 million. The net amount of revenue protected totals nearly $575,000. If both outlier cases are removed, the revenue protected would increase to nearly $700,000.³⁹ If the undecided case with the refund claimed of over $1.2 million is determined to be due the taxpayer, however, the net revenue protected will be a negative amount, meaning that the IRS refund freeze will have actually cost the government over $650,000.⁴⁰ Over $81,000 of interest is included in the final refunds issued to the taxpayers because of the delay in receiving the refund due. Additional interest may also be due taxpayers for the 75 undecided cases.

The average amount of protected revenue for all sample cases is slightly more than $1,200. This amount is significantly less than the average dollars recommended on returns selected for examination.⁴¹

 FIGURE 9: NET REVENUE PROTECTED (ALL CASES THAT HAVE BEEN DECIDED)

Analysis of STARS Cases

CI uses the Scheme Tracking and Referral System (STARS) to estimate the amount of revenue protected. When CI determines that a return is fraudulent, after manual review and verification, it places the account in "Z" freeze status and includes the return in the STARS system.

CI appears to have made an initial determination of fraud on 171 of the sample cases. TAS Research identified these cases by finding the earliest instance of a Z freeze and comparing the input cycle of the CI freeze to the return posting cycle. In all 171 cases, the earliest CI freeze occurs on or after the return posting cycle, indicating that the return was frozen based on its characteristics, rather than due to a determination of fraud in a prior year.

Among decided cases (142 out of 171), about 46 percent (65 out of 142 cases) reflect an erroneous determination of fraud by CI, as demonstrated by Figure 10. Although these cases were placed in Z freeze status by CI, the taxpayer ultimately received a full refund.

 FIGURE 10: DETERMINATION OF CASES WHERE MANUAL REVIEW INDICATES FRAUD

Objective VI: Determine if CI cases are correctly coded on TAMIS.

An analysis of the Master File data clearly indicates that the majority of the CI refund cases referred to TAS do ultimately result in the issuance of a full refund. Overall, a re view of all sample cases showed that about 90 percent were correctly coded by TAS.

The following facets of this objective will be analyzed:

• A comparison of TAS relief coding to the final case resolution.

• An analysis of TAS accuracy of relief coding.

There is a significant discrepancy between CI and TAS statistics regarding the outcome of CI refund freeze cases. In general, TAS records indicated that the majority of taxpayers received relief (i.e., the refund was released), while CI records indicated that most taxpayers were appropriately denied the claimed refund. While TAS does not have access to CI data regarding the outcome of the frozen refund returns, Master File data supports TAS' contention that most of the frozen refund cases in its inventory do receive a full refund. In fact, as previously indicated, about 56 percent (263 of 473) of the taxpayers in our sample received a full refund. When considering only the decided cases, about two-thirds (263 of 398) of the taxpayers received a full refund. An analysis of the overall TAS coding on the sample cases indicated that the vast majority were coded correctly.⁴³ Figure 11 compares the TAS coding of relief (if any) provided to the taxpayer to the actual outcome of the case.

 FIGURE 11: ANALYSIS Of TAMIS CODING

No Sample Year Fraud and TAMIS coded "No Relief"

There were 37 cases where no sample year fraud was detected (full refund was released), but TAS coded the closed case as "no relief." TAS closed 29 (78 percent) of these cases prior to the release of the refund, and therefore coded them correctly. Of the remaining eight cases, five were coded as "no relief" as a result of the taxpayer not receiving the requested relief even though the refund was released (e.g., frozen refund was released, but offset to another debt, so taxpayer never received refund). The remaining three cases were coded incorrectly on TAMIS.

Sample Year Undecided and TAS Coded "Full Relief"

Only two cases have these characteristics. TAS reviewed both cases and determined that they were coded incorrectly.

No Sample Year Fraud Indication and TAS Coded "Full Relief"

Of the 42 cases in this category, 20 (48 percent) cases included the input of an amended return. Therefore, the TAS coding is correct, but the return is still considered to have an "indication" of fraud since all of the refund claimed on the original return was not received. The remaining 22 cases (52 percent) were coded incorrectly by TAS.

No Sample Year Fraud and TAS Code "Partial Relief"

A manual review of these cases shows that eight of the 13 cases (62 percent) were coded correctly by TAS. For the correctly coded cases, the "partial relief" designation is because the taxpayer had come to TAS with other issues (often other tax year refunds which were frozen by CI) which could not be resolved at the time of case closing.

CONCLUSIONS

I. Describe the characteristics of taxpayers coming to TAS because their refunds were frozen by CI.

Most of the taxpayers in the sample were in the lower income strata, and on the average, the frozen refund represented over 25 percent of their yearly incomes. About 75 percent of the taxpayers in the sample were eligible for the Earned Income Tax Credit.

Nearly all of the cases are referred to TAS by Accounts Management (AM), and this finding has several important implications for the IRS. One, it suggests the important role that AM plays in ensuring that taxpayers who need assistance from TAS are properly directed to TAS. Two, it suggests that opportunities to improve taxpayer awareness of TAS exist within the other organizations responsible for TAS referrals.

The taxpayers in our sample who claimed the EITC were no more likely to have committed fraud than those taxpayers who did not claim the EITC. This finding suggests that taxpayers who claim EITC are no more likely to engage in fraud than taxpayers who do not claim the EITC.

We determined that in six percent of the sample cases, the taxpayers were victims of identity theft.⁴⁴ This finding demonstrates another problem that identity theft has created for taxpayers and tax administration, in general.

II. Determine the number of taxpayer refunds held effectively (in cases of fraud) or ineffectively (in cases where fraud was not substantiated).

We found that nearly two-thirds (66 percent) of the decided cases in our sample were not fraudulent. This is a concern, since TAS had over 28,000 CI related cases in FY 2005 and CI receipts have increased rapidly over the last several years. Moreover, it is likely that at least some portion of taxpayers who have their refunds frozen by CI, but do not seek TAS assistance, also experience significant delays in receiving their refunds.

From FYs 2002-2005, the number of cases referred to TAS as a result of CI actions to deter fraud has grown by over 400 percent. Given our sample results, we are 95 percent confident that between 62 percent and 70 percent of all taxpayers who came to TAS because of CI actions did not commit fraud. Given the FY 2005 case count, this finding suggests that between 17,535 and 19,797 of the taxpayers who came to TAS for assistance with frozen refunds during the fiscal year had not committed fraud.⁴⁵ The definition TAS Research used to determine the presence of fraud is very broad. In this study, any reduction in the refund originally claimed by the taxpayer is considered as an indication of fraud. It is likely that in some cases the adjustments to the refund amount resulted from honest taxpayer error. In addition, the refund amount was based solely on the amount actually received by the taxpayer. Data was not obtained to determine those instances in which the originally claimed refund was awarded, but the amount actually received by the taxpayer was reduced because of an offset of part or all of the refund to another tax or obligation collected by the IRS. It is likely that some taxpayers in the sample who were classified as fraudulent actually had their refunds reduced by offset and were not fraudulent.

In almost 18 percent of the sample cases, CI prohibited TAS from contacting the taxpayer in response to the taxpayer's inquiry about the frozen refund. In over 20 percent of these cases the taxpayer ultimately received a full refund. It is possible that more taxpayers would have received a full refund if TAS had been given permission to contact the taxpayer. In light of this finding, TAS believes that CI should reevaluate its policy concerning communication with taxpayers in freeze status.⁴⁶

III. Determine the length of the CI freeze that holds taxpayer refunds.

Typically, it took about nine months for taxpayers to receive their refunds. The delay is longer for some, as one out of every five taxpayers in this group did not receive their refunds for over a year. TAS is concerned that this delay imposes a severe economic hardship on many taxpayers.

Over one half of the taxpayers in the sample had an adjusted gross income (AGI) below $13,000, and many ultimately received the EITC. The value of the taxpayers' refunds also represents over 25 percent of the income for the average taxpayer and much more for those with lower incomes. The extended delay undoubtedly forces some of these taxpayers to seek other high cost sources of funds (e.g., high interest payday loans, etc.).

TAS believes that additional consideration should be given to funding limitations that exist within CI and Examination. CI is currently not limiting the number of cases in which it freezes refunds to only those that can be processed within a reasonable timeframe.

IV. Determine the likelihood that taxpayers will become repeat filers of fraudulent returns.

Only six percent of taxpayers in the sample were found to have repeat indicia of fraudulent actions.⁴⁷ In over 85 percent of these cases, the CI identification of fraud was too late to stop issuance of the refund in the initial year. This suggests that CI's efforts to prevent future fraud are effective if the compliance action materially affects the taxpayer. This finding also suggests that CI's current policy of freezing future refunds for a certain number of years may be overly restrictive and administratively inefficient.

V. Compute the amount of revenue actually protected and compare to interest that is paid on subsequently released refunds,

At most, after considering interest paid on delayed refunds, CI protected less than 20 percent of the refund amounts originally frozen.⁴⁸ The amount of interest paid to taxpayers who ultimately received their refunds was more than 15 percent of the amount protected. Depending on the outcome of one large undecided case, the amount of revenue protected may actually be negative.

CI uses the STARS system to estimate the amount of revenue protected. When CI determines that a return is fraudulent, after manual review and validation, it places the account in "Z" freeze status and includes the return in the STARS system. Given that among the decided STARS cases in our sample, approximately 46 percent of the taxpayers ultimately received full refunds, TAS believes that the CI manual review process should be improved to increase accuracy and reduce taxpayer burden.

VI. Determine If TAMIS is correctly coding CI cases.

An analysis of the Master File data clearly indicates that the majority of the CI refund cases referred to TAS do ultimately result in the issuance of a full refund. Overall, a review of all sample cases showed that about 90 percent were correctly coded by TAS.

RECOMMENDATIONS

TAS recommends that CI and the Business Operating Divisions determine the marginal revenue gained by working different categories of returns identified as fraudulent. The marginal revenue gained by working these cases can then be compared to the yield gained in Examination by working non-fraudent cases to see if the expenditure of resources on some of these cases is merited.

CI should consider alternative treatments for addressing noncompliance among taxpayers suspected of fraud. These alternative treatments may encompass methods similar to those used currently by Examination. For example, a "soft" letter may be an effective tool for addressing noncompliance among taxpayers suspected of fraud, where the dollar amount of the alleged noncompliance is relatively small or there is evidence of a "facilitator" (e.g., a preparer).

In collaboration with the SB/SE and W&I Operating Divisions and TAS, CI should consider conducting a similar research study on a sample of refund freeze cases worked without TAS assistance. Study design presents special challenges, since many taxpayers, especially those in the lower income strata, may not have received refunds they were entitled to because they were unable to effectively pursue their claims. This study, combined with the results herein, would allow the IRS to better assess the effectiveness of CI's refund freeze program. The IRS should also consider establishing a cross functional team to review the findings and recommendations from both studies. A balanced review will lead to improved program administration and reduced taxpayer burden.

CI should notify taxpayers soon after their refunds are frozen that their refund claims are being reviewed and will not be released until a determination is made. Notification should include complete coverage of taxpayer rights (e.g., their right to file suit in tax court to claim their refunds) and a list of resources available to taxpayers for assistance (e.g., TAS and the Low Income Taxpayer Clinics).

CI should consider using its scoring algorithm to rank cases suspected of fraud. This would enable CI to freeze only those cases that have the highest scores, greatest risk of fraud, and for which the IRS has sufficient resources to ensure timely processing.

CI should review its policy of freezing refunds for a certain number of subsequent tax years after fraud is initially detected. Study findings show that the likelihood of taxpayers repeating fraudulent behavior is very small. Additionally, the resources devoted to these subsequent periods could be better utilized to work new inventory more expeditiously. Also, subsequent returns will be reviewed automatically by CI's ongoing fraud detection procedures. TAS is concerned that this practice could potentially be harming taxpayers who qualify for a refund by discouraging them from filing (e.g., taxpayers who qualify for the EITC).

CI should consider shortening the 180 day period during which other IRS organizations cannot assist the taxpayer to 60 or 90 days from the date the refund freeze begins. This would help alleviate taxpayer burden by enabling TAS to assist taxpayers much sooner. TAS also recommends that CI develop expedited procedures for certain circumstances, including economic hardship, EITC, and identity theft cases.

BADGES OF FRAUD

Fraud is established by demonstrating that a taxpayer intended to evade a tax that was due and owing by conduct intended to conceal, mislead or otherwise prevent the collection of such tax.⁴⁹ Mere negligence is not fraud.⁵⁰ Courts have established certain indicia of fraud (often referred to as "badges of fraud"), including:

• Understating income;

• Maintaining inadequate records;

• Giving implausible or inconsistent explanations;

• Concealing income or assets;

• Failing to cooperate with tax authorities;

• Providing incomplete or misleading information to one's tax preparer; and

• Filing false documents and dealing primarily in cash.⁵¹

The federal government has an array of civil and criminal statutes it can employ to punish and deter fraud,⁵² including:

• Civil fraud penalty for fraudulent underpayment of tax of 75 percent of the underpayment;⁵³

• Civil penalty for fraudulent failure to file of 75 percent of the amount required to be shown as tax on the return;⁵⁴

• Criminal sanctions of imprisonment for up to 3 years and monetary fines of up to $100,000 ($500,000 for corporations) for filing false returns, statements or other documents under penalties of perjury with the IRS;⁵⁵

• Criminal sanctions of imprisonment for up to 5 years and fines for filing false, fictitious or fraudulent claims with any U.S. government agency or office.⁵⁶

Burden of Proof in Fraud Cases

Unlike in deficiency cases, where the Commissioner is entitled to the presumption of correctness, the IRS bears the burden of proving civil and criminal fraud. Civil and criminal Federal income tax fraud contain identical elements, but the degree of required proof differs.⁵⁸ Criminal conviction of a tax offense requires the Department of justice to establish each element of the crime beyond a reasonable doubt.⁵⁹ A civil fraud case requires the government to prove both the underpayment of tax and fraud by clear and convincing evidence.⁶⁰ A criminal conviction for income tax evasion collaterally stops the taxpayer from relitigating the issue in the civil proceedings.⁶¹

Proof of fraud with intent to evade a tax due and owing may be established either by direct evidence or by indirect or circumstantial evidence. Some of the methods the IRS uses to prove fraud include the so-called "badges of fraud," the net worth⁶² method, the expenditures method,⁶³ and the bank deposits method.⁶⁴

CI selects "for investigation all information items which, upon evaluation and screening, are deemed to warrant inquiries . . ." ⁶⁵ Thus, when the QRP utilizes data mining software to search for returns claiming refunds and containing data elements that CI believes are indicators of fraud, the returns selected then warrant additional inquiries and consequently the refunds are frozen pending investigation. The initial investigation "should involve such inspection of the taxpayer's books and records or other related inquiries as is necessary to initially determine whether the case possesses criminal potential."⁶⁶

If the government is unable to sustain the greater burden of proof in criminal cases of guilt beyond a reasonable doubt, it still may be able to proceed to prove the taxpayer's civil fraud by the lesser "clear and convincing" standard. Once CI has completed its criminal investigation, any information gathered during the administrative criminal investigation can be used to prove the taxpayer's civil fraud.⁶⁷

² TAS's database, TAMIS, is dedicated to the recordation, control, and processing of TAS taxpayer cases and to the analysis of core tax issues, laws, policies, and internal IRS functional processes that are the sources of significant taxpayer hardship and other critical problems. Cases are classified by primary core issue codes (PCIs) identifying the main concern or issue at question. CI cases are designated with PCI 950.

³ The accuracy of Master File data is improved through the use of programs that check for math and transcription errors. Although the data collected from the Data Collection Instrument (DCI) was checked for errors, it is possible that some errors were committed when the data was recorded.

⁴ The Earned Income Tax Credit (EITC) is the nation's largest need-based anti-poverty program. Frank Sammartino, Eric Toder and Elaine Maag, Providing Federal Assistance for Low-Income Families Through the Tax System: A Primer, The Urban-Brookings Tax Policy Center 8 (July 2002) at http://www.urban.org/uploadedpdf./410526.pdf.

⁵ There were 29 such cases out of 454 cases.

⁶ 18.9 percent. See Figure 8 ($574,960 / $3,035,710).

⁷ When CI places a Z Freeze on a tax module indicating that CI believes the refund claim is fraudulent, TAS is precluded from informing a taxpayer of CI's involvement in a case. IRM § 13.1.10.9.

⁸ Figure 1 is based only on the TAS primary issue code. Including secondary issue code case counts, the FY 2005 count of Cl cases received into TAS inventory is 31,627.

⁹ IRM § 9.8.1.3.1.4 assigning freeze code functions to FDCs.

¹⁰ IRS Overview, Questionable Refund Program, at http://www.irs.gov/compliance/enforcement/article/ id =117528,00.html.

¹¹ Software that develops algorithms by detecting trends and patterns in data.

¹² 1,051,154. The "IRS Overview, Questionable Refund Program, at http://www.irs.gov/compliance/enforcement/article/ 0,,id=117528,00.html. The "processing year" is the calendar year in which a taxpayer files his or her return.

¹³ In FY 2003, CI verified 130,579 cases and identified 96,953, or 74 percent as fraudulent. In FY 2004, CI verified 179,138 cases and identified 118,075, or 66 percent, as fraudulent.

¹⁴ TAMIS data as recorded in IRS Business performance Management System, March 2005.

¹⁵ TAS Research selected nearly as many cases from the first half of FY2005 as from all of 2004, to assure that the sample reflected current trends in CI processing and procedures.

¹⁶ TAS database, TAMIS, is dedicated to the recordation, control and processing of TAS taxpayer cases and to the analysis of core tax issues, laws, policies and internal IRS functional processes that are the sources of significant taxpayer hardship and other critical problems. Cases are classified by primary core issue codes (PCIs) identifying the main concern or issue at question. CI cases are designated with PCI 950.

¹⁷See appendices II and III.

¹⁸ TAS does not have the statutory or delegated authority to release refunds frozen by Cl and must therefore forward the case to CI for a determination on the release of the refund.

¹⁹ The accuracy of Master File data is improved through the use of programs that check for math and transcription errors. Although the data collected from the DCI was checked for errors, it is possible that some data errors were committed when the data was recorded.

²⁰ Information obtained from the DCI was also used to determine what function referred the case to TAS. Master File data was utilized to describe the return characteristics.

²¹ The amount claimed is after IRS programs have corrected math and transcription errors. EITC math error authority also includes some compliance checks to better ensure age, relationship and Social Security Number validity.

²² The average refund was well over $6,000.

²³ Sample cases were determined to have no indicia of fraud if an amount equal to or exceeding the frozen refund claimed on the original return was received by the taxpayer. The amount of refund actually received by the taxpayer was determined by netting the Master File transaction codes for refund issuances and reversals. Data on refund offsets was not captured. It is likely that some taxpayers in the sample who were classified as fraudulent actually had their refunds reduced by off-set and were not fraudulent.

²⁴ While 56 percent of the sample cases were without fraud, when applying the weights from the stratified sample, the weighted number of cases without fraud was 55 percent. Accordingly,at the 95 percent confidence level between 51 percent and 59 percent of all taxpayers did not commit fraud. Excluding undecided cases, at the 95 percent confidence level between 62 percent and 70 percent of all taxpayers did not commit fraud.

²⁵ Total exceeds 100 percent due to rounding.

²⁶ Chi square analysis showed no significant difference between the EITC and non-EITC return fraud rates.

²⁷ The additional frozen tax years were determined to be those with a claimed refund (IRS determined amount after return processing) with no indication of the issuance of a refund based on the Master File transactions codes and with no other transactions which would reduce the claimed refund (i.e., audit referral, or audit adjustment of tax, EITC, or withholding).

²⁸ 118 different tax refunds were frozen (nine prior to sample year and 109 subsequent to sample year).

²⁹ CI is concerned that administrative contact of a taxpayer under CI investigation could adversely impact possible subsequent criminal prosecution of the taxpayer.

³⁰ The length of time that a taxpayer's refund was frozen was computed by subtracting the return posting cycle from the cycle when the refund was actually issued to the taxpayer. The cycle of refund issuance was determine to be the cycle of the latest, unreversed transaction code generating the refund.

³¹ Identify theft cases were identified from the DCI's recordation of the manual review of the sample cases.

³² Additional research is required to determine the disposition of the two cases where taxpayers did not receive refunds.

³³ Some taxpayers had refunds frozen more than one additional year.

³⁴ For a fuller discussion of identity theft issues see Most Serious Problem: Criminal Investigation Refund Freezes; and Most Serious Problem: Identity Theft, Vol. I.

³⁵ Analysis is based on 454 of 473 samples cases. Sample year cases for Tax Year 2001 were removed from the analysis because data to determine possible prior fraud was not available. Tax Year 2004 cases in which 2004 was the first year evidencing fraud were also removed from the analysis (Sample data was not available for tax years prior to 2001 or after 2004).

³⁶ The amount of revenue actually protected could be overstated, because data on refund offsets was not captured. It is likely that some taxpayers in the sample who were otherwise entitled to a full refund had their refunds reduced by offset.

³⁷ Includes interest paid by IRS to taxpayer as a result of the delay in refund issuance.

³⁸ An outlier is an individual observation that stands apart from the usual pattern where the unusual behavior may be attributable to factors other than those being studied. In this instance it represents an original return claiming a modest refund (about $5,000), which is ultimately awarded a refund in excess of $1.3 million.

³⁹See Non-outlier row in Figure 8.

⁴⁰ $574,960 - $1,231,136 = $656,176.

⁴¹ IRS, FY 2004 Data Book Table 10, Average dollars recommended for non-business returns with total positive income less than $25,000 is over $10,000.

⁴² Includes interest paid by IRS to taxpayer as a result of the delay in refund issuance.

⁴³ This analysis was based on DCI data. Sonic of the differences in outcomes reported by TAS and CI are likely attributable to instances where a reduced refund amount was ultimately issued. For example, if a taxpayer filed an amended return reducing the refund on a module frozen by CI to the correct amount, this case would be coded as a "full relief" case in accordance with TAMIS coding conventions.

⁴⁴ There were 27 cases of identity theft out of 473 cases.

⁴⁵ 66 percent +/-4 percent = 62 percent /70 percent respectively. .62*28,282= 17,535;.70*,282=19,797.

⁴⁶ Many of these taxpayers may benefit from knowing the location of the Low Income Taxpayer Clinics (LITCs) that may be able to offer legal assistance.

⁴⁷ There were 29 cases out of 454 cases.

⁴⁸ 18.9 percent. See Figure 8 ($574,960 / $3,035,710).

⁴⁹Clayton v. Comm'r, 102 T.C. 632, 647 (1994). Neither the Internal Revenue Code nor Treasury Regulations specifically define fraud.

⁵⁰Mitchell v. Comm'r, 118 F.2d 308, 310 (5th Cir. 1941) (applying the definition of fraud with respect to the fraud penalty under the predecessor of IRC § 6663). The Court stated: Negligence, whether slight or great, is not equivalent to the fraud with intent to evade tax named in the statute. The fraud meant is actual, intentional wrongdoing, and the intent required is the specific purpose to evade a tax believed to be owing.

⁵¹Spies v. U.S., 317 U.S. 492, 499 (1943); Conti v. Comm'r, 39 F.3d 658, 662 (6th Cir. 1994); Douge v. Comm'r, 899 F.2d 164, 168 (2d Cir. 1990); Bradford v. Comm'r, 796 F.2d 303, 307 (9th Cir. 1986).

⁵² The distinction between civil and criminal punishment for fraudulent conduct under the Internal Revenue Code and other federal statutes relates to the extent of sanctions that can be imposed, the burden of proof and the intended purpose of the sanction. Civil fraud penalties under the Internal Revenue Code are monetary in nature, and the government must establish culpability by clear and convincing evidence. IRC § 7454(a); see also Tax Ct. R. 142(b). Under statutes imposing criminal sanctions, punishment can include fines and imprisonment, and the government must prove all elements of the offense by a standard of "beyond a reasonable doubt." U.S. v. Boulware, 384 F.3d 794 (9th Cit. 2004). Criminal sanctions are primarily intended to punish, while civil penalties are intended to draw a distinction between acceptable and unacceptable conduct and to deter unacceptable behavior. Helvering v. Mitchell, 303 U.S. 391, 399 (1938); see also Jay Soled, Third Party Civil Tax Penalties and Professional Standards, 2004 Wis. L. Rev. 1611, 1635 (2004).

⁵³ IRC § 6663.

⁵⁴ IRC § 6651(f).

⁵⁵ IRC § 7206.

⁵⁶ 18 U.S.C. § 287.

⁵⁷ Treasury Inspector General for Tax Administration, Ref. No. 2004-10-189, Improvements Are Needed to Ensure Information Developed During Criminal investigations is Referred for Civil Action I (Sept. 2004).

⁵⁸ The Tax Division of the Department of Justice prosecutes criminal tax fraud cases following referral from the IRS and defends the IRS in civil tax fraud cases in the district courts and Court of Federal Claims. The IRS Office of Chief Counsel defends civil tax fraud cases in the Tax Court.

⁵⁹Holland v. U.S., 348 U.S. 121, 126 (1954); IRM § 9.5.1.2.2(1).

⁶⁰ IRC § 7454(a); Tax Ct. R. 142(b).

⁶¹Arctic Ice Cream Co. v. Comm'r, 43 T.C. 68, 74-75 (1964), acq., 1966-2 C.B. 4.

⁶² Unexplained deposits or acquired property (increases in net worth) in excess of reported income and cash reserves are considered the equivalent of unreported income. See IRM § 9.5.9.5. 1.1 for a discussion of the legal requirements to establish a prima facie net worth investigation.

⁶³ If a taxpayer's expenditures during a given period exceed his/her reported income and the source of the funds used to make those expenditures is unexplained, the IRS may infer that such expenditures represent unreported income. See IRM § 9.5.9.6.

⁶⁴ In some circumstances, proof of deposits can be substantial evidence of income. See Gleckman v. U.S., 80 F.2d 394, 399 (8th Cit. 1935); IRM § 9.5.9.7.

⁶⁵ IRM § 9.5.1.3.1(1).

⁶⁶ IRM § 9.5.1.3.2(1). The focus of the investigation is ascertaining the authenticity of the return and supporting documents and the responsibility of the filer. See IRM § 9.5.3.2.4(1).

⁶⁷ IRM § 9.5.14.3.1(1). However, there are limitations on using information developed during the course of a grand jury investigation for civil purposes.