GAO IRS AUDIT RATES. Rate for Individual Taxpayers Has Declined But Effect on Compliance Is Unknown



Similar documents
GAO HOME MORTGAGE INTEREST DEDUCTION. Despite Challenges Presented by Complex Tax Rules, IRS Could Enhance Enforcement and Guidance

DlNBVRGH + Sickness Absence Monitoring Report. Executive of the Council. Purpose of report

2001 Attachment Sequence No. 118

Lump-Sum Distributions at Job Change, p. 2

Health insurance exchanges What to expect in 2014

a GAO GAO COLLEGE COMPLETION Additional Efforts Could Help Education with Its Completion Goals Report to Congressional Requesters

Recognition Scheme Forensic Science Content Within Educational Programmes

Small Businesses Decisions to Offer Health Insurance to Employees

Health insurance marketplace What to expect in 2014

ClearPeaks Customer Care Guide. Business as Usual (BaU) Services Peace of mind for your BI Investment

Enterprise Risk Management Software Buyer s Guide

STATUS OF LAND-BASED WIND ENERGY DEVELOPMENT IN GERMANY

Utilization of Smoking Cessation Benefits in Medicaid Managed Care,

GAO HIGHER EDUCATION. Improved Tax Information Could Help Families Pay for College. Report to the Committee on Finance, U.S.

July 22, The Honorable Henry A. Waxman Chairman Committee on Energy and Commerce House of Representatives

FDIC Study of Bank Overdraft Programs

2015 EDITION. AVMA Report on Veterinary Compensation

Quick Reference Guide: One-time Account Update

Anthem Blue Cross Life and Health Insurance Company University of Southern California Custom Premier PPO 800/20%/20%

PROPERTY AND CASUALTY INSURANCE. Effects of the Nonadmitted and Reinsurance Reform Act of 2010

Health insurance exchanges What to expect in 2014

Vendor Rating for Service Desk Selection

Small Business Networking

Humana Critical Illness/Cancer

S a l e s Ta x, U s e Ta x, I n c o m e Ta x W i t h h o l d i n g a n d M i c h i g a n B u s i n e s s Ta x E s t i m a t e s

Econ 4721 Money and Banking Problem Set 2 Answer Key

An Undergraduate Curriculum Evaluation with the Analytic Hierarchy Process

How To Set Up A Network For Your Business

MAX. As an increasingly larger share of Medicaid enrollees MEDICAID POLICY BRIEF

How To Network A Smll Business

Corporate Compliance vs. Enterprise-Wide Risk Management

Contextualizing NSSE Effect Sizes: Empirical Analysis and Interpretation of Benchmark Comparisons

GAO HIGHER EDUCATION. Federal Science, Technology, Engineering, and Mathematics Programs and Related Trends

Small Business Networking

Guide to Reading Your Morgan Stanley Smith Barney Statement

Small Business Networking

Economics Letters 65 (1999) macroeconomists. a b, Ruth A. Judson, Ann L. Owen. Received 11 December 1998; accepted 12 May 1999

Small Business Networking

Treatment Spring Late Summer Fall Mean = 1.33 Mean = 4.88 Mean = 3.

Reasoning to Solve Equations and Inequalities

A National Look at the High School Counseling Office

Pay over time with low monthly payments. Types of Promotional Options that may be available: *, ** See Page 10 for details

File Storage Guidelines Intended Usage

How To Get Low Wage Workers Covered By Insurance Through Their Employer

Active & Retiree Plan: Trustees of the Milwaukee Roofers Health Fund Coverage Period: 06/01/ /31/2016 Summary of Benefits and Coverage:

San Mateo County ACCEL Adult-Education College and Career Educational Leadership AB 86 Adult Education Consortium Project Management Plan 24,

DEPARTMENT OF DEFENSE 6000 DEFENSE PENTAGON WASHINGTON, o_c

persons withdrawing from addiction is given by summarizing over individuals with different ages and numbers of years of addiction remaining:

Data quality issues for accounting information systems implementation: Systems, stakeholders, and organizational factors

STATE OF MONTANA Developomental Disabilities Program Comprehensive Evaluation Hi-Line Home Programs, Inc Adult Services

Why is the NSW prison population falling?

Unit 29: Inference for Two-Way Tables

Marketplace Insurance Premiums in Early Approval States: Most Markets Will Have Reductions or Small Increases in 2015

2006 IPCC Software for National Greenhouse Gas Inventories: Application and use for India

2011 Statistics on Social Work Education in the United States

BUSINESS OWNERS PACKAGE INSURANCE APPLICATION

Introducing Kashef for Application Monitoring

Small Business Cloud Services

How To Study The Effects Of Music Composition On Children

Your duty, however, does not require disclosure of matter:

Helicopter Theme and Variations

Network Configuration Independence Mechanism

Prescriptive Program Rebate Application


Policy Brief. Receipt of Public Benefits and Private Support among Low-income Households with Children after the Great Recession.

Factoring Polynomials

UNDERGRADUATE PREPARATION AND DISSERTATION METHODOLOGIES OF ACCOUNTING PHDS OVER THE PAST 40 YEARS

WIC Participant and Program Characteristics 2012 Final Report

Pre-Approval Application

JaERM Software-as-a-Solution Package

COMPARISON OF SOME METHODS TO FIT A MULTIPLICATIVE TARIFF STRUCTURE TO OBSERVED RISK DATA BY B. AJNE. Skandza, Stockholm ABSTRACT

How To Understand And Understand The Benefits Of Accounting

Estimating Exchange Rate Exposures:

AntiSpyware Enterprise Module 8.5

Portfolio approach to information technology security resource allocation decisions

City of Dade City AGENDA MEMO. participants permitted. participants Currently as the plan was approved loans to are not

2. Transaction Cost Economics

Office of Student Financial Aid Haven-Warren Room 210

Learner-oriented distance education supporting service system model and applied research

I I I I I I I I I 1 I I I 1 I I. Regional Economic Multipliers. A Consumer s Guide to. Cletus C. Coughlin and Thomas B. Mandelbaum

Basic Analysis of Autarky and Free Trade Models

UNITED STATES DEPARTMENT OF AGRICULTURE Washington, D.C ACTION BY: All Divisions and Offices. FGIS Directive 2510.

Self-Insurance In Times Of Growing And Retreating Managed Care. When enrollment in indemnity insurance declines, so does self-insurance.

Make the change today. Just have look at our loan calculator on page and see how much you could

Polynomial Functions. Polynomial functions in one variable can be written in expanded form as ( )

Total Data Protection for Sensitive Data Wherever It May Flow

Health Information Systems: evaluation and performance of a Help Desk

Lecture 3 Gaussian Probability Distribution

Participation and investment decisions in a retirement plan: the influence of colleagues choices


SPECIAL PRODUCTS AND FACTORIZATION

Threshold Population Levels for Rural Retail Businesses in North Dakota, 2000

Application Bundles & Data Plans

Experiment 6: Friction

Current Topics. Implications for Compliance Organizations. Presentation to the IIB. October, 2015

Navy Asbestos Medical Surveillance Program : Demographic Features and Trends in Abnormal Radiographic Findings

PEARLS MONITORING SYSTEM

Long Term Financial Planning

Commercial Cooling Rebate Application

The Economic Footprint of Michigan s Fifteen Public Universities

Transcription:

GAO United Sttes Generl Accounting Office Report to the Chirmn, Subcommittee on Oversight, Committee on Wys nd Mens, House of Representtives April 2001 IRS AUDIT RATES Rte for Individul Txpyers Hs Declined But Effect on Complince Is Unknown GAO-01-484

Contents Letter 1 Appendix I Individul Income Tx Audit Rte Trends for Fiscl Yers 1996 Through 2000 17 Appendix II Assessment of Resons Affecting Audit Rte Chnges for Fiscl Yers 1996 Through 2000 22 Appendix III Audit Rtes by Geogrphic Loction for Fiscl Yers 1996 Through 1999 23 Appendix IV Comments From the Internl Revenue Service 25 Appendix V GAO Contcts nd Stff Acknowledgments 27 Tbles Tble 1: Audit Rtes for Individuls 7 Tble 2: Audit Rtes for Individul Txpyers by Primry Income Source nd Income Levels 8 Tble 3: Individul Returns Audited by Type of Audit nd Auditor 9 Tble 4: Audit Stffing Levels for Individul Audits 10 Tble 5: Direct nd Nondirect Stff Yers nd Percent Chnge by Type of Auditor 11 Tble 6: Returns Filed nd Percent Chnge by Income Level 13 Tble 7: EIC Audits s Percent of Totl Audits of Lower Income Individuls 13 Tble 8: Audit Rtes for Individul Txpyers Overll nd by Income Levels 17 Tble 9: Individul Income Tx Returns Filed nd Percent Chnge by Income Level 18 Tble 10: Individul Income Tx Returns Audited nd Percent Chnge by Income Level 19 Tble 11: Combined Audit Clsses of Nonbusiness nd Business Schedule C Returns nd Percent Chnge 19 Pge i

Tble 12: Clcultion of Fiscl Yer 1999 Audit Rte for Combined Nonbusiness nd Business Returns for Lower Income nd Higher Income Individuls 20 Tble 13: Audits of Lower Income nd Higher Income Individuls by Type of Audit 21 Tble 14: Correspondence nd Fce-to-Fce Audit Rtes by Income Level 21 Tble 15: Direct Audit Hours nd Percent Chnge by Type of Auditor 22 Tble 16: Direct Audit Hours Per Audit nd Percent Chnge by Type of Auditor 22 Tble 17: Stff Yers Spent on Nonudit Activities nd Percent Chnge by Type of Auditor 22 Tble18: IRS Individul Audit Rtes by Region nd District Office 23 Tble 19: IRS Districts With the Highest nd Lowest Audit Rtes 24 Tble 20: Comprison of IRS Individul Audit Rtes, Including nd Excluding Service Center Audits 24 Figures Figure 1: Audit Rtes for Lower Income nd Higher Income Individuls 2 Figure 2: Audit Rtes for Lower Income nd Higher Income Nonbusiness nd Business Returns 3 Pge ii

United Sttes Generl Accounting Office Wshington, DC 20548 April 25, 2001 The Honorble Amo Houghton Chirmn, Subcommittee on Oversight Committee on Wys nd Mens House of Representtives Every yer, the Internl Revenue Service (IRS) performs vriety of complince checks to ensure the ccurcy of informtion reported by txpyers on their tx returns. These checks include verifying computtions on returns, mtching informtion reported by third prties to income reported by txpyers on returns, nd requesting more informtion bout items on tx return. Some of the checks were performed on virtully every one of the pproximtely 125 million returns filed in 1999; others were performed on smll subset of returns. IRS clssified bout 620,000 of these complince checks s udits, which rnge from contcts with txpyers tht re simple enough to be udited through the mil to fce-to-fce contcts over items tht re too complex to be udited through correspondence. 1 In recent yers, the udit rte the proportion of tx returns tht IRS udits ech yer hs drwn ttention. The ttention is due to long-term decline in udit rtes nd the difference in udit rtes for lower nd higher income individuls. The decline in udit rtes hs rised concerns bout whether this could led to decline in txpyers ccurtely reporting their tx libilities (i.e., their voluntry complince). The 1999 udit rte for lower income txpyers hs rised concerns bout whether txpyers re being selected for udits in n equitble mnner. Becuse of the ttention over udit rtes, you sked us to (1) describe the chnges in udit rtes for individul income tx returns overll nd for ctegories, such s mjor sources (i.e., nonbusiness versus business) 2 nd levels of income for fiscl yers 1996 through 2000; (2) obtin IRS resons 1 Under Internl Revenue Code Section 7602, IRS cn exmine the txpyer s books nd records nd solicit testimony from the txpyer nd relevnt prties regrding the ccurcy of tx return. 2 Nonbusiness sources include wges, dividends, nd interest. Business sources include self-employment income, such s income reported by individul sole proprietors on schedule C, which is used to compute profit or loss. Pge 1

nd relted dt explining the chnges in the udit rtes; nd (3) describe wht is known bout the effects of chnges in the udit rtes on tx complince. Results in Brief Compring fiscl yers 1996 nd 2000, the overll income tx udit rte of individuls declined bout 70 percent from 1.67 percent to 0.49 percent s shown in figure 1. Rtes declined regrdless of the individul txpyer s income level. Figure 1: Audit Rtes for Lower Income nd Higher Income Individuls 3.5 Audit rte (percent) 3 2.5 2 1.5 1 0.5 0 1996 1997 1998 1999 2000 Fiscl yer Higher income Lower income Overll udit rte Note 1: Lower income includes individuls reporting income less thn $25,000 nd higher income includes individuls reporting income of $100,000 or more. Note 2: The overll udit rte flls below the other two lines becuse it lso includes udits of other individuls, such s those reporting moderte income on their tx returns. Pge 2

In ny given yer, figure 1 lso shows greter udit rte for higher income individuls thn for lower income individuls. When the higher nd lower income groups re divided by mjor source of income, s shown in figure 2, some exceptions to this pttern re pprent. For individuls relying on nonbusiness income, those with lower incomes were udited t greter rte in fiscl yer 1999 thn those with higher incomes. For individuls relying on business income, those with lower incomes were udited t greter rte in fiscl yers 1996, 1999, nd 2000. Figure 2: Audit Rtes for Lower Income nd Higher Income Nonbusiness nd Business Returns 4.5 Audit rte (percent) 4 3.5 3 2.5 2 1.5 1 0.5 0 1996 1997 1998 1999 2000 Fiscl yer Nonbusiness-lower income Nonbusiness-higher income Business-lower income Business-higher income Note: Lower income includes individuls reporting income less thn $25,000 nd higher income includes individuls reporting income of $100,000 or more. Pge 3

Compring the type of udits nd the income level of the udited individuls, most udits were correspondence udits nd most lower income individuls were udited through correspondence. Most udited higher income individuls were udited through fce-to-fce udits. According to IRS officils, udit rtes declined for fiscl yers 1996 to 2000 for three min resons. First, over this period, the number of IRS uditors for individul returns declined by more thn hlf for resons such s decline in totl stff nd decisions to chnge stffing priorities to better serve txpyers before they file their returns. Second, IRS ws more likely to use the remining uditors in other duties, such s ssisting txpyers. Third, udits took longer due to dditionl udit requirements, such s more written communictions with txpyers bout the sttus of their udit. To explin the chnges in the udit rtes by income levels, IRS officils cited increses in the number of high-income tx returns nd n udit focus on noncomplince by erned income credit (EIC) climnts, who re usully lower income individuls. IRS rw dt were generlly consistent with IRS resons. We could not estblish the reltive influence of ech reson on the chnges in the udit rte. The specific effect of the recent decline in the udit rte on the level of voluntry complince is not known for severl resons. First, IRS does not hve current relible informtion on levels of voluntry complince. Even with this informtion, IRS would still need to tke number of steps to try to determine the specific link between complince nd udit rtes. Second, becuse nonudit nd udit progrms re designed to ensure tx complince, it is possible tht expnsion of IRS nonudit complince progrms could compenste to some degree for declining udit rtes. Third, improvements in ssisting nd educting txpyers bout their tx obligtions could similrly compenste to some degree for declining udit rtes. Becuse IRS does not hve relible, updted informtion on voluntry complince, we do not know the net effects on tx complince of the declining udit rtes, chnges in the volume of nonudit checks, nd ny improvements in IRS eductionl efforts. Bckground Annully, IRS udits some tx returns to determine whether txpyers complied with the tx lws. IRS ttempts to select returns for udit tht hve n indiction of potentil noncomplince bsed on, for exmple, its formul for flgging suspicious returns. IRS believes tht credible thret of being udited deters some noncomplince. IRS udits check complince in reporting income, deductions, nd other return items s well s in pying the correct tx libility. To conduct these complince Pge 4

checks, IRS uditors sk txpyers for documenttion bout specific items on their returns. IRS conducts two types of udits, fce-to-fce nd correspondence, using three clsses of uditors revenue gents, tx uditors, nd tx exminers. Fce-to-fce udits cn be either (1) field udits, in which n IRS revenue gent visits n individul who hs business income or complex return, or (2) office udits, in which n individul who hs less complex return visits tx uditor t n IRS office. Correspondence udits, s the nme suggests, re done by tx exminers who correspond with txpyers through the mil. Correspondence udits usully involve one line item on return. Becuse correspondence udits involve fewer nd usully simpler tx return items, they re less likely to burden txpyers in terms of time, contcts with IRS, nd documenttion provided to IRS. IRS lso checks complince nd contcts individul txpyers through nonudit enforcement progrms. For exmple, IRS mth error progrm checks returns for mth nd consistency errors nd contcts txpyers if such errors re found. IRS underreporter progrm mtches the income reported on tx returns with the informtion returns (e.g., W-2 forms) filed by third prties, such s employers who py wge income. If discrepncies re found, then txpyers re miled notice. Although such contcts cn be similr to correspondence udit contcts, IRS does not define them s udit contcts. 3 Over the yers, IRS hs shifted contcts between the udit nd nonudit ctegories. For exmple, in fiscl 1997, IRS shifted over 700,000 cses involving missing or invlid socil security numbers on tx returns from the correspondence udit progrm to the mth error progrm. Chnges in the definition of n udit could contribute to decreses or increses in the udit rte. Scope nd Methodology To describe chnges in udit rtes for individuls (s opposed to prtnership or corporte txpyers), we used IRS method for computing udit rtes. For ll txpyers nd by txpyer ctegories, the udit rte equls the proportion of IRS udits closed in fiscl yer compred to returns filed in the previous clendr yer. We used dt from IRS Dtbook, Audit Informtion Mngement System, nd Sttistics of 3 See Tx Administrtion: IRS Use of Nonudit Contcts (GAO/GGD-00-7, Mr. 16, 2000) for further discussion of nonudit contcts nd their reltionship to udit contcts. Pge 5

Income bout individul income tx returns filed in clendr yers 1995 through 1999 nd udits of the returns tht closed in fiscl yers 1996 through 2000. This llowed us to describe the chnges nd updte the udit rte trends in erlier reports. 4 We lso described udit rtes by vrious ctegories. One ctegory ws the income reported on individul income tx returns, which IRS divides into brod groups. Under IRS grouping, lower income individuls report income under $25,000 nd higher income individuls report $100,000 or more of income on their tx returns. Other ctegories included the types of IRS udit, IRS office loctions, nd the mjor income sources. Nonbusiness sources include individuls who generted most of their income from sources such s wges, dividends, nd interest. Business sources include individuls who generted most of their income from selfemployment nd reported tht income on schedule C (nonfrm income) or schedule F (frm income). For comprisons of lower nd higher income by source of income, we excluded schedule F income becuse IRS dt only split schedule F income into groups under nd over $100,000. We did include Schedule F txpyers in the overll udit rte. We interviewed officils from IRS Exmintion Division nd the Brookhven Service Center to discuss IRS resons for chnges in the udit rtes from fiscl yers 1996 through 2000. We lso obtined vilble IRS dt relted to the resons given by IRS officils. For exmple, we obtined IRS dt on chnges in the number of uditors nd number of hours spent doing udits. We checked for inconsistencies between the rw dt nd the resons tht IRS officils gve us. However, due to time constrints, we did not do ny more detiled nlyses to determine the extent to which IRS resons explined the chnges in udit rtes. Nor did we ttempt to identify resons beyond those offered by IRS. To describe wht is known bout the potentil effects of chnges in the udit rtes on tx complince, we used our previous nd ongoing work on IRS udits, other IRS enforcement progrms, nd tx complince. We lso used informtion from our discussions with IRS officils. 4 See Tx Administrtion: Audit Trends nd Results for Individul Txpyers (GAO/GGD-96-91, Apr. 26, 1996), which discussed udit rtes for individuls during fiscl yers 1988-95. We issued report (Tx Administrtion: IRS Audit nd Criminl Enforcement Rtes for Individul Txpyers Across the Country, GAO/GGD-99-19, Dec. 23, 1998) on geogrphic udit rtes. Appendix III updtes such udit rtes for fiscl yers 1996-99. Pge 6

We did our work t IRS ntionl office in Wshington, D.C., between September 2000 nd Mrch 2001 in ccordnce with generlly ccepted government uditing stndrds. We requested comments on drft of this report from the Commissioner of Internl Revenue. We received written comments from the Commissioner in letter dted April 19, 2001. The comments re reprinted in ppendix IV nd discussed t the end of this letter. Individul Audit Rtes Hve Been Declining Overll nd Across Vrious Ctegories From fiscl yers 1996 through 2000, the overll income tx udit rte of individuls declined. As tble 1 shows, IRS nnul udit rtes for individuls declined from 1.67 percent to 0.49 percent bout 70 percent. The tble lso shows tht the udit rtes fell for ll mjor sources of income nonbusiness s well s schedule C nd schedule F business returns over the 5 yers. Tble 1: Audit Rtes for Individuls Returns nd udits in thousnds Individul txpyers 1996 1997 1998 1999 2000 Percent chnge Individul income tx returns 116,059.7 118,362.6 120,342.4 122,546.9 124,887.1 8 Individul udits 1,941.6 1,519.2 1,192.8 1,100.3 617.8-68 Audit rtes (percent) 1.67 1.28 0.99 0.90 0.49-70 Nonbusiness 1.54 1.15 0.88 0.82 0.42-73 Schedule C 3.60 3.15 2.35 2.03 1.55-57 Schedule F 2.29 1.82 1.21 0.90 0.54-77 Tble 1 lso shows tht the udit rte ptterns for ech yer chnged little from fiscl yers 1996 to 2000. Schedule C business returns were more thn twice s likely to be udited thn the nonbusiness returns in ech yer. Tble 2 shows tht udit rtes declined bout eqully 67 percent nd 70 percent, respectively for lower nd higher income individuls. When txpyers re seprted into nonbusiness nd business income groups, udit rtes declined t lest 42 percent from fiscl yers 1996 to 2000 for lower nd higher income individuls in the two groups. Pge 7

Tble 2: Audit Rtes for Individul Txpyers by Primry Income Source nd Income Level Primry income source nd income level 1996 1997 1998 1999 2000 Percent chnge Nonbusiness nd schedule C business incomes Lower income 1.91% 1.46% 1.12% 1.24% 0.63% -67 Higher income 3.19 2.74 2.03 1.40 0.96-70 Nonbusiness income Lower income 1.82 1.39 1.06 1.18 0.55-70 Higher income 2.85 2.27 1.66 1.14 0.84-70 Schedule C business income Lower income 4.21 3.19 2.37 2.69 2.43-42 Higher income 4.09 4.13 3.25 2.40 1.48-64 Note: See tble 12 for n explntion of how udit rtes were clculted for fiscl yer 1999. Lower income includes individuls reporting income less thn $25,000 nd higher income includes individuls reporting income of $100,000 or more. Tble 2 lso shows tht higher income individuls were more likely to be udited thn lower income individuls in ech of the 5 yers. However, exceptions to this pttern emerged when these udit rtes by income level were nlyzed by source of income. First, in the nonbusiness group, IRS ws more likely to udit lower income individuls only in fiscl yer 1999. Second, in the business group (schedule C), the rtes fluctuted by income levels. IRS ws more likely to udit lower income individuls in fiscl yer 1996, higher income individuls in fiscl yers 1997 nd 1998, nd lower income individuls in fiscl yers 1999 nd 2000. Most udits of lower income individuls were correspondence udits, with the proportion of udits of lower income individuls tht were correspondence udits rnging from 69 to 84 percent over the 5 yers. Audits of higher income individuls were mostly fce-to-fce udits, rnging from 62 to 75 percent over the 5 yers. (See tble 13 in pp. I for detils.) Correspondence nd fce-to-fce udit rtes lso vried by txpyer income. For exmple, in fiscl yer 2000, the fce-to-fce udit rte (fceto-fce udits divided by ll returns filed) for higher income individuls ws 0.60 percent compred with 0.13 percent for lower income individuls. For correspondence udits in fiscl yer 2000, the udit rte for higher income individuls ws 0.37 percent nd for lower income individuls ws 0.50 percent. (See tble 14 in pp. I for detils.) Pge 8

Tble 3 shows tht both types of fce-to-fce udits (field nd office) nd correspondence udits declined by similr rtes from 1996 to 2000. Tble 3 lso shows tht correspondence/tx exminer udits ccounted for over hlf of ll udits in ech yer (rnging from 54 percent to 67 percent) nd tht the number of udits declined ech yer for ll types of udits/uditors except for correspondence udits in fiscl yer 1999. Tble 3: Individul Returns Audited by Type of Audit nd Auditor Individul returns udited Audit/uditor type 1996 1997 1998 1999 2000 Percent chnge Field/revenue gents 252,430 209,781 168,054 124,518 91,586-64 Office/tx uditors 509,420 505,834 383,366 235,625 145,975-71 Correspondence/tx exminers 1,179,696 803,628 641,360 740,130 380,204-68 Totl 1,941,546 1,519,243 1,192,780 1,100,273 617,765-68 Note: With very few exceptions, revenue gents do field udits, tx uditors do office udits, nd tx exminers do correspondence udits. Tx exminer totls for fiscl yers 1998, 1999, nd 2000 lso include service center totls from IRS dt. The declines in udit rtes were spred uniformly cross IRS four regions. However, udit rtes vried by region. The udit rtes declined bout 50 percent in ech of IRS four regionl offices from fiscl yers 1996 to 1999. 5 For ech of these 4 yers, the rnge of udit rtes ws highest in the Western Region (1.09 to 0.47 percent) compred to the Northest Region (0.44 to 0.23 percent), the Southest Region (0.58 to 0.28), nd the Midsttes Region (0.62 to 0.32 percent). (See tbles 18, 19, nd 20 in pp. III.) IRS Officils Cited Severl Resons for the Chnges in Individul Audit Rtes According to IRS officils, overll udit rtes declined for fiscl yers 1996 to 2000 for three min resons. First, IRS hd fewer uditors for individul returns for resons tht include decline in stff nd decisions to chnge stffing priorities to focus on customer service. Second, IRS ws more likely to use the remining uditors in other duties, such s ssisting txpyers. Third, udits took longer due to dditionl requirements, such s more written communictions with txpyers bout the sttus of their udit. With respect to chnges in the udit rte by income levels, IRS 5 Fiscl yer 1999 return filings by geogrphicl loction were not vilble for us to compute the relted udit rtes for fiscl yer 2000. Pge 9

officils cited n increse in the number of high-income tx returns nd n udit focus on noncomplince by erned income credit climnts, who re usully lower income individuls. IRS rw dt were generlly consistent with ll these resons. However, due to time constrints, we did not nlyze the dt to determine the extent to which IRS resons explined the chnges in udit rtes. IRS Auditor Stffing Levels Declined According to IRS officils, IRS did fewer udits between fiscl yers 1996 nd 2000, in prt becuse it hd fewer uditors. IRS officils explined tht uditor stff levels declined for two resons. First, tight budgets in the 1990s reduced overll stffing levels. Second, IRS put more stff in positions to serve txpyers nd generlly hs not hired revenue gents or tx uditors since 1995. As shown in tble 4, the number of revenue gent nd tx uditor positions ssigned to udit individul income tx returns declined stedily since 1996. By fiscl yer 2000, the number of these positions declined bout 54 percent for revenue gents nd bout 61 percent for tx uditors. This represents loss of over 2,000 stff yers for udit stff devoted to field nd office udits. On the other hnd, tx exminer positions, which do the simpler correspondence udits, incresed 13 percent, or 200 positions, between fiscl yers 1997 nd 2000 (dt for fiscl yer 1996 were not vilble). Tble 4: Audit Stffing Levels for Individul Audits Fiscl yer Chnge Type of uditor 1996 1997 1998 1999 2000 Number Percent Revenue gents 2,441 2,121 1,751 1,407 1,116-1,325-54 Tx uditors 1,173 1,045 797 621 461-712 -61 Tx exminers 1,515 b 1,772 b 1,740 1,715 200 c 13 c Totl 4,681 4,320 3,768 3,292 IRS ws not ble to provide tx exminer dt for fiscl yer 1996. b Tx exminer totls for fiscl yers 1997 nd 1998 re IRS estimtes. c Only covers chnge from fiscl yer 1997 to 2000. Auditors Spent More Time on Other Activities IRS officils lso sid tht its uditors spent less time uditing in fiscl yers 1996 through 2000. Our nlysis of IRS dt, s shown in tble 5, indictes tht for individul income tx returns, the verge mount of direct udit time ctul time doing udit work hs declined in Pge 10

comprison to time spent on nondirect udit ctivities. Nondirect udit ctivities include txpyer ssistnce, other detils, nd trining. Tble 5: Direct nd Nondirect Stff Yers nd Percent Chnge by Type of Auditor Type of uditor 1996 1997 1998 1999 2000 Percent chnge Revenue gents Direct udit 7,924 7,765 6,840 5,903 5,265-34 Nondirect udit 7,159 6,634 6,807 7,158 7,285 2 Totl stff yers 15,083 14,399 13,647 13,061 12,550-17 Direct udit percent 0.53 0.54 0.50 0.45 0.42-20 Tx uditors Direct udit 1,280 1,164 906 725 566-56 Nondirect udit 1,205 1,154 1,207 1,204 1,136-6 Totl stff yers 2,485 2,318 2,113 1,930 1,702-32 Direct udit percent 0.51 0.50 0.43 0.38 0.33-35 Prt of the reson for the decline in uditing is tht revenue gents nd tx uditors spent incresingly more time providing txpyer ssistnce between fiscl yers 1996 nd 2000. The mount of time spent on txpyer ssistnce by revenue gents incresed from bout 1.0 percent of vilble stff yers in 1996 to bout 4.4 percent of vilble stff yers in 2000. The mount of time spent on txpyer ssistnce by tx uditors incresed from bout 1.4 percent of vilble stff yers in 1996 to bout 12.3 percent of vilble stff yers in 2000. IRS did not hve comprble dt for ssistnce provided by tx exminers who hd been slted to do udits. In ddition, revenue gents nd tx uditors hd less time to udit becuse of incresed time in trining. (See tble 17 in pp. II for dditionl informtion on revenue gent nd tx uditor trining.) Considering the 54-percent decrese in the number of revenue gents, the trining time per revenue gent incresed bout 227 percent. The trining time per tx uditor over the sme 5 yers incresed bout 95 percent. IRS did not hve comprble trining dt on tx exminers. IRS Auditors Spent More Time to Finish Ech Audit Finlly, IRS officils sid tht uditors generlly took longer to finish udits during fiscl yers 1996 to 2000. Our nlysis of IRS dt for this period (see tble 16 in pp. II) showed tht the verge time to finish n udit incresed for ll types of uditors, including bout Pge 11

37 percent (20.2 hours to 27.6 hours) for revenue gents (field udits), 56 percent (4.6 hours to 7.1 hours) for tx uditors (office udits), nd 153 percent (0.7 hours to 1.8 hours) for tx exminers (correspondence udits). IRS officils told us tht Internl Revenue Service Restructuring nd Reform Act of 1998 requirements incresed udit time. 6 Among other things, these requirements resulted in IRS uditors hving to send more notices to txpyers nd third prties tht provide informtion bout the txpyer being udited. New requirements to explin innocent spouse provisions nd to protect txpyers under udit hve generted more review work. These officils sid the ct hs creted mny new tsks during udits. Other fctors, such s the experience level of the uditor nd complexity of the udit, lso ffected udit time per return. For exmple, IRS officils sid tht they lost mny experienced uditors to higher grded positions elsewhere in IRS. Becuse multiple fctors ffect udit time per return, determining the contribution of ech fctor to chnges in udit time could be difficult. Becuse of time constrints, we did not ttempt such n nlysis. IRS Cited Other Resons for the Vrying Individul Audit Rtes by Income Levels IRS officils offered two resons why the udit rtes for lower income individuls exceeded the rtes for higher income individuls in selected yers mong the nonbusiness nd business groups. First, s tble 6 shows, the number of higher income returns filed in clendr yers 1995 through 1999 tht were subject to udits in fiscl yers 1996 to 2000 significntly incresed compred with the number of lower income returns filed. For nonbusiness returns, the number of higher income returns filed rose 80 percent compred with 5-percent decrese for lower income returns filed. For business returns, the number of higher income individul returns incresed bout three times the rte of lower income business returns. 6 P.L. 105-206. Pge 12

Tble 6: Returns Filed nd Percent Chnge by Income Level Returns filed in clendr yer Income level 1995 1996 1997 1998 1999 Percent chnge Nonbusiness returns Lower income 59,211,700 58,790,700 58,266,600 57,432,900 56,247,800-5 Higher income 4,540,800 5,260,500 6,044,700 7,025,000 8,151,600 80 Schedule C business returns Lower income 2,436,300 2,464,700 2,530,100 2,546,800 2,541,000 4 Higher income 1,738,300 1,770,700 1,835,500 1,876,000 1,948,900 12 Note: Returns filed in clendr yers1995 through 1999 re used to compute the udit rtes for fiscl yers 1996 through 2000. Lower income includes individuls reporting income less thn $25,000 nd higher income includes individuls reporting income of $100,000 or more. Second, IRS udits in fiscl yers 1997 through 2000 hve continued to focus on EIC noncomplince, usully by lower income individuls. 7 As tble 7 shows, EIC udits, usully correspondence udits, ccounted for lrge percent of the udits of lower income txpyers, regrdless of their mjor source of income. In fct, the EIC portion of ll udits for lower income txpyers in fiscl yer 2000 ws more thn double the fiscl yer 1997 EIC portion of these udits. Tble 7: EIC Audits s Percent of Totl Audits of Lower Income Individuls Fiscl yer Audit type 1997 1998 1999 2000 Totl udits of lower income individuls 895,625 679,088 745,614 369,912 Service center EIC udits b 360,101 290,010 572,594 325,654 EIC s percent of totl udits 40% 43% 77% 88% Note: IRS ws unble to provide EIC dt for fiscl yer 1996. Excludes Schedule F filers becuse IRS dt do not show how mny of these filers reported less thn $25,000. b District office udits might cover EIC in prt, but IRS does not uniformly cpture tht informtion. 7 Congress encted EIC in 1975, s refundble tx credit vilble to lower income, working txpyers, to offset the effect of socil security txes on lower income fmilies nd to encourge these txpyers to seek employment rther thn welfre. Pge 13

IRS officils lso sid tht project to ddress noncomplince by schedule C filers who climed EIC explined the greter udit rtes for lower income business filers compred with those with higher incomes during fiscl yers 1999 nd 2000. Effect of the Recent Declining Audit Rte on Tx Complince Levels Is Not Known The specific effect of the recent decline in the udit rte on the level of voluntry complince is not known. One reson is tht IRS does not hve current relible informtion on the levels of voluntry complince. IRS lst mesured overll income tx complince for tx yer 1988. IRS nd others re concerned tht chnges in the tx lws, economy, nd demogrphics since 1988 hve mde the complince informtion out of dte. Even if IRS hd this informtion, IRS would still need to tke number of steps to try to determine the specific link between chnges in udits nd chnges in voluntry complince levels. Historiclly, mesuring the specific impct of udit rte chnges on voluntry complince hs been difficult. It is difficult to collect dt on nonudit fctors tht lso cn ffect voluntry complince levels, nd then to control for these fctors in order to isolte the impct of udit rte chnges. For exmple, it is difficult to determine the effect of declining udit rtes on voluntry complince when IRS nonudit checks could offset to some degree ny negtive effects of declining udit rtes on complince. Since the 1970s, for exmple, the underreporter progrm hs grown, covering more types of income especilly mong nonbusiness txpyers. IRS lso uses the mth error progrm to help ensure txpyer complince. Since the mth error nd underreporter checks cn be similr to correspondence udits, growth in these progrms my offset to some degree the decline in the udit rte. Furthermore, it hs lso been difficult to mesure how improvements in ssisting nd educting txpyers bout their tx obligtions compenste for declining rtes. These IRS efforts, lthough not designed to find noncomplince, could help txpyers to voluntrily comply. To the extent tht eduction efforts succeed in promoting complince, overll complince would not necessrily decline if the udit rte declines. IRS hs been llocting more resources to txpyer ssistnce nd eduction. One exmple is the incresed use of revenue gents nd tx uditors to provide txpyer ssistnce. Becuse IRS does not hve mesure of voluntry complince, we do not know the net effects on tx complince levels of the declining udit rtes, Pge 14

chnges in the volume of nonudit checks, nd ny improvements in IRS eductionl efforts. Agency Comments On April 19, 2001, we received written comments on drft of this report from the Commissioner of Internl Revenue (see pp. IV.). The Commissioner sid tht IRS grees with our presenttion nd nlysis of the udit rte dt s well s with the need for current nd relible dt on voluntry complince. The Commissioner greed tht chnges in the economy nd tx lws hve rendered IRS complince dt obsolete. The Commissioner s comments lso expnded on wht IRS officils told us during our work bout the resons for the udit rte decline. Specificlly, the comments sid tht two provisions (sections 1203 nd 1204) of the Restructuring nd Reform Act of 1998 creted cutionry environment tht led to udits tking longer. 8 The Commissioner lso sid tht the report did not cknowledge historicl dt in two studies--one by IRS nd one by externl reserchers--on the effects of the decline in udit rtes on voluntry complince. 9 We did not cknowledge these studies in the report becuse, while they estimte reltionship between udits nd complince, they re not bsed on current dt. The most recent of the studies used dt from 1982 through 1991. Becuse of the possibility tht chnges over time in the economy, tx lws, demogrphics, nd IRS complince progrms hve chnged the reltionship between udit rtes nd voluntry complince, we did not cite the studies. These two studies did report positive reltionship between udit rtes nd voluntry complince. This finding is consistent with the concern, which we describe in the report, tht decline in udit rtes could led to decline in voluntry complince. As rrnged with your office, unless you publicly nnounce its contents erlier, we pln no further distribution of this letter until 30 dys from its 8 Section 1203 requires the Commissioner of Internl Revenue to terminte ny IRS employee tht ws proven to hve committed certin violtions in connection with the performnce of officil duties. Section 1204 prohibits the use of enforcement sttistics in imposing or suggesting production quots or in evluting employee performnce. 9 Upon checking with IRS officils, we found out tht these two studies re Dubin, Gretz, nd Wilde, The Effect of Audit Rtes on the Federl Income Tx, 1977-1986, Ntionl Tx Journl (1990), nd The Determinnts of Individul Income Tx Complince: Estimting the Impcts of Tx Policy, Enforcement, nd IRS Responsiveness (IRS Pub. 1916, 1996). Pge 15

dte of issue. We will then send copies to Representtive Willim M. Thoms, Chirmn, nd Representtive Chrles B. Rngel, Rnking Minority Member, House Committee on Wys nd Mens; Representtive Willim J. Coyne, Rnking Minority Member, Subcommittee on Oversight, House Committee on Wys nd Mens, nd Sentor Chrles E. Grssley, Chirmn, nd Sentor Mx Bucus, Rnking Minority Member, Sente Committee on Finnce. We will lso send copies to the Honorble Pul H. O Neill, Secretry of the Tresury; the Honorble Chrles O. Rossotti, Commissioner of Internl Revenue; the Honorble Mitchell E. Dniels, Jr., Director, Office of Mngement nd Budget; nd other interested prties. Copies of this report will be mde vilble to others upon request. If you hve ny questions, plese contct me or Tom Short t (202) 512-9110. Key contributors to this report re cknowledged in ppendix V. Jmes R. White Director, Tx Issues Pge 16

Appendix I: Tx Audit Rte Appendix I: Individul Income Tx Audit Rte Trends for Fiscl Yers 1996 Through 2000 Trends for Fiscl Yers 1996 Through 2000 Tble 8: Audit Rtes for Individul Txpyers Overll nd by Income Level Fiscl yer udit rte Income level 1996 1997 1998 1999 2000 Percent chnge Overll udit rte for individuls 1.67% 1.28% 0.99% 0.90% 0.49% -70% Nonbusiness returns TPI under $25,000 1.82 1.39 1.06 1.18 0.55-70 TPI $25,000 under $100,000 1.03 0.73 0.60 0.36 0.22-78 TPI $100,000 nd over 2.85 2.27 1.66 1.14 0.84-70 Subtotl 1.54 1.15 0.88 0.82 0.42-73 Business returns Schedule C TGR under $25,000 4.21 3.19 2.37 2.69 2.43-42 TGR $25,000 under $100,000 2.85 2.57 1.82 1.30 0.93-67 TGR $100,000 nd over 4.09 4.13 3.25 2.40 1.48-64 Subtotl 3.60 3.15 2.35 2.03 1.55-57 Schedule F TGR under $100,000 1.59 1.28 0.93 0.68 0.35-78 TGR $100,000 nd over 3.61 2.75 1.63 1.23 0.80-78 Subtotl 2.29 1.82 1.21 0.90 0.54-77 Legend TPI = totl positive income (income from positive sources only) Schedule C-TGR = totl gross receipts (profit or loss from business) Schedule F-TGR = totl gross receipts (profit or loss from frming) Note: We combined two of IRS income levels ($25,000 to $50,000 nd $50,000 to $100,000) into one income group ($25,000 to $100,000) becuse their udit rtes were similr. Pge 17