Odds of Audit in Your Community:
IRS Removes Essential Tracking Data from Public View
Sweeping organizational changes in the IRS are undermining the ability of the American people to track the way the nation’s largest and most powerful enforcement agency is handling taxpayers, according to an analysis of IRS data.
For the first time in recent history, the IRS refused to provide the geographic location of many corporate and selected individual tax audits conducted last year, and has announced that this year it has completely discontinued collecting information required to assess the odds of audit that taxpayers face in different parts of the country.
Such data have in the past highlighted serious administrative failures in assuring that similarly situated taxpayers in different parts of the country are treated in similar ways. For example:
Corporations: Three out of four of the largest corporations in northern New York were audited in FY 2000. During the same year in Delaware and Maryland, however, only one out of ten of the largest corporations were targeted. Even more surprising, the additional taxes recommended for the heavily-audited corporations in northern New York was only $390,000 per return. This compared with an average additional tax of $28 million per return recommended for the corporations in Maryland and Delaware. (See table.)
Individuals: The district audit rate for individual taxpayers earning $100,000 or more in Los Angeles, North Dakota, South Dakota, and Minnesota was seven times the rate in FY 2000 for those living in Georgia, Delaware, Maryland and Virginia. (See table.)
While the identification of these kinds of discrepancies was possible in fiscal year 2000, changes in IRS data collection practices have made it impossible in FY 2001. In 1992, geographic identifiers were collected regarding more than 99% of the additional taxes that the IRS said were owed by corporations. In 2001, such information was collected for only 2% of these taxes. For individuals, there has been a continuing decline in the completeness of audit information over the last decade. In 1992, geographic information was collected about 82% of the additional taxes the IRS claimed were owed by individual taxpayers. In 2001, the information was collected for 39%. Next year, the agency will collect no information about the geographic location of taxpayers it is targeting. (See graph and table.)
Up until last year, it thus was possible to ask whether the poor allocation of resources and other administrative failings by the IRS had contributed to a tax system in which corporations and individuals located in different parts of the nation were effectively living under different tax laws. Now, the absence of solid information comparing the district-to-district or state-by-state way that taxpayers are being treated makes it impossible for both agency managers and the American people who pay their salaries to judge whether this powerful agency is functioning in an effective and equitable way. Expressed another way, to see whether the constitution's promise of equal treatment under the law is being kept.
Fundamental Problems in IRS Computer Management Systems
The loss of this essential information is symptomatic of more wide-spread problems in the ability of IRS computer systems to provide agency managers the data they need to assure the effective operation of the agency. The key challenge that IRS Commissioner Charles O. Rossotti was brought in to deal with was the IRS’s massive collection of obsolete computer systems. As a successful businessman with extensive experience in managing large systems, his appointment broke with the long tradition of naming tax attorneys to head one of the largest systems of them all.
Since becoming IRS Commissioner in 1997, Mr. Rossotti has repeatedly promised that by spending hundreds of millions of dollars on new computer systems and a top-to-bottom reorganization of the agency that the IRS would soon offer the same kind of service that the major credit card companies and the big package shipping companies now provide their customers. Unlike their counterparts in the private sector, however, the IRS data systems are not designed to give IRS managers the information they need to assess how they are treating their customers -- the American taxpayers.
In customer-based systems, information is collected about how the customers are being treated by the organization. But that is not how IRS management data systems are designed. Instead, they continue to focus inward -- upon the IRS organizational unit and its characteristics. (See note.) When, as was done in the past, the basic organizational units were handling all the taxpayers within their boundaries, the problem was not acute. But with the increasing centralization of many enforcement functions into fewer and fewer compliance and call centers, as well as the reorganization of the IRS along customer -- rather than geographic -- lines, these old ways of handling things no longer serve the agency’s goals.
For example, under the new system, almost all audits completed during fiscal year 2001 by IRS’s new Large and Mid-Size Business Division were recorded as taking place in Brooklyn. Along with the audited corporations were audits of individuals -- even a few low income wage-earners -- who also had been examined by that division. This accounting failure occurred because the IRS’s management reports that were furnished TRAC did not record where the corporation or individual taxpayer was located, or even where the audit took place, but the location of the IRS office that processed the forms. Furthermore, the audit of a corporation is always categorized under the industry unit where the IRS employee conducting the audit is assigned when the case is closed. This means, for example, that if an IRS employee conducting an audit of a natural resources corporation is transferred to the division handling financial companies, the audit of the natural resources corporation will end up being improperly categorized on the IRS books. For these and a variety of other reasons, including the interconnectedness of financial matters in a wider world that does not fit neatly into the confines of the IRS’s organizational sections, the actual function of the corporation may differ from the assigned mission of the agency unit doing the audit.
The profound shortcomings of the agency’s current information management system were dramatically illustrated recently when the IRS was forced to retract basic data the agency had provided TRAC that it initially had argued would offer the American people a new kind of accountability. The data withdrawn by the agency allegedly were simple counts of the number of audits it had made of the largest companies involved in five major segments of the economy -- communications, retailing, heavy manufacturing, natural resources, and financial and health care services. When TRAC raised questions about the resulting audit rates for each of the five business areas, the IRS acknowledged that the data were not correct and that under the new reorganization such counts could not now be determined. (See note.)
In the wake of the collapse of Enron and several other very large businesses in the last few years, the IRS’s inability to provide the public with accurate segment-by-segment data about how it is performing its basic responsibilities toward the largest corporations -- including those involved in natural resources -- is not reassuring.
IRS Goes After Fewer and Fewer Wealthy Americans
The IRS has long looked favorably on what it calls Service Center (correspondence) audits. This is because they are relatively inexpensive to conduct and bring in a lot of revenue. In a time of limited IRS staffing, the relative importance of the program has grown. Last year, for example, 40% of all additional taxes through audit claimed by the IRS from individual taxpayers was the result of correspondence as distinguish from face-to-face (district) audits. (See graph and table.)
In the last ten years, the basic nature of this Service Center program has undergone substantial change as the number of wealthier taxpayers selected for a correspondence audit crumbled. In 1992, high income taxpayers had the highest correspondence audit rates. Today, they are below average. The IRS has drastically cut back on the number of these cost-effective audits directed at individuals reporting $100,000 or more although they are the most productive in locating tax underreporting. While there has been a major increase in the number of such taxpayers, IRS audits of this group went from 74,566 in 1992, to 29,086 in 2001. The decline in high income audits occurred even though the highest level of tax underreporting was uncovered among this group. (See graph and table.)
A similar dramatic also occurred for high income business (Schedule C) filers reporting gross receipts of $100,000 or more. In 1992, such high income business taxpayers had the second highest correspondence audit rates. Today, they are well below average. IRS audits of these high income business filers went from 12,743 in 1992, to a mere 5,556 in 2001 -- yet these were second only to high income taxpayers in turning up high levels of tax underreporting. (See graph and table.)
Audit Rates for Rich Again Fall Below Rates for Poor
Audit rates for all individuals and business returns have fallen sharply over the last decade. In 1992 slightly more than one percent (1.06%) of individual returns were audited by the IRS. By 2001, the figure was slightly over half of one percent (0.58%) -- a slight rebound over the rate in 2000 (0.49%) because of a jump in correspondence audits. (See graph.)
However, largely as a result of the dramatic change in targeting of correspondence audits, the odds of audit for high income taxpayers was in fact less than those of low income taxpayers -- those with incomes less than $25,000 filing a simple return. And the odds of audit of high income business (Schedule C) files was less than that for low income businesses with less than $25,000 of gross receipts.
Other Enforcement Changes
In the weeks immediately proceeding April 15, IRS announced several new enforcement initiatives that may have led some taxpayers to conclude that the agency was on the rebound from a multi-year period of shrinking tax enforcement activities. But the national-level enforcement data still collected by the agency indicate that for most categories of taxpayers this assumption would be wrong.
While a few areas show a small and very recent increase in IRS enforcement actions, a ten-year comparison presents a picture of an agency in retreat.
First, considering the 1992 and 2001 period, here is the changing picture concerning the chances that a taxpayer would face four basic kinds of IRS administrative actions. In 1992, 1 out of 6 taxpayers was the subject of an IRS penalty. In 2001, 1 in 7. Concerning the chances of receiving an information matching notice, the odds went from 1 in 21 in 1992 to 1 in 50 in 2001. For most Americans, the fear of an IRS audit has always loomed large. Here, for all income groupings, the odds went from 1 in 94 in 1992 to 1 in 174 in 2001. Then, for taxpayers who owe the IRS some money, the IRS has the authority to impose its will through liens, levies or seizures. In 1992, 1 out of 24 taxpayers faced these sanctions. By 2001, only 1 in 145.
(See chart and table.)
In terms of actual counts, in 1992 there were 4,700,000 levies, seizures or liens. In 2001, there were only 875,000 such actions. (See graph and table.)
In addition to its civil enforcement authority, the IRS can asked the Justice Department to bring criminal charges against knowing taxpayers. Because such charges can result in prison sentences, they in some ways represent the agency’s ultimate weapon. For that reason, the odds of IRS criminal actions are much longer than the odds of administrative actions. But once again, the data show a declining IRS presence. According to information collected by the Justice Department, the IRS recommended the prosecution of 1 in 21,441 taxpayers in 1992. By 2001, only one taxpayer in 61,818 was subject to such an IRS recommendation. Actual prosecutions also fell -- 1 in 41,513 in 1992, 1 in 100,952 in 2001. (See chart and table.)
The Commissioner’s View. In a recent statement about the agency’s enforcement actions, Commissioner Rossotti contended that while the agency’s in-person audit rate was still declining, that the IRS during the last year had begun “to stabilize compliance activities.” He acknowledged that although the agency still had more work to do, that the FY 2001 audit effort represented “a good step toward a tax system where everyone pays a fair share.”
Several years ago, Congress ordered the IRS to abandon a special scientific program designed to measure the extent to which American taxpayers complied with the law. Partly because of this decision, there today is no good evidence concerning the impact of the dramatic collapse of IRS enforcement on how well taxpayers are meeting their lawful obligations. And because of a variety of special government powers such as its right to withhold payroll taxes, the importance of enforcement is not all that clear. But experience in other areas suggest that in the long run there well may be problems. When police officers in a congested city cut back on issuing parking tickets, double parking and other forms of illegal parking are almost certain to increase.
The IRS in recent years has paid more attention to taxpayer services. Although consistent ten-year data are not available for this area, IRS staff years devoted to such subjects as toll-free phones, walk-in assistance and adjustments and taxpayer relations have inched up from 13,169 in 1997 and 15,976 in 2001.
Hits on the IRS web site, for example, went from 102 million in 1996 to 2.6 billion in 2001. Also sharply up were the counts on downloaded forms and instructions -- from 3.8 million in 1996 to 317 million in 2001. (See graph.) Perhaps partly because of the availability of these web-based services, the total number of toll-free telephone calls IRS answered were down. Calls answered by IRS peaked at 74 million in 1997 and had fallen to only 51 million last year. While the IRS said that the percent of callers who were successful in their attempts to reach an assistor has improved markedly since 1995, last year this service indicator slipped from 61% in 2000 to 54% in 2001. (See graph and table.)
Overall, the number of IRS employees is down 12 percent since 1992.
The total, however, has begun to inch back up. There were 114,819 in 1992, 97,375 in 1998 and 100,577 in 2001. The decline in permanent employees, however, has been more precipitous -- 113,028 in 1992, 83,007 in 2001 -- a drop of 27 percent. To partly make up for this loss, the IRS has returned to filling in the blanks with seasonal and temporary employees. They jumped from 1,791 in 1992 to 17,570 in 2001. (See graph and table.)