The IRS district audit -- the face-to-face meeting between a taxpayer and an auditor or revenue agent -- is a mythic event in American life. For the IRS, the public perception of the agency's audit program has long been thought of as a critical force in persuading taxpayers to comply with the nation's tax laws. For the politicians, especially in recent years, audits have been held up as an example of how the federal government sometimes acts in abusive ways. (Not many years ago some of the same politicians were regularly voting to increase IRS audit budgets and sanction authority.) For comedians, the audit has always been a good source for jokes and skits, especially around April 15.

Despite its important place in the public mind, the public knows surprisingly little about IRS audit practices.

• How frequently do they actually occur?
• What are the odds of audit for different income groups?
• How long does the average audit take to complete?
• At the end of the audits, what percent of them lead the IRS to claim the taxpayer owes additional taxes and penalties?
• Considered by various income groups, how much does the IRS claim the average taxpayer owes the government?
• How does the performance of the IRS revenue agents compare with that of the auditors? (Because revenue agents have more experience and training, they usually work on larger and more complicated matters.)
• Looked at according to the IRS's 33 districts, is the IRS audit program effective and fair? Are similarly situated taxpayers being treated in similar ways?
• One of the great and continuing mysteries in American life is the reason for the audit --how come the IRS decided to pick on me?

Now, agency data tapes can provide some general answers to many of these questions.The IRS in 1998 conducted more than face-to-face audits of individuals and corporations. The IRS divides federal tax returns into a variety of different classes -- by type of return filed, by the income/or business receipts recorded, etc. It then assigns revenue agents to examine the more complicated returns and tax auditors for the less complicated returns. Under agency procedures, the IRS collects detailed information about each transaction.

Consider the question why was I audited? For many years the IRS has emphasized that a significant proportion of audits were initiated as a result of a selection process controlled by a computerized mathematical formula called "DIF." The agency's emphasis on DIF was comforting to the public because it implied a neutral, scientific, fact-based tax selection process that was not influenced by political factors or the whimsy of an individual employee. Data tapes from the IRS, however, indicate that DIF is playing a diminishing role in the audit process. The decline is especially noticeable for individual and taxpayers. In 1992, DIF was cited as the initiating factor in just under half the audits (46.4%).In 1998, slightly more than one quarter of individual district audits (28.8%) were triggered for this reason. For corporations, except for a dip in 1994 and 1995, the use of DIF has been more consistent -- 34.6% in 1992, compared with 29.6 percent in 1998.

According to the IRS, the decision of an individual not to file any tax return at all was also a significant reason for audit. In 1998, this factor was cited as the trigger in 12% of the face-to- face audits. (This area has been the subject of special IRS attention. In 1995, for example, a refusal to file prompted 39% of all audits.)

Special IRS research projects and studies prompted slightly more than one in ten of the audits of individuals. In a special project, the IRS selects certain kinds of returns or economic groups that it believes may present unusual compliance problems. A random sample of the individuals who have been identified as falling into these groups are then subject to audit. One such special project looked at individuals working in restaurants and bars in the Cleveland area. Another focused on those inidividuals seeking benefits under the Earned Income Credit program.

The IRS appears to have some concerns about the work of certain tax preparers. In 1998, 6% of individual audits were initiated because they had been prepared by a particular accountant, tax lawyers or other kind of tax professional.

The IRS has a massive computerized program that matches almost all payments made to individuals from different sources (salaries, interest from saving banks, dividends from corporations, etc.) with the income tax returns of the individuals. Surprisingly, the agency's computerized matching program was cited as prompting only 2% of the 1998 IRS district audits. (These figures do not include computer notices sent from Service Centers from IRS matching programs. These are much more frequent. But the district (face-to-face) audits are relatively rare.)

Information from states, including from state tax authorities, also proves to be a relatively unimportant source for audit selection by the feds. Only 3 out of every 1,000 IRS district audits arise for this reason (0.3%). Suspicion of fraud also is not a frequent reason for audit selection. Eight out of every 1,000 IRS audits arise for this reason (0.8%).