Advertisement
YOU ARE HERE: LAT HomeCollections

Budget Crunch Forces IRS to Cancel Random Audits

October 24, 1995|ROBERT A. ROSENBLATT | TIMES STAFF WRITER

WASHINGTON — Strapped by a tight budget, the Internal Revenue Service is canceling a planned line-by-line audit of the tax returns of 153,000 randomly chosen individuals and businesses.

One of the toughest of all IRS audits, the Taxpayer Compliance Measurement Program--sometimes fearfully referred to by taxpayers as the audit from hell--helps the agency find out how taxpayers are cheating.

The program forces taxpayers to justify every expense and deduction, certify every dependent and include every cent of income under the watchful gaze of IRS agents assigned exclusively to the task. It provides the most complete picture available of how taxpayers "do and do not comply with the law," Frank Keith, the IRS chief of media relations, said in an interview Monday.

But the latest version of the audits, scheduled to begin in December, with agents examining 1994 returns for as long as 30 months, will not be carried out, Keith said.

"Because of the budget situation facing us in 1996, we will have program cuts and curtailments," Keith said. A funding bill now before a House-Senate conference committee would cut the IRS' 1996 budget from $8.2 billion to $7.4 billion. The agency would save $1.4 billion by eliminating the compliance program, Keith said.

This would have been the first time in seven years that the government conducted its most detailed inquiry into the cheating patterns of individual taxpayers. The last taxpayer compliance program for individual taxpayers was in 1988.

Under the program, taxpayers are investigated at random to help the IRS learn how to program its computers in its search for cheaters. The program has been carried out periodically for the last 30 years.

A previous taxpayer compliance program inquiry indicated that there were widespread false claims of children as dependents. As part of the examination, taxpayers were asked for the birth certificates of their children.

The IRS subsequently decided in the late 1980s to ask taxpayers to include a Social Security number for each child claimed as a dependent. The next year, the number of children claimed dropped by 7 million.

The IRS asked parents claiming the child-care credit to provide the Social Security number of the person watching their children or the tax identification number of the day-care business. The next year, hundreds of thousands of claims for the child-care credit disappeared, indicating to the IRS that there had been widespread false claims.

There was also greater compliance among those actually providing child care, according to Keith. Realizing that parents were listing their names and numbers, thousands of child-care providers began reporting more income to the IRS.

Before the program began in the 1960s, audits produced changes in tax liabilities only 54% of the time. But the program gave the government potent ammunition for its audits by revealing patterns of cheating. Now 81% of the audits result in tax-liability changes.

The line-by-line audits produce less revenue than regular audits because IRS agents spend more time on them and because the random selection of taxpayers for audit makes it less likely that tax violations will be uncovered.

The regular audit system depends on secret computer formulas suggesting which taxpayers are good subjects for audit. Certain types of businesses, certain amounts of deductions or expense claims can trigger these regular audits.

This year's program covering 1994 returns--those filed in April of this year--originally had been scheduled to start Oct. 1. Then the project was delayed until December. It would have covered individuals, partnerships and corporations--a total of 153,000 returns.

Keith said that it is "impossible to speculate" when the program might be revived.

Meanwhile, he said, the IRS will begin seeking other ways to determine how taxpayers are failing to comply with the law.

Advertisement
Los Angeles Times Articles
|
|
|