"God does not play dice," Albert Einstein famously said. The great physicist was speaking about the degree of order in a universe that can seem random - not, of course, about tax law.
But Einstein's rhetorical statement raises a relevant question: How does the IRS decide when to subject someone to a tax audit?
In the 1990s, a math professor analyzed over 1,000 tax returns and came to a simple conclusion. The professor, Amir Aczel, said that the IRS looks mainly at the relationship between expenses and income in deciding when to conduct an audit.
If expenses seem overly large compared to income, Aczel thought, the IRS computers may target the taxpayer for an audit - or at least flag a tax return for further review.
Aczel's analysis was not hidden away in an obscure academic journal. He published it in book form, with the attention-getting title of "How to Beat the IRS at its Own Game."
The analysis continues to attract considerable interest today. One commentator, Douglas Greenberg, recently suggested that Aczel was able to be so accurate in anticipating the IRS' actions because the IRS uses such old computers.
Some of these computers date back nearly fifty years. They are capable of assigning a "DIF' score that separates returns that will continue processing from returns that are to be audited. But the score a return gets appears to be based not on advanced algorithms, but on simple ratios.
These ratios include, for example, itemized deductions compared to adjusted gross income for Schedule A.
Source: "How the IRS Selects Returns for Audit," SFGate, Douglas Greenberg, 6-11-12