Tax Loss Estimations for Sentencing Purposes

In United States v. Poltonowizc (3rd Cir. 2009), an unreported, nonprecedential decision, the court approved tax loss estimations for a convicted return preparer (former IRS CI analyst). His tax evasion scheme was unsophisticated. In a sting operation,

Although the agent never mentioned charitable contributions, and provided him with no evidence whatsoever of any such contributions, he included $2,190 in cash ontributions to charity and $495 in non-cash contributions to charity on the agent's return. As a result, the agent's tax return showed that she was entitled to a $12 refund, instead of reflecting that she owed $1,012 in additional taxes. Subsequently, the agent requested a meeting with Poltonowicz to discuss a letter she received from the IRS informing her that she would be audited. Again, the agent wore a recording device. He admitted to the preparation of a false tax return and that he included the false deductions to save her from paying additional taxes (as he operated under the assumption that she would not be audited). He reassured her that she would not get in trouble for the fraudulent return.

Poltonowicz pled to one count of filing a false tax return. He thereafter continued his pattern of conduct through another company in the name of a female, described as his "long-time roommate and housekeeper." In a second trial, a jury convicted him of unspecified tax crimes. Moving to the sentencing phase, the defendant's position was that a particularized inquiry should be made into the dollars included in the estimated tax loss calculation under the Sentencing Guidelines. The Government, however, calculated the tax loss in a less precise way. The Government's calculations included only tax losses for returns personally prepared by Poltonowicz. (Specifically, it excluded returns prepared by employees who had, according to the testimony, claimed similar false deductions at Poltonowicz's direction or teaching.) Of that set,

Of that subset of tax returns, the government included only those that contained one of the methods of falsifying tax returns established at trial, such as fictitious cash and non-cash charitable contributions, employee non-reimbursed expenses, and claims of eligibility for the earned income tax credit. The government filtered that subset to include two types of returns: (1) returns for which the IRS had conducted an audit and had subsequently assessed the taxpayer with additional tax liability based on the tax payer's inability to substantiate their return, or (2) returns for taxpayers interviewed, who confirmed that they did not provide any evidence of the deductions at issue or request that they be included. The estimate of $419,853.20, in the manner calculated, was actually under inclusive.

Addressing Poltonowicz's arguments on appeal, the court of appeals said:

The district court relied on evidence presented at trial and the sentencing hearing to reach its conclusion. The government established the modus operandi -- preparing tax returns with fictitious data for charitable contributions, employee non-reimbursed expenses, and claims of eligibility under the earned income tax credit. It did not err in including tax returns in the tax loss calculation which had been subject to and had failed an audit by the IRS, even if the government did not interview the tax payer. Poltonowicz is on audiotape informing a potential client that he knew exactly how to claim fictitious deductions without getting caught. Indeed, the evidence suggests a much larger tax loss. He personally prepared 20,000 to 25,000 tax returns, yet the government calculated its tax loss based on just 225 of those returns. One former employee testified that at least 25% of the returns Poltonowicz filed contained fictitious deductions. The government excluded from its calculation any returns that were prepared by employees, even though several employees testified that he directed them to add fictitious deductions to the returns they filed. On average, 50-54% of returns claim charitable contributions; whereas, 98% of Poltonowicz's clients claimed such deductions. Notably, his clients uniformly claimed to donate in one of three precise amounts: $490, $495, and $500.

Poltonowicz also challenges the government's calculation of additional losses by comparing his average claims for certain deductions, such as the charitable deduction, with that of the national average. He asserts that it was improper to compare his clients to the national average because his clients were not average tax payers; rather, his clients consisted of blue-collar, religious, conservative tax payers who were far more likely to make charitable contributions than the average tax payer. He makes a similar argument with respect to the government's comparative information on employee non-reimbursed expenses. These arguments lack merit. The District Court did not rely on the government's comparative data in reaching its conclusion that the tax loss exceeded $ 400,000. The District Court based its conclusion on the audited returns and mentioned the additional statistical evidence in noting that the government's calculation was extremely conservative. There is no error with a District Court's consideration of statistical evidence in a case involving upwards of 20,000 tax returns.

This type of estimation would appear to be appropriate under the Guidelines in setting a reasonable minimum tax loss for sentencing purposes. There is a related, but quite different, issue of whether anything less than actual proof of a substantial tax loss due for purposes of the evasion element of tax due and owing is appropriate in the case in chief. I have previously argued in my blogs in the context of criminal prosecutions of tax enablers where the taxpayers are absent such estimations are not appropriate.