Court Affirms Loss Calculations Based on Sampled Data

US v. Mehta: Mehta was a tax preparer in Maryland who would prepare tax returns for clients that contained phantom deductions beyond those which the clients related to him. He would then submit the returns electronically to the IRS and, via BankOne, participate in a program whereby the bank cut a check for the taxpayer, with the bank later taking possession of the actual refund from the IRS. Mehta was convicted at trial of multiple counts of aiding and abetting the preparation of false tax returns and wire fraud. He was sentenced to 48 months in prison.

On appeal, Mehta challenged both his wire fraud convictions and his sentence. The Fourth Circuit affirmed on all counts. First, Mehta argued that the district court should have granted his motion for acquittal as to the wire fraud counts both because the evidence was insufficient to support convictions and that there was a variance between the indictment and the proof at trial. After reviewing the evidence, the court concluded that it was sufficient to sustain the convictions. As for the variance, the court concluded that the discrepancy as to which particular states the defrauded wires travelled through was not prejudicial to Mehta.

The court was more divided, in reasoning but not outcome, with regards to the proper calculation of Mehta's advisory Guideline range. At sentencing, the district court based Mehta's tax loss amount (which drove the Guideline range) on an IRS audit of only a portion of the total number of returns at issue. Specifically, the loss was based on the amount of additional tax owed that the taxpayers involved agreed to pay to the IRS. First, Mehta argued that district court should not have relied on those agreements by taxpayers to determine the amount of loss. The court concluded that those agreements were sufficient evidence to use in the loss calculation. Second, Mehta argued that the district court erred by taking the average tax liability from the IRS sample and multiplying it by the total number of fraudulent returns in order to determine loss. The court agreed, but found the error to be harmless, as the record still supported a loss amount sufficient to trigger the same advisory Guideline range. Judge Shedd concurred as to the result, but argued that there was no error.