Fifth Circuit Affirm Lengthy Jail Sentence for “Sophisticated” Tax Fraud Conviction against a Tax Preparer. The Same Cannot Be Said When Banks Commit Fraud. Not One Banker Went to Jail in Texas

Appeal from the United States District Court for the Northern District of Texas ; USDC No. 4:16-CR-267-1

Before SOUTHWICK, WILLETT, and OLDHAM, Circuit Judges. PER CURIAM:*

SOUTHWICK

WILLETT

OLDHAM

Pritesh Patel pled guilty to two counts of aiding and assisting in the preparation of false tax returns. On appeal, his only claims of errors concern his sentence. He challenges the district court’s application of the sophisticated-means enhancement and its calculation of tax loss. We AFFIRM.

FACTUAL AND PROCEDURAL BACKGROUND

Patel operated a tax-preparation business in Arlington, Texas. During tax years 2007 through 2011, Patel prepared fraudulent tax returns for his clients that improperly claimed education credits. After an Internal Revenue Service (“IRS”) investigation, Patel was indicted on 12 counts of aiding and assisting in the preparation of false tax returns. Patel pled guilty to two counts and the other counts were dismissed.

The Government’s first sample of returns estimated a $4,805,820.50 tax loss. To arrive at this figure, the Government reviewed 100 income tax returns prepared by Patel in 2010. All claimed an education credit. A key part of the government’s evidence was the absence of the T-1098 Tax Form for many who claimed the credit. Educational institutions are required to file that form when a taxpayer pays educational expenses. The Government found that only 46 of the reviewed returns — or 46 percent — had a corresponding Form 1098-T on file with the IRS. Therefore, it concluded that 54 percent of the returns were fraudulent. Accordingly, the Government determined that of the total education credits of $9,802,317 claimed by Patel’s clients for tax years 2007 through 2010, 54 percent — or $4,805,820.50 — were fraudulent.

Patel pointed out that the sample contained 99 tax returns, not 100, but more importantly, that 63 of the taxpayers were actually entitled to the education credits even though a Form 1098-T for some of them was missing from the sample. The Government conceded that the sample contained 99 returns and that there were inaccuracies about missing Form 1098-T’s in that sample. To remedy these errors, the Government had the IRS Scheme Development Center (“SDC”), which was not the source of the data for the original sample, conduct a second sample. This second sample yielded an estimated tax-loss amount of $6,707,011.50. We discuss later the relevant details of this second sample.

Patel objected to the second sample and to the sophisticated-means enhancement. The district court overruled his objections and applied the sophisticated-means enhancement. Patel appealed.

DISCUSSION

I. Tax-loss calculation

The base-offense level for a tax-fraud offense is determined by calculating the loss that was the object of the offense. UNITED STATES SENTENCING GUIDELINES §§ 2T1.1(a), (c), 2T4.1 (2018). The Guidelines’ commentary states that when “the amount of the tax loss may be uncertain,” the district court may “make a reasonable estimate” of the loss “based on the available facts.” Id. § 2T1.1, cmt. n.1. A district court’s loss calculation is a finding of fact reviewed for clear error. United States v. Johnson, 841 F.3d 299, 303 (5th Cir. 2016). A district court’s method of determining the amount of loss is an application of the Guidelines which we review de novo. United States v. Harris, 597 F.3d 242, 250–51 (5th Cir. 2010). “To prevail on an argument that the district court’s calculation of tax loss was clearly erroneous, a defendant must introduce evidence to contradict or rebut the alleged improper computation of the loss.” Johnson, 841 F.3d at 303.

The Government estimated the tax loss in the following manner. An IRS agent testified that he identified the Patel-prepared tax returns by using an electronic filing identification number (“EFIN”), an employer identification number, and a preparer taxpayer identification number (“PTIN”) registered to Patel and his business by the IRS. According to the agent, for the 2007 tax year, Patel filed 1019 tax returns claiming education credits in the amount of $1,289,281. For tax year 2008, the agent determined that Patel filed 1631 tax returns claiming education credits in the amount of $2,058,341.

Next, the agent determined the validity of these claimed education credits for those two tax years. To do so, the agent examined an IRS database for the Form 1098-T, which would reflect that an educational institution received a relevant payment from the taxpayer. When the agent could not locate a Form 1098-T on file to correspond with the education credit claimed by a taxpayer, the education credit was categorized as false.

Based on this process, the agent determined that 509 of the 1019 tax returns Patel filed for tax year 2007, or almost exactly half, did not have a corresponding Form 1098-T on file with the IRS. For tax year 2008, 59 percent did not have a corresponding form. Using 50 percent as the figure for each year, the agent calculated the tax loss for 2007 through 2011 to be $6,707,011.50.

Patel objected to the Presentence Report’s (“PSR”) recommendation that a tax loss of over six million dollars had occurred. The district court found the figure to be “a reasonable estimate based on the available facts.” Patel renews his objection on appeal. The Government counters that the method it used to estimate Patel’s tax loss produced a reasonable estimate, which is all that is required by the Guidelines.

According to Patel, the Government’s premise for its calculations that the absence of a Form 1098-T equates to a fraudulent education tax credit is incorrect. Patel relies on the first sample performed by the Government, which made the same assumption and overlooked many of the forms. Thus, Patel reasons that the second sample must be flawed too.

Further, two documents produced by the National Association of College and University Business Officers show that educational institutions sometimes lose or misidentify a taxpayer identification number (“eTIN”) on the Form 1098-T. Finally, Patel argues the second sample itself contained errors where data on an IRS spreadsheet was incorrectly entered.

In our analysis, we start with the fact that, though Patel produced evidence that he had Form 1098-T’s that were missing from the first sample, he offered no similar evidence for the second sample that surveyed different tax-year returns. We do not agree with the district court that the errors in the first sample were irrelevant to the second. Indeed, those errors raise questions about the significance of a missing Form 1098-T. Nevertheless, Patel needed to show that the Government’s second sample did not bear sufficient indicia of reliability on which to base an estimate.

To support that the second sample contains significant errors, Patel relied on the National Association of College and University Business Officers Advisory Report, which indicated that educational institutions sometimes fail to provide a student’s correct TIN number on a Form 1098-T. The Government introduced evidence that the highest taxpayer identification error rate due to educational institutions’ mistakes from tax year 2009 through 2012 was just 3.6 percent. That figure does not correlate at all to the error rate in the initial sample, but it is affirmative evidence supporting the methodology that was not directly rebutted. Thus, although error rates are relevant and indeed indicate that the tax-loss estimate is imperfect, Patel has not shown that the tax-loss estimate was unreasonable. See Johnson, 841 F.3d at 305.

Finally, at sentencing, Patel questioned the IRS agent about the purported errors in the second sample’s spreadsheet. The agent defended the spreadsheet entries and explained why the spreadsheet did not contain the kinds of errors Patel claimed. After Patel concluded this line of questioning, the district court asked the agent whether the “objections or the substance of them ha[d] caused [him] to change [his] opinion on the fairness and reasonableness of the estimate.” The agent still believed it to be “a fair estimate.” Patel fails to support that any error was material to the final tax-loss estimate. See Johnson, 841 F.3d at 305.

II. Sophisticated-means enhancement

A defendant’s offense level is increased two levels if the offense involved “sophisticated means.” U.S.S.G. § 2T1.4(b)(2). That phrase is defined as “especially complex or especially intricate offense conduct pertaining to the execution or concealment of an offense.” Id. § 2T1.4, cmt. n.3. The determination is a factual finding, which this court reviews for clear error. United States v. Clements, 73 F.3d 1330, 1340 (5th Cir. 1996). The PSR itself “generally bears sufficient indicia of reliability” to be competent evidence. United States v. Narviz-Guerra, 148 F.3d 530, 537 (5th Cir. 1998). “Bald, conclusionary statements do not acquire the patina of reliability by mere inclusion in the PSR.” Id. (quoting United States v. Elwood, 999 F.2d 814, 817–18 (5th Cir. 1993)).

According to the PSR, Patel filled out applications for several PTINs using employees’ information without their knowledge. Patel would use these PTINs to prepare tax returns. “This allowed [Patel’s business] to prepare a large number of returns each year without the IRS knowing Patel was the preparer.”
Patel says the district court erred in finding that the offense involved sophisticated means.

Specifically, he argues that “there is no evidence as to how many returns used the PTINs in question or whether they were used on any of the returns allegedly claiming a false education credit.” He also asserts that prior to 2011, all tax preparers had to use the PTIN of the business owner. As such, he applied for PTINs in the name of others only for the 2011 tax year. Finally, Patel argues that according to the IRS agent’s testimony, registering multiple PTINs did not aid in masking his scheme.

First, although Patel used only his and his wife’s PTINs in filing returns from 2008 through 2010, Patel used the PTINs of two other individuals to file

approximately 270 tax returns in 2011, in addition to the 870 returns he filed using his wife’s PTIN that year. The PSR does not state which of the 270 tax returns filed under the other two employees’ names, if any, fraudulently claimed an education credit. This, though, does not show clear error. The district court reasonably estimated that 50 percent of the tax returns Patel filed were fraudulent.

It would therefore be unreasonable to find that all 270 tax returns were not fraudulent. Even if none of the 270 returns he prepared under the others’ names were fraudulent, he still took steps to “obscure the link between the” actual false tax returns and himself by decreasing the total number of prepared tax returns with his PTIN. Clements, 73 F.3d at 1340.

Next, at the sentencing hearing, the agent testified that regardless of which PTIN is used by a tax preparer, a return can be traced back to the tax preparer by using an EFIN. Indeed, the agent traced all the returns filed by Patel via one EFIN registered to him. The agent’s testimony was in the context of explaining how, after the scheme was discovered, he was able to verify the fraudulent returns Patel was responsible for filing. Patel ineffectively relies on the agent’s testimony.

Patel has failed to show that the district court’s finding was clearly erroneous. One circuit has concluded that the “essence of [sophisticated means] is merely deliberate steps taken to make the offense more difficult to detect.” United States v. Kontny, 238 F.3d 815, 821 (4th Cir. 2001). We have upheld sophisticated-means enhancements when defendants have used the names of other people to make it more difficult for their offenses to be detected, even if that method was not by itself particularly sophisticated. E.g., United States v. Conner, 537 F.3d 480, 492 (5th Cir. 2008); United States v. Wright, 496 F.3d 371, 379 (5th Cir. 2007); Clements, 73 F.3d at 1340.

Patel took deliberate steps to make his offense more difficult for the IRS  to detect. He used PTINs of other employees without their knowledge to repare tax returns. This allowed Patel to prepare a larger number of tax returns in 2011 without the IRS knowing that he was the sole preparer.

AFFIRMED.

I am text block. Click edit button to change this text. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

approximately 270 tax returns in 2011, in addition to the 870 returns he filed using his wife’s PTIN that year. The PSR does not state which of the 270 tax returns filed under the other two employees’ names, if any, fraudulently claimed an education credit. This, though, does not show clear error. The district court reasonably estimated that 50 percent of the tax returns Patel filed were fraudulent.

It would therefore be unreasonable to find that all 270 tax returns were not fraudulent. Even if none of the 270 returns he prepared under the others’ names were fraudulent, he still took steps to “obscure the link between the” actual false tax returns and himself by decreasing the total number of prepared tax returns with his PTIN. Clements, 73 F.3d at 1340.

Next, at the sentencing hearing, the agent testified that regardless of which PTIN is used by a tax preparer, a return can be traced back to the tax preparer by using an EFIN. Indeed, the agent traced all the returns filed by Patel via one EFIN registered to him. The agent’s testimony was in the context of explaining how, after the scheme was discovered, he was able to verify the fraudulent returns Patel was responsible for filing. Patel ineffectively relies on the agent’s testimony.

Patel has failed to show that the district court’s finding was clearly erroneous. One circuit has concluded that the “essence of [sophisticated means] is merely deliberate steps taken to make the offense more difficult to detect.” United States v. Kontny, 238 F.3d 815, 821 (4th Cir. 2001). We have upheld sophisticated-means enhancements when defendants have used the names of other people to make it more difficult for their offenses to be detected, even if that method was not by itself particularly sophisticated. E.g., United States v. Conner, 537 F.3d 480, 492 (5th Cir. 2008); United States v. Wright, 496 F.3d 371, 379 (5th Cir. 2007); Clements, 73 F.3d at 1340.

Patel took deliberate steps to make his offense more difficult for the IRS  to detect. He used PTINs of other employees without their knowledge to repare tax returns. This allowed Patel to prepare a larger number of tax returns in 2011 without the IRS knowing that he was the sole preparer.

AFFIRMED.