Opinion
20-CR-272 (PKC)
06-26-2023
UNITED STATES OF AMERICA, v. DONAL O'SULLIVAN, HELEN O'SULLIVAN, and PADRAIG NAUGHTON, Defendants.
MEMORANDUM & ORDER
PAMELA K. CHEN, United States District Judge:
Before the Court is the issue of the proper amount of loss to be used for purposes of determining the United States Sentencing Guidelines (“Guidelines,” “U.S.S.G.”) range that applies to Defendants Donal O'Sullivan, Padraig Naughton, and Helen O'Sullivan. For the reasons that follow, the Court finds that the appropriate loss amount for purposes of sentencing is $1,214,976.65.
RELEVANT PROCEDURAL HISTORY
Following a 14-day trial in October 2021, Defendants were convicted of all counts in the Indictment, charging them with conspiracy and substantive crimes relating to mail and wire fraud, and embezzlement from employee benefit funds. (Verdict Form, Dkt. 262.) These charges stemmed from Defendants' participation in a multi-year fraud scheme to evade making required payroll contributions to certain benefits funds for employees of Navillus Tile, Inc. d/b/a Navillus Contracting (“Navillus”)-one of the largest construction companies in New York City-for work that was covered by collective bargaining agreements (“CBAs”). (Indictment, Dkt. 1, ¶¶ 1, 6-7.) At trial, the Government established that, between 2011 to 2017, Defendants fraudulently funneled payroll funds for 97 of Navillus's employees through a consulting firm, D.E.M. Consulting LLC d/b/a/ as Allied (“Allied”), which was not a party to the CBAs, to give the appearance that Navillus was not required to make benefits contributions or submit remittance reports to the benefits funds with respect to those employees' “covered work,” i.e., work for which benefits were owed under the CBAs. (See Mem. & Order, Dkt. 302, at 2, 37, 42, 45-46.)
The Guidelines calculation prepared by the Probation Department was based on the Government's loss calculation of $1,915,533.31. (See, e.g., Presentence Report for Donal O'Sullivan, Dkt. 338, ¶ 28; see Gov't. Letter, Dkt. 360, at 1.) Defendants Donal and Helen O'Sullivan (the “O'Sullivan Defendants”) objected to the Government's loss calculation and sought a loss hearing. (Def. Letter, Dkt. 328.) The parties submitted briefing in advance of the hearing, which was held on April 26, 2023. (Gov't. Letter, Dkt. 333; Def. Pre-Hr'g Mem., Dkt. 351.) At the hearing, the Defense called two expert witnesses and an attorney from the Defense team, and introduced multiple exhibits. (See Tr. of Fatico Hr'g (“Tr.”), April 26, 2023, Dkt. 364.) The Court directed the parties to submit post-hearing briefing, which was completed on June 6, 2023. (Gov't. Letter, Dkt. 360; Def. Letter, Dkt. 365.)
Although the letter filed by Defendant Donal O'Sullivan indicated that the request was being made on behalf of all three Defendants, who had agreed on “a joint defense position on loss amount” (Dkt. 328 at 1, n. 1), Defendant Naughton later clarified that he did not join in his codefendants' request for a loss hearing. (Dkt. 348.)
LEGAL STANDARD
The Government bears the burden of proving by a preponderance of the evidence the facts supporting the applicability of provisions of the Guidelines. See United States v. Archer, 671 F.3d 149, 161 (2d Cir. 2011); see also United States v. Desimone, 119 F.3d 217, 228 (2d Cir. 1997). With respect to the loss enhancement under U.S.S.G. § 2B1.1, the Government must prove “both the existence and amount of the loss attributable to the offenses of conviction.” United States v. Cuti, No. 08-CR-972 (DAB), 2011 WL 3585988, at *4 (S.D.N.Y. July 29, 2011) (citing United States v. Williams, 247 F.3d 353, 358 n. 7 (2d Cir. 2001)). However, in determining the amount of loss, the Court “need not establish the loss with precision but rather need only make a reasonable estimate of the loss, given the available information.” United States v. Uddin, 551 F.3d 176, 180 (2d Cir. 2009) (internal quotation marks and citations omitted); see United States v. Carboni, 204 F.3d 39, 46 (2d Cir. 2000); U.S.S.G. § 2B1.1, Application Note 3(C) (“The court need only make a reasonable estimate of the loss.”). Finally, the Court has considerable discretion in deciding what evidence or information to consider at sentencing. United States v. Kuyumcu, No. 16-CR-308 (DLI), 2017 WL 3995576, at *2 (E.D.N.Y. Sept. 8, 2017) (citing United States v. Pugliese, 805 F.2d 1117, 1122 (2d Cir. 1986)); see also United States v. Salemi, 46 F.3d 207, 210 (2d Cir. 2005).
The same standard applies to the Court's determination of the amount of restitution owed to victims under the Mandatory Victim Restitution Act (“MVRA”). See United States v. Donaghy, 570 F.Supp.2d 411, 423 (E.D.N.Y. 2008) (“The government bears the burden of proving the amount of loss sustained by the victim by a preponderance of the evidence....A restitution award need only to be a reasonable estimate of the victim's actual losses.”) (internal citations omitted); see also United States v. Gushlak, 728 F.3d 184, 196 (2d Cir. 2013) (“We reiterate that the MVRA requires only a reasonable approximation of losses supported by a sound methodology.”).
DISCUSSION
The Court finds that, as a starting point, the Government's methodology for calculating the loss caused by Defendants' illegal conduct is inherently sound. The Government relies upon Defendants' own trade classifications, as reflected in Navillus's payroll records, to identify those employees who were paid by Allied and as to whom the union benefits funds should have received contributions for covered work performed by those employees. While, as discussed below, there is no guarantee that a Navillus employee with a particular trade classification (e.g., bricklayer, carpenter, etc.) actually or always performed that type of work, it is certainly a fair and reasonable inference that these employees did so most of the time, since that was the skill they were employed for by Navillus and those were the types of skills needed for the construction jobs at issue. Indeed, as the Government notes and Defendants' payroll compliance auditing expert, David Evangelista, acknowledged, union benefits fund auditors use the same methodology when reviewing employers' payroll records to determine if employers owe contributions. (Dkt. 360 at 3-4; Evangelista Testimony, Tr. 41-43, 74.) Mr. Evangelista testified that “if [he] were conducting a payroll audit on behalf of a union benefit fund for a bricklayers union and . . . found payroll records that explicitly classified workers as bricklayers, [he] would claim that work for the benefit fund in the first instance[.]” (Tr. 74.) He further testified that “auditors typically would not interview . . . an employee of an employer[,]” but would instead “deal with the representative of an employer and then, you know, the process would ferret out if there were interviews or statements.” (Id.) Thus, the Court finds that the Government's loss amount of $1,840,873.71, could be a “reasonable estimate” in the absence of any other relevant information. Uddin, 551 F.3d at 180.
In the Court's view, the term “covered work” has two aspects: first, that the employee performed the type of work covered by a CBA or Project Labor Agreement (“PLA”) (e.g., bricklayer), and second, that the project was covered by a CBA and/or PLA signed by the employer, i.e., Navillus.
This amount represents a reduction of $74,659.60 from the Government's original loss amount, which the Government made after reviewing the Defense's loss-related materials and analyses. (Dkt. 360, at 2.)
Here, the O'Sullivan Defendants have presented voluminous evidence they claim shows that the vast majority of hours of work performed by the 97 Allied-paid Navillus employees that the Government claims was “covered” work was, in fact, “non-covered” work, and that the more accurate loss amount is $193,736.17 (Dkt. 351 at 1; Dkt. 349-6, at 2)-approximately one-tenth of the Government's loss calculation. The two principle bases for the Defense's challenge are that (1) based on Defense interviews of many, if not a majority, of the 97 employees (some now former employees) or their supervisors (where the employee could not be located or interviewed), the employee did not actually perform work covered by a CBA or PLA at the job site; and (2) much of the work counted as covered by the Government was, in fact, non-covered, either because it was either yard work, which is not covered by any CBA or PLA, or offsite work excluded under the School Construction Authority's PLAs with Navillus (“SCA PLAs”), which constituted much of the work at issue in the case. (Dkt. 365 at 1-2.) The Court finds that, giving some benefit of the doubt to the Defense (for reasons discussed more fully below), the first basis warrants a reduction in the loss amount in this case, but that the second basis does not.
It is unclear to the Court how many of the 97 employees Defense investigator, Laureen Fitzgerald, interviewed. While her affidavit indicates that she interviewed “44 individuals during [the] pre-trial period,” i.e., October 2020 through August 2021, and then “approximately 50 individuals during [the] post-trial period,” i.e., from January 2022 through August 2022, her affidavit does not indicate whether there was any overlap between the first and second groups. (Decl. of Laureen Fitzgerald, Dkt. 349-4 ¶¶ 6, 28.) Ms. Fitzgerald further states that the “second round of interviews” focused on “individuals who either worked in the yard or who were supervisors.” (Id. ¶ 27.) Although the Defense has provided an interactive spreadsheet explaining the basis for its proposed loss amount reductions, by employee, that includes links to interview statements taken by Ms. Fitzgerald, there appears to be no statement as to how many of the 97 employees at issue were interviewed.
I. Reduction to Loss Amount Based on Supervisory/Safety Work
The Court addresses the first basis. As discussed, the Government's method for calculating loss, while sound as an initial matter, does not take into account whether the Navillus employee actually performed covered work as suggested by the company's payroll records. As with a benefits fund audit and a loss determination under the Guidelines, the amount of contributions owed or loss caused should be based on all available competent evidence and information. See Uddin, 551 F.3d at 180 (loss amount can be based on a “reasonable estimate, given the available information”). Thus, here, the Court should consider whether the employee or supervisor interview statements proffered by the Defense undercut the otherwise reasonable inference that employees assigned to certain trade classifications in Navillus's payroll records, in fact, performed work in that trade while at the covered work site.
The Court does not fault the Government for not conducting the interviews performed by the Defense. Logically, the Government does not have the same access as Defendant Donal O'Sullivan and his legal team do to current Navillus employees and supervisors (who presumably want to keep their jobs) or former employees (who presumably might want to work for Navillus in the future). It also is unrealistic to think that Navillus's current employees or supervisors, or even their former ones, would agree to speak to the Government, and thus short of subpoenaing them (which it is unclear the Government has the power to do in this context), the Government simply does not have the ability to interview these individuals or obtain the same information proffered by the Defense. Relatedly, the Court notes that the vitriol directed at the Government- e.g., Dkt. 365 at 2 (“Had the government done the work, it too would have come to the conclusion that the loss to the ERISA benefit funds over the six-year period charged in the indictment totaled less than $200,000. . . . That the government digs its heels in on its nearly $2,000,000 loss figure demonstrates the prosecutors' intellectual dishonesty or laziness.”)-is neither warranted in this case nor helpful to the Court in any case.
As an initial matter, the Court shares the Government's skepticism about the reliability of the interview statements, though (like the Government) the Court does not question Ms. Fitzgerald's or the Defense team's integrity or methods. Rather, the Court's skepticism is based on several factors. First, there is the obvious incentive for these current employees and supervisors, and former employees (who could be re-hired at any time or sporadically by Navillus) to want to maintain or curry goodwill with Defendants, particularly Navillus president Donal O'Sullivan, especially given that there were no apparent consequences for inaccuracies in their statements. Second and relatedly, the Court has had no opportunity to assess the veracity or reliability of these interviewees' statements and recollections-in contrast to the trial testimony of Navillus employees or former employees about their performance of covered work, consistent with their trade classifications in Navillus's payroll records. Third and in any event, the individuals interviewed by the Defense were asked to remember the specific tasks they performed on particular job sites anywhere from three to eleven years earlier. (Dkts. 349-4 ¶¶ 5, 27 (interviews conducted either October 2020 to August 2021 or January 2022 to August 2022); Indictment, Dkt. 1 (fraud scheme occurred between 2011 to 2017)). The Court finds it unlikely that the employees', former employees', or supervisors' memories were so precise or clear.
Nonetheless, despite the Court's general skepticism about the reliability of the proffered employee/supervisor statements disputing their performance of covered work, the Court finds that some reduction of the loss amount is warranted, given the likelihood, suggested by the volume of Defense evidence, that at least some of the work counted by the Government as covered was noncovered. In the demonstrative exhibit presented at the Fatico hearing by attorney Taylor Bleistein, the Defense identifies and summarizes the number of hours for non-covered work it believes occurred based on employee/supervisor statements indicating that the actual work performed was that of a supervisor or safety worker. (Tr. at 102-09 (introducing Def. Ex. G); Dkt. 349-6 at 2.) The total number is 35,127 hours. (Id.) Because the Court cannot determine the precise amount of contributions tied to that number of hours, it has applied a percentage approach to reducing the Government's proposed loss amount. (Id.; see also Dkt. 349-6, at 2.) Dividing 35,127 hours by 102,585-which is the total number of hours disputed by the Defense-results in a percentage of 34%. When multiplied against the Government's revised loss amount of $1,840,873.71, that percentage results in a reduction of $625,897.06 and, consequently, a total loss amount of $1,214,976.65. For the above-stated reasons and given that this approach does not factor in the different contribution rates applicable to each employee, the Court considers this reduction to be generous to Defendants.
II. Work Covered by SCA PLAs
Turning to the Defense's second basis for challenging the Government's loss calculation, the Court rejects the Defense's argument that SCA PLAs applicable to a large portion of the work deemed to be covered by the Government superseded the applicable CBAs and rendered this work non-covered. (See Dkt. 349-6 (55,997 hours of work challenged by Defense as yard work or work excluded by SCA PLAs); Tr. at 107 (attorney Bleistein testifying that “over 60 . . . school jobs are under [the second-to-last column in Defense Exhibit G titled “Yard Work/SCA PLA]” and acknowledging that this category constitutes the “lion's share” of the discrepancy between the Government's and Defense's covered work/contributions owed/loss figures).) First, putting aside the parties' dueling arguments about the purpose or effect of the SCA PLAs-i.e., to level the playing field for employers who use union labor as opposed to companies who do not (Defense) versus “to subject all construction contractors retained to perform work on school projects to the contribution requirements set forth in [CBAs] of the participating unions regardless of whether those contractors were independent signatories to those agreements” (Government)-the Court finds that the plain language of the Supremacy Clause in the SCA PLAs does not create a conflict with the corresponding CBAs; rather, it exempts employees who perform offsite tasks from the application of the SCA PLAs, but not from any applicable CBA.
The parties agree on the relevant portions of the SCA PLAs. First, the SCA PLAs include a section entitled “Excluded Employees” (“Section 3”) that identifies “persons [who] are not subject to the provisions of [the PLA], even though performing” work covered by the PLA. (SCA PLA for 2009-2014 at Article 3, Section 3, Dkt. 365-2 at ECF 20.) Persons excluded from the provisions of the SCA PLAs include:
“ECF” refers to the pagination generated by the Court's electronic filing system or “ECF” and not the document's internal pagination.
Employees and entities engaged in offsite manufacturing, modifications, repair, maintenance, assembly, painting, handling or fabrication of project components, materials, equipment or machinery or involved in deliveries to and from the Program site.(Id. at Article 3, Section 3, subsection c, Dkt. 365-2, at ECF 16.) Second, the SCA PLAs contain a “Supremacy Clause,” which provides in relevant part:
This agreement, together with the local Collective Bargaining Agreements appended hereto as Schedule A, represents the complete understanding of all signatories and supersedes any national agreement, local agreement or other collective bargaining agreement of any type which would otherwise apply to this Program Work, in whole or in part. Where a subject covered by the provisions of this Agreement is also covered by a Schedule A, the provisions of this agreement shall prevail.... It is further agreed that, where there is a conflict, the terms and conditions of this Project Labor Agreement shall supersede and override terms and conditions of any and all other national, area, or local collective bargaining agreements.(SCA PLA for 2009-2014, at Article 2, Section 4, Dkt. 365, at 6 (emphasis added).)
The Defense argues that where a CBA requires an employer to pay contributions for “offsite” work performed by one of its employees, it is in conflict with, and therefore “trumped” by, the Supremacy Clause and Section 3-the Excluded Employees provision-of the SCA PLA, so as to excuse employers from making contributions otherwise required by the applicable CBA for offsite work performed as part of a job covered by a SCA PLA. (Dkt. 365 at 4-7.) The Court disagrees.
As the Government contends, Section 3 only excludes from “the provisions of Agreement [i.e., the SCA PLA]” those employees who, inter alia, engage in offsite work. (SCA PLA for 2009-2014 at Article 3, Section 3, Dkt. 360, at 7.) Section 3 neither states or even suggests that these employees are excluded from the provisions of the applicable CBA, which, under the Supremacy Clause, “together with” the SCA PLA, “represents the complete understanding of all signatories and supersedes any national agreement, local agreement or other collective bargaining agreement of any type which would otherwise apply to this Program Work, in whole or in part.” (SCA PLA for 2009-2014, at Article 2, Section 4, Dkt. 360, at 7.) Thus, based on Section 3's plain language, Navillus employees who performed offsite work for covered SCA jobs “were not subject to the provisions of the SCA PLA,” yet they were otherwise subject to the provisions of CBA applicable to the work that they performed in connection with the project. Section 3 therefore does not create a “conflict” between the SCA PLAs and the corresponding CBAs so as to trigger the application of the SCA PLA's Supremacy Clause.
The Court recognizes that the 55,997 hours in the second-to-last column of Defendants' demonstrative exhibit (Dkt. 349-6) includes hours that the Defense argues should be deducted because it was for yard work, which is not covered by any CBA. Because the Court cannot determine how much of the 55,997 was purportedly for yard work (versus for offsite or other work exempted by the SCA PLAs) and because the 35,127 Supervisory/Safety Workers hours likely overstates the number of non-covered hours, the Court does not further reduce the loss amount by any portion of the Yard Work/SCA PLA hours. In addition, the Court does not find that the Defense's argument about the Government applying the wrong fringe benefit rate (e.g., apprentice versus journeyman) warrants any reduction in the loss amount because the Court simply does not have sufficient information regarding what each of the 97 employees' proper benefit rate was or should have been. The Court agrees with the Government that “[t]he willingness of the defendants (through Navillus) to compensate [Jorge] Mejia at an hourly rate well above that of an apprentice [i.e., the rate relied upon by the Government] reflects the defendants' own assessment of his experience level and is a better proxy for determining the correct fringe benefit rate than the total number of hours reflected in the payroll records for Navillus and Allied, which . . . do not reflect hours Mr. Mejia may have worked in the industry for other employers.” (Dkt. 360 at 9-10.)
In closing, the Court notes that the calculation of loss in a case like this will necessarily be imperfect and approximate, given the myriad individualized facts that might affect that calculation, but which, practically speaking, cannot be fully considered by the Court, e.g., the impossibility of the Court convening a Fatico hearing at which each of the 97 employees testify about the work they actually performed (assuming they remember) on covered job sites. Such absolute precision, however, is not what is required. Rather, the Court need only be satisfied that the loss calculation is a “reasonable estimate, given the available information.” Uddin, 551 F.3d at 180. The Court finds that the reduced loss amount of $1,214,976.65 meets that standard.
SO ORDERED.