Bayer Corp. & Subsidiaries v. United States

Full title: BAYER CORPORATION AND SUBSIDIARIES, Plaintiffs, v. The UNITED STATES…

Court: United States District Court, W.D. Pennsylvania.

Date published: Feb 6, 2012

850 F. Supp. 2d 522 (W.D. Pa. 2012)

109 A.F.T.R.2d (RIA) 2012

Opinion

Civil Action No. 09–351.

2012-02-6

BAYER CORPORATION AND SUBSIDIARIES, Plaintiffs, v. The UNITED STATES, Defendant.

Clifton B. Cates, Ivins, Phillips & Barker Chartered, Washington, DC, John M. McIntyre, Alexandria C. Samuel, Reed Smith LLP, Pittsburgh, PA, for Plaintiffs. Jan M. Geht, Christopher J. Williamson, U.S. Department of Justice, Washington, DC, Jennifer R. Andrade, United States Attorney's Office, Pittsburgh, PA, for Defendant.

WILLIAM L. STANDISH

Clifton B. Cates, Ivins, Phillips & Barker Chartered, Washington, DC, John M. McIntyre, Alexandria C. Samuel, Reed Smith LLP, Pittsburgh, PA, for Plaintiffs. Jan M. Geht, Christopher J. Williamson, U.S. Department of Justice, Washington, DC, Jennifer R. Andrade, United States Attorney's Office, Pittsburgh, PA, for Defendant.

OPINION

WILLIAM L. STANDISH, District Judge.

This civil action arises out of the denial of federal income tax credits for qualified research expenses claimed by-Plaintiffs, Bayer Corporation and Subsidiaries (“Bayer”), for the years 1990–2006. Before the Court is Bayer's Amended Motion for a Case Management/Protective Order Based on Statistical Sampling (“Amended Sampling Motion”). (Docket No. 71). For the reasons set forth below, the Amended Sampling Motion will be denied.

FINDINGS OF FACT

After consideration of the testimony and evidence presented by the parties during a hearing on the Amended Sampling Motion, the Court makes the following findings of fact: ¹

1 Due to the complexity and lengthy procedural history of this case, the Court also has included applicable provisions of the Internal Revenue Code and undisputed facts in its findings for background purposes. The undisputed facts are derived from the docket, evidence submitted in support of other motions filed by the parties and Bayer's informational Power Point presentation to the Court on May 4, 2010.

Internal Revenue Code

1. A federal income tax credit for qualified research expenses, commonly referred to as “QREs,” was established by the Economic Recovery Tax Act of 1981. The purpose of the tax credit for QREs was to “stimulate a higher rate of capital formation and to increase productivity.” S.Rep. No. 97–144, at 76–77 (1981), 1981 U.S.C.C.A.N. 105, 182.

2. QREs are a subset of research expenses in general and are limited to salaries and wages, supplies and contract research performed by third parties. ² (Docket No. 81, pp. 45–46). In general, the credit for QREs for any taxable year is 20% of the excess QREs for the taxable year over a specified base period.³ 26 U.S.C. § 41(a) and (c).

2 QREs are defined in the Internal Revenue Code as the sum of in-house research expenses and contract research expenses paid or incurred by the taxpayer during the taxable year in carrying on any trade or business of the taxpayer. “In-house research expenses” include “any wages paid or incurred to an employee for qualified services performed by such employee,” and “any amount paid or incurred for supplies used in the conduct of qualified research.” 26 U.S.C. § 41(b)(2)(A)(i). The term “contract research expenses” means “65 percent of any amount paid or incurred by the taxpayer to any person (other than an employee of a taxpayer) for qualified research.” 26 U.S.C. § 41(b)(3)(A).

3 Excess (currently unusable) QRE credits may be carried backward or forward by a taxpayer to offset tax liability for other years.

3. Due to concerns that taxpayers were interpreting the tax credit for QREs too broadly and that “some taxpayers ... claimed the credit for virtually any expenses relating to product development,” S.Rep. No. 99–313, at 694–95 (1986), the credit for QREs was amended by the Tax Reform Act of 1986, Pub.L. 99–514, section 231(b), 100 Stat. 2173, to provide a definition of “qualified research.”

4. Under Section 41(d) of the Internal Revenue Code, “qualified research” means research “which is undertaken for the purpose of discovering information-(i) which is technological in nature, and (ii) the application of which is intended to be useful in the development of a new or improved business component of the taxpayer, ...,” and the term “business component” means “any product, process, computer software, technique, formula, or invention which is to be-(i) held for sale, lease, or license, or (ii) used by the taxpayer in a trade or business of the taxpayer.” 26 U.S.C. § 41(d)(1). The test for determining whether an expense was incurred in connection with qualified research is to be applied separately to each business component of the taxpayer. 26 U.S.C. § 41(d)(2)(A).

5. In 2003, the Treasury Regulation relating to recordkeeping requirements for QRE credits was changed to read, in relevant part, as follows;

§ 1.41–4 Qualified research for expenditures paid or incurred in taxable years ending on or after December 31, 2003

* * *

(d) Recordkeeping for the research credit. A taxpayer claiming a credit under section 41 must retain records in sufficiently usable form and detail to substantiate that the expenditures claimed are eligible for the credit ....

* * *

Treas.Reg. § 1.41–4(d).

Although the 2003 amendment to recordkeeping requirements for QRE credits under Section 41 of the Internal Revenue Code is prospective, taxpayers may opt for retroactive application of the regulation and Bayer has done so in this case.⁴ (Hearing Tr. 7/18/11, Docket No. 81, pp. 103–04).

4 Records of Bayer's research expenses are maintained in two forms: paper and electronic, (Hearing Tr. 7/18/11, Docket No. 81, pp. 52–54).

Background

6. Bayer is a multifaceted company employing more than 20,000 people in the United States. It is headquartered in Pittsburgh, Pennsylvania. Bayer is comprised of the following divisions: healthcare, material science and crop science. Bayer Healthcare includes animal health, biological, consumer care and pharmaceutical units; Bayer MaterialScience LLC is one of the leading producers of polymers and high performance plastics in the United States; and Bayer CropScience is one of the world's leading innovative companies in the areas of crop protection, non-agricultural pest control, and seed and plant biotechnology. (Bayer Power Point presentation, 5/4/10).

7. Bayer engages in research activities at numerous sites throughout the United States. Bayer also conducts research in other countries including Germany, Japan, India, France, the United Kingdom and Singapore. The availability of the tax credit for QREs in Section 41 of the Internal Revenue Code is a factor in the decision of Bayer's parent, Bayer AG, a publicly traded German company, as to where research will be conducted. (Hearing Tr. 7/18/11, Docket No. 81, pp. 38–41).

8. Commencing in the 1960s and continuing through the 1990s, Bayer maintained two main frames, i.e., big, bulky IBM computers, in the United States. Bayer purchased or leased software to keep general ledgers, sales ledgers, payroll, inventory and logistical information on the main frames. In the late 1990s, Bayer changed to the SAP Enterprise Accounting System which utilizes servers that connect personal computers, i.e., desktop and laptop computers, rather than large main frame computers. To reduce high software and maintenance costs, the electronic information stored on Bayer's two main frames in the United States was consolidated on the main frame in Pittsburgh at some point. Thereafter, in 2005, the Pittsburgh main frame was decommissioned and the electronic information moved to Bayer's main frame in Germany. In late June 2011, the Germany main frame, which was Bayer's last remaining main frame, was decommissioned. At the time, the electronic information stored on the main frame was examined and certain data was retained in light of this litigation. (Hearing Tr. 7/18–19/11, Docket No. 81, pp. 52–54, Docket No. 82, pp. 94–97, 102–03).

9. Bayer does not account for research expenses by individual projects. Rather, Bayer uses an accounting system that is based on the nature of activities performed. Similar activities are grouped into “cost centers” and individual expenses, such as payroll, supplies and outside services, are charged to the appropriate cost centers and then to specific expense classes within those cost centers.⁵ (Docket No. 71–4, p. 2, ¶ 6).

5 Examples of Bayer's cost centers include “formulation development” (for crop protection agents) and “clinical studies” (for pharmaceuticals). (Docket No. 72, p. 6).

10. In addition to accounting records, Bayer substantiates QREs for the purpose of claiming tax credits under Section 41 of the Internal Revenue Code with internal status reports, emails, correspondence with federal agencies, EXCEL files, Word documents, Power Point presentations, access databases, lab notebooks, patent applications, and standardized and centrally maintained personnel and payroll records.⁶ Bayer also substantiates QREs through evidence provided by the employees and former employees who performed the research. (Hearing Tr. 7/18/11, Docket No. 81, pp. 54–56).

6 Prior to the initiation of this litigation, Bayer's legal department issued hold notices to individuals identified by Bayer's tax department as working in relevant cost centers to retain all paper and electronic documents that could in any way relate to research. (Hearing Tr., 7/18/11, Docket No. 81, pp. 61–62).

11. Paul F. Wright has been employed in Bayer's Tax Department for 28 years. Mr. Wright, who currently serves as Vice President, Tax for Bayer, oversees the preparation of all tax returns filed by Bayer and its affiliates (approximately 5,000 returns annually). (Hearing Tr. 7/18/11, Docket No. 81, p. 33).

12. In the mid–1990s, Mr. Wright became aware through discussions with various Bayer employees that Bayer's claims for QRE credits under Section 41 of the Internal Revenue Code were limited to expenses incurred in cost centers involving pure or indisputable qualified research activities, and that expenses charged to other cost centers, such as quality assurance, also may include expenses for qualified research activities. (Hearing Tr. 7/18/11, Docket No. 81, pp. 93–94).

13. In 1997, based on Mr. Wright's recommendation, Bayer retained Deloitte & Touche LLP, an accounting firm, to conduct a study to determine whether Bayer was claiming all of the QRE credits to which it was entitled under Section 41 of the Internal Revenue Code (“the Deloitte study”). At the time, the Internal Revenue Service (“IRS”) had completed its examination of QRE credits claimed by Bayer for two audit cycles, i.e., 1990–1991 and 1992–1994, and the parties had reached an agreement concerning those credits.⁷ Following completion of the Deloitte study in 1998,⁸ Bayer filed a claim with the IRS for additional QRE credits for all open tax years which included 1990 to 1997.⁹ The IRS denied not only Bayer's claim for additional QRE credits for the open tax years, it denied the credits for QREs that had previously been agreed upon for the 1990–1991 and 1992–1994 audit cycles. (Hearing Tr. 7/18/11, Docket No. 81, pp. 33–34, 90–91, 93–96).

7 The IRS typically audits Bayer for 2 or 3 years at a time. (Hearing Tr. 7/18/11, Docket No. 81, p. 90).

8 As part of the Deloitte study, 23 of Bayer's 49 research sites generating the QREs that are the subject of this case were visited. (Hearing Tr. 7/18/11, Docket No. 81, p. 96).

9 Bayer submitted the report of the Deloitte study, which was comprised of sixty 3–ring binders, to the IRS in support of the claim for additional QRE credits, (Hearing Tr. 7/18/11, Docket No. 81, p. 96).

Commencement and History of Litigation to Date

14. On March 23, 2009, Mr. Wright initiated this civil action against the Government on Bayer's behalf seeking a refund of federal income taxes in the amount of $49,236,589.¹⁰ The refund claim arises out of the IRS's complete or partial denial of QRE credits claimed by Bayer or its predecessors-in-interest for the years 1990–2006 inclusive (the “credit years”). ¹¹ Bayer alleges that as a result of the IRS's complete or partial denial of its claims for credits for QREs incurred during the credit years, Bayer overpaid the federal income tax due for the years 1987–1990, 1995 and 2006.¹² (Docket No. 1, No. 81, p. 33).

10 Prior to the commencement of this case, Bayer–Onyx, a partnership, by Bayer, its “tax matters partner,” filed a Complaint for Readjustment of Partnership Items under Section 6226 of the Internal Revenue Code, and the case was assigned Civil Action No. 08–693. In that case, Bayer–Onyx alleges that the Government erred in disallowing $178,271 of QRE credits claimed in 1994, $115,427 of QRE credits claimed in 1995 and $63,593 of QRE credits claimed in 1996. At the request of the parties, Civil Action No. 08–693 was stayed until April 13, 2009. Prior to the expiration of the stay, the present related case was filed. On April 20, 2009, the parties filed a joint motion to consolidate Civil Action No. 08–693 with this case for discovery purposes. The motion was granted the next day.

11 According to Mr. Wright, after the filing of its claim for additional QRE credits for 1990 to 1997 based on the Deloitte study, the IRS disallowed approximately 98% of the credits for QREs previously agreed upon for the 1990–1991 and 1992–1994 audit cycles. Thus, “[Bayer] ended up much worse off than when [Bayer] actually started this process.” (Hearing Tr. 7/18/11, Docket No. 81, pp. 101–02).

12 With respect to the years for which no refund is claimed, QREs were incurred but could not be used because Bayer incurred no tax liability for those years. The unusable credits were carried backward or forward to other tax years by Bayer. (Bayer Power Point presentation, 5/4/10).

15. The Government denies that Bayer is due a refund of federal income taxes for the years 1987–1990, 1995 and 2006. Moreover, the Government has asserted a counterclaim against Bayer for federal income taxes for the year 2006 in the amount of $80,361,674, together with accrued interest totaling $13,323,180, which was assessed against Bayer by the IRS on September 9, 2009. The assessment arises out of the disallowance of Bayer's claim “for research tax credits under 26 U.S.C. § 41 for the year 2006 (including research credit carry forwards from prior years applied to 2006), as well as other non-research tax credit adjustments the [IRS] made during the audit of [Bayer] for the 2006 tax year.” ¹³ (Docket No. 25, ¶ 18).

13 Bayer estimates the total amount at stake in this tax litigation to be $175,000,000. (Bayer Power Point presentation, 5/4/10, Hearing Tr. 7/18/11, Docket No. 81, p. 93).

16. On March 15, 2010, the Government served Bayer with its First Set of Interrogatories. In Interrogatory No. 26, Bayer was asked to “[i]dentify and describe each new or improved business component Bayer contends it incurred qualified research expenses to develop during the credit years.” ¹⁴ (Docket No. 31–1). Bayer responded to Interrogatory No. 26 as follows:

14 As noted in Finding of Fact No. 4, Section 41(d)(2)(A) of the Internal Revenue Code provides that the test for determining whether a research expense is a QRE for purposes of the tax credit is to be applied separately to each business component of the taxpayer. Thus, in Interrogatory No. 26, the Government sought the identity of the business components underlying the QREs claimed by Bayer for the credit years. In this connection, the Court notes that the parties have reached an agreement regarding the QRE credits claimed by Bayer for the years 2007 and 2008, and that prior to the IRS audit for the tax years 2007 and 2008, Bayer had never been asked to identify the business components giving rise to claimed QRE credits. (Hearing Tr. 7/18/11, Docket No. 81, pp. 139, 150).

Response. In addition to its general objection, Bayer objects on the basis that this Interrogatory is overbroad and unduly burdensome without the adoption of a suitable sampling method. During the credit years, Bayer estimates that it developed more than 100,000 business components, which Bayer's books and records do not, and are not required to, track individually.¹⁵
(Docket No. 31–2).

15 As noted in Finding of Fact No. 9, Bayer uses a cost center system of accounting for research expenses rather than a project (or business component) based system. The cost centers at issue did not remain constant throughout the credit years. Cost centers were added, deleted or renumbered when Bayer's accounting system was modified or business units were sold, divided or closed. (Hearing Tr. 7/18/11, Docket No. 81, pp. 46–47).

17. On May 25, 2010, the Government filed a Motion to Compel Plaintiff to AnswerInterrogatory No. 26 (“Motion to Compel”). (Docket No. 30). Noting the Court's previously stated preference for the parties to develop a methodology to streamline this prodigious case, the Government indicated in its memorandum in support of the Motion to Compel that the parties “have a fundamental disagreement that affects their views on a suitable methodology,” and that it was necessary for the Court to resolve the disagreement to enable the parties to engage in discussions regarding such a methodology. With respect to Bayer's objection to identifying business components until the parties had adopted a sampling method, the Government contended that Bayer had it “exactly backwards.” ¹⁶ That is, until Bayer identifies the business components for which it claims the QRE credits at issue, it is impossible for the Government to determine if this case could be streamlined through some form of sampling.¹⁷ (Docket No. 31, pp. 9–10).

16 Bayer has never undertaken an analysis to determine whether it can establish a nexus between the QRE credits at issue and the business components generating the QREs. The parties disagree on Bayer's ability to do so. (Hearing Tr. 7/18/11, Docket No, 81, pp. 148–49).

17 In a footnote, the United States stated: “To be clear, the United States does not concede that sampling is permissible or appropriate; that said, unless Bayer identifies its business components, it will be impossible to determine if sampling is even workable.” (Docket No. 31, p. 10, n. 4).

18. By Order dated June 17, 2010, the parties were directed to be prepared to discuss discovery matters, including the appointment of a Special Master to oversee discovery, during a status conference scheduled for June 22, 2010. (Docket No. 40). Thereafter, on August 24, 2010, with the parties' consent, the Court appointed Thomas D. Arbogast, Esquire to serve as Special Master (“SM”) with regard to discovery disputes in this case. (Docket No. 50).

19. In a Joint Status Report filed October 13, 2010, the parties indicated that SM Arbogast heard argument on the Government's Motion to Compel on September 20, 2010. Based on discussions during the hearing and feedback from SM Arbogast, the Government agreed to “table” its Motion to Compel and to serve additional discovery requests on Bayer to better understand the scope of Bayer's claims and the organization of Bayer's records to enable the Government to evaluate any future sampling proposals, and Bayer agreed to reevaluate its response to Interrogatory No. 26 to determine if it could further clarify the response. (Docket No. 52).

20. Despite the Government's willingness to “table” its Motion to Compel, by Order dated October 26, 2010, the Court denied the Motion to Compel without prejudice to the Government's right to renew the motion at an appropriate time if necessary. (Docket No. 55).

21. By letter dated November 29, 2010, the Government proposed that the parties engage in a “pilot sample” to determine the feasibility of utilizing a sampling method to resolve this case. In its responsive letter dated December 15, 2010, Bayer rejected the idea of a pilot sample because “whether sampling is possible at all ... does not depend on the results of a prior mini-sample,” and a pilot sample “would require two separate discovery periods and two separate trials” and “consume substantial additional time and resources.” (Gov't Hearing Exh. 115 (Tab 15)).

22. On March 3, 2011, the Government renewed its Motion to Compel with SM Arbogast. In the transmittal letter, Government counsel stated: “After extensive discussions with Bayer, the United States has determined that it will not consent to sampling.” (Docket No. 54, No. 66, No. 71–3).

23. The next day, Bayer filed a Motion for a Case Management/Protective Order Based on Statistical Sampling (“Sampling Motion”) that would “enable the parties to conduct all of the necessary discovery and allow the Court to decide the entire case in the next two to three years.” (Docket No. 63). Bayer attached two documents to the Sampling Motion: (1) a memorandum dated March 4, 2002 from the IRS's Director of Field Specialists to the Directors of the IRS's Large and Mid–Sized Business Division regarding a Field Directive that provided guidance on the use of statistical sampling in audits of taxpayers' QRE credits, and (2) the Field Directive attached to the March 4, 2002 memorandum. (Docket No. 63–1, No. 63–2).

24. In the memorandum filed in support of its Sampling Motion, Bayer stated:

* * *

During the years at issue, Bayer's research spending exceeded $6 billion at 49 separate sites across the country. The research was performed by tens of thousands of individual Bayer employees. It consisted of millions of individual expenditures that were charged to more than 1300 cost centers. The vast scope of this enterprise is illustrated by the fact that Bayer has already collected more than one billion (1,000,000,000) pages of electronic records that are potentially relevant to its claims from just four of the forty-nine sites at issue and has already turned over more than 3 million pages of responsive documents to the government.

Given the massive size and scope of the activities and expenditures at issue, it is essential that the parties and the Court be able to focus their analysis on a manageable universe of information. If the parties were to try to conduct a detailed investigation of every one of the millions of expenditures, at every one of the forty-nine sites, for each of the more than twenty years at issue,¹⁸ discovery alone would require decades. Even if the parties were to depose all of the likely more than ten thousand relevant current and former Bayer employees located all across the country, it would be impossible to introduce more than a tiny fraction of their testimony at trial. Some form of sampling is absolutely essential to comply with the Court's directive that the parties find a way to streamline this case.
(Docket No. 64, p. 6).

18 With respect to the statement that “more than twenty years [are] at issue,” 1984 to 1988 is the base period for computing Bayer's QRE credits for 1990 through 2006. Thus, those years also are relevant in this case.

Bayer also stated that its proposed sampling method “is a framework that will govern the scope of discovery and the facts Bayer must prove at trial. The proposal does not, however, oblige the government or the Court to accept any of these facts as true. As usual in tax cases, it will be Bayer's responsibility to prove them.” (Docket No. 64, p. 3).

25. In a memorandum filed April 1, 2011, SM Arbogast denied the Government's Motion to Compel as premature and burdensome.¹⁹ Noting that Bayer maintains a cost center system of accounting for research activities, as opposed to a project based system, SM Arbogast denied the Motion to Compel “based upon the practical conclusion that compelling Bayer to identify all business components at this stage of the discovery process is not feasible because of the inherent limitations in Bayer's recordkeeping.” SM Arbogast cautioned, however, that “nothing in this Order should be read as excusing Bayer from disclosing to the United States during the course of discovery the business components associated with Bayer's identified research activities.” ²⁰ SM Arbogast emphasized that the denial of the Government's Motion to Compel did not equate with the imposition of statistical sampling upon the Government for discovery purposes, which he clearly viewed as an entirely separate issue and the subject of Bayer's pending Sampling Motion. (Docket No. 66).

19 Because the Government had not had the opportunity to respond to the Sampling Motion filed by Bayer, SM Arbogast declined to address the merits of the Sampling Motion at that time. (Docket No, 66, p. 3).

20 As noted previously, the Internal Revenue Code provides that the test for determining whether an expense was incurred in connection with qualified research is to be applied separately to each business component of the taxpayer. 26 U.S.C. § 41(d)(2)(A). Bayer does not dispute its obligation to identify the business components to which the claimed QRE credits apply. Due to its method of accounting, however, Bayer claims that it cannot do so without interviews of numerous researchers and extensive document collection.

26. On April 4, 2011, the Government filed an objection to SM Arbogast's ruling on its Motion to Compel based on Bayer's alleged failure to provide evidence supporting its claim that responding to Interrogatory No. 26 was unduly burdensome. (Docket No. 67).

27. Three days later, Government counsel sent a letter to SM Arbogast requesting clarification of the following statement in his ruling on the Motion to Compel:

However, nothing in this Order should be read as excusing Bayer from disclosing to the United States during the course of discovery the business components associated with Bayer's identified research activities, whatever form that discovery process takes, as that information is available.

28. On April 22, 2011, SM Arbogast responded to the letter of Government counsel and Bayer's response thereto as follows:

* * *

My proposed order was founded upon the practical conclusion that it would not be feasible, in light of the Court's instructions, to compel Bayer to disclose all of its business components over a 17 year period for all research sites. I did, however, purposely make clear that I viewed Bayer as obligated to disclose the business components supporting its claims for qualified research and that I would support a reasonable effort by the Defendant to secure identification of business components within the context of an efficient discovery plan.

The United States has now identified four large research sites that Bayer has visited (footnote omitted) and for which substantial amounts of electronic records have been collected and produced. On behalf of the United States Mr. Geht now requests that I clarify whether my March 15, 2011 Order requires Bayer to identify business components from these four major research sites or, alternatively, to clarify Bayer's current obligation with respect to Interrogatory Number 26 “in light of the [data] collection efforts undertaken by Bayer to date.”

In specific response to Mr. Geht's question, at this time I am aware of no outstanding discovery request, except for Interrogatory No. 26, the enforcement of which has been denied, which asks Bayer to identify its business components. If the United States wishes to tailor such a request, such as that suggested in Mr. Geht's letter (assuming a renewed objection to such a discovery demand as outlined in Mr. McIntyre's April 11, letter) your Special Master could again consider a revised motion to compel a response to a revised interrogatory. However, because of the alleged magnitude of data at these four my suggestion is that any renewed demand for a compilation of business component information be directed initially at specific costs centers for specific years at specific sites. Such a more limited request might enable the United States, and certainly the Special Master, to gain some understanding of the procedure by which business components can be identified under Bayer's “cost center” accounting system. Were Bayer to claim that such a limited discovery request presents an undue burden, I might be more inclined to issue an order compelling compilation and disclosure.

* * *

Gov't Hearing Exh. 106 (Tab 7).

29. With regard to the Government's assertion in its objection to SM Arbogast's ruling on the Motion to Compel Bayer to answer Interrogatory No. 26 that the denial lacked evidentiary support, Bayer noted in its brief in opposition that Mr. Wright was present during oral argument on the Government's initial Motion to Compel before SM Arbogast on September 20, 2010; that Mr. Wright was available for questioning; and that Mr. Wright did, in fact, answer questions regarding the burden presented by the Government's Interrogatory No. 26. (Docket No. 69, p. 3). Bayer also attached an affidavit of Mr. Wright to its opposition in which he attested to Bayer's estimates regarding the number of current and former Bayer employees who performed qualified research that is relevant in this case (10,000); the method by which Bayer accounts for its research activities (cost centers rather than individual projects); the amount of research expenses incurred by Bayer during the credit years ($6,000,000,000); the number of cost centers that are relevant in this litigation (1,300); the number of research sites at issue (49); and the number of documents that had been turned over in response to the Government's discovery requests (3,000,000). (Docket No. 69–2). On June 23, 2011, the Government's objection to SM Arbogast's ruling on the Motion to Compel was overruled by the Court. (Docket No. 77).

30. On May 3, 2011, Bayer filed the Amended Sampling Motion referred to in SM Arbogast's April 22, 2011 letter. (Docket No. 71). In addition to the two documents attached to its original Sampling Motion, Bayer attached the following documents: (1) an expert report by a statistician regarding a sampling proposal (Docket No. 71–1); (2) the Government's March 3, 2011 letter to SM Arbogast regarding its renewal of the Motion to Compel in which counsel stated that “[a]fter extensive discussions with Bayer, the United States has determined that it will not consent to sampling” (Docket No. 71–3); and (3) a declaration of Mr. Wright to which was attached a memo by Deloitte & Touche LLP regarding the methodology used in the 1997–1998 study of Bayer's QREs on which its claim for additional tax credits for the credit years is based (Docket No. 71–4, No. 71–5). In the memorandumin support of its Amended Sampling Motion, Bayer re-iterates its position that the proposed sampling method “is a framework that will govern the scope of discovery and the facts Bayer must prove at trial.” (Docket No. 72, p. 3).

31. In denying the Government's Motion to Compel on April 1, 2011, SM Arbogast stated, among other things: “Whether the Defendant's view on sampling ultimately prevails is yet to be determined, but in the first instance the government is free to take some other approach to compel the production of the data it believes is necessary to test Bayer's claims. Whether that request is broad or narrow is the government's call.” (Docket No. 66, p. 8). In accordance with this statement, as well as SM Arbogast's April 22, 2011 letter to counsel responding to the Government's request for a clarification of the ruling on the Motion to Compel, the Government served Bayer with a Fourth Set of Interrogatories on May 3, 2011 which included a similar, but much narrower, interrogatory to Interrogatory No. 26 which was the subject of its Motion to Compel. Specifically, Bayer was asked by the Government in Interrogatory No. 68 to identify the business components relating to QREs claimed for 3 cost centers at 4 of the research sites where document retrieval efforts had been undertaken for one year each. (Gov't Hearing Exh. 105 (Tab 6)).

32. Bayer responded to the Government's narrowed discovery request as follows:

Response. Objection. This interrogatory is premature and improper given the Plaintiffs' Amended Motion for a Case Management/Protective Order Based on Statistical Sampling dated May 3, 2011, which is presently pending before the Special Master. In that Motion, as amplified in the supporting Memorandum, Plaintiffs ask the Court to order sampling methodology that defines both the facts to be proven at trial and the scope of discovery. Until the Motion is decided, both are unknown. Fed.R.Civ.Pro. 26(c)(2) contemplates that when a motion for a protective order is made, the court rule on that motion before the discovery complained of is permitted to proceed. Responding to Defendant's latest discovery would require Plaintiff to forego the very relief it seeks in its pending Motion before the Court has rendered a decision thereon. If the protective order sought in Plaintiff's Motion is granted, then except for those documents that pertain to the sample elements chosen by the Court, the documents now sought by Defendant will be outside the scope of permissible discovery, as defined in Fed.R.Civ.Pro. 26(b). Defendant's Request is therefore objectionable....
(Gov't Hearing Exh. 105 (Tab 6)).²¹

21 During his deposition, Mr. Wright was asked whether Bayer included in its proposed sampling method the specific procedure by which the business components underlying Bayer's claimed QRE credits could be identified from the sampled data as suggested by SM Arbogast in his letter to counsel on April 22, 2011, Mr. Wright testified that Bayer had not. When asked “why not?”, Mr. Wright testified that “business component is common sense,” “it's not necessary.” Docket No. 73–1, p. 31 (Depo. p. 62).

33. On June 13, 2011, SM Arbogast sought guidance from the Court regarding his authority to rule on Bayer's Amended Sampling Motion. Specifically, SM Arbogast noted:

* * *

Bayer's initial and amended Sampling Motions request that the Court devise a discovery plan based upon statistical sampling and further requests that the results of the sample be extrapolated to the entirety of Bayer's claims without the requirement to introduce further evidence. Under Bayer's motion there would be no required proof of any underlying facts on claims or taxable years not examined as part of the statistical sample. The United States has consistently and vigorously opposed the idea that it can be forced to accept statistical sampling as a means of discovery or as a basis by which Bayer can meet its burden of proof at trial.

As stated in my May 2, 2011 Second Report my plan is to frame a report and recommendation on or before August 30, 2011 to resolve this discovery dispute. However, in light of the scope of the Bayer Amended Sampling Motion the issue is presented as to whether any proposed order should be limited to matters of discovery or may, in my discretion, outline a basis by which such discovery should govern the scope of facts which may be introduced at trial.

Both parties agree that unless I have the authority to issue a proposed (footnote omitted) order dealing with the evidence to be introduced at trial, I cannot fully dispose of Bayer's Amended Sampling Motion....

I have made no judgment at this stage of the proceedings as to whether Bayer's Amended Sampling Motion has merit or whether the Court has the power to compel the United States to accept such a proposal over its objection. The United States argues vigorously in the negative on both points. Further, the United States objects to the Special Master's seeking to propose any ruling over what evidence can be introduced at trial....

My objective is to seek clarification of whether I have authority from this Court to deal with all aspects of the Bayer Amended Sampling Motion. In order to fully address the issues raised by the Motion, both parties agree that more than discovery issues are involved, but apparently disagree whether the existing orders of this Court provide me with such authority. Thus, I seek guidance from you on this question.
(Docket No. 74).

34. A week later, the parties' filed a Joint Motion for Hearing by the Court which stated in parts

* * *

7. Bayer's Amended Sampling Motion asks that the Court issue an Order that would: a) compel the parties to randomly select 100 components of Bayer's claim for intensive analysis; b) limit discovery, with certain minor exceptions, to the 100 components selected; c) set a comprehensive schedule for the remainder of the case; and d) provide that the proof at trial will be limited to the evidence subject to discovery.

8. The parties agree that the Special Master has the authority to decide whether to order (a), (b), and/or (c) above.

9. Federal Rule of Civil Procedure 53 places certain limitations on the authority of Masters where the parties have not consented.

10. While Bayer has no objection to the Special Master deciding its Amended Sampling Motion in its entirety, the United States is not willing to provide its consent.

11. Because Bayer's Amended Sampling Motion asks the Court to establish certain procedures for trial in this matter (step (d) above), the parties have agreed that the Motion seeks relief beyond the scope of the authority granted to the Special Master under the appointment order and Rule 53.

12. The United States believes that if the Court decides that step (d) is premature at this point in the litigation, the Special Master can decide the remainder of the pending motion....

* * *

14. Bayer believes that its entire motion must be decided at this time because of its view that the evidence at trial must come from the same pool of evidence that will be examined in discovery. Without such an assurance at this point, the parties might spend the next two to three years in discovery only to learn that the trial would require the examination of a raft of other, previously unexplored evidence. Bayer believes that such a result would be remarkably inefficient and a misuse of judicial resources.

* * *

(Docket No. 75).

35. The parties' Joint Motion for a Hearing on Bayer's Amended Sampling Motion was granted and a hearing held on July 18, 19 and 20, 2011. SM Arbogast was present throughout the hearing. (Hearing Tr. 7/18/11, Docket No. 81, No. 82, No. 83).

Substantiation of Claimed QRE Credits

36. Luis Rodriguez is the lead Project Manager for Bayer Business & Technology Services LLC (“BBTS”). Since July 2009, Mr. Rodriguez has been assigned to Bayer's Tax Department to collect and preserve evidence relating to the QREs incurred by Bayer during the credit years. (Hearing Tr. 7/19/11, Docket No. 82, pp. 72–73, No. 78–1, ¶¶ 1–2).

37. To begin retrieval of documents to substantiate Bayer's claim for a refund in this case, Mr. Wright chose 5 of the 49 research sites generating QREs for the credit years. The sites include 4 current research sites, Berkeley, California (“Berkeley”), Kansas City, Missouri (“Kansas City”), Research Triangle Park in Raleigh, North Carolina (“RTP”) and Pittsburgh, Pennsylvania (“Pittsburgh”), and 1 research site that was closed in 2006, West Haven, Connecticut (“West Haven”).²² (Hearing Tr. 7/18–19/11, Docket No. 81, pp. 63–66, No. 82, p. 76). The 5 sites comprise 49% of the total QRE credits claimed by Bayer for the credit years (Berkeley–10%, Kansas City–6%, RTP–2%, Pittsburgh–7% and West Haven–24%), and at least 42% of the total cost centers generating QREs during the credit years (Berkeley–6%, Kansas City–?%, ²³ RTP–9%, Pittsburgh–15% and West Haven–12%). (Hearing Tr. 7/19/11, Docket No. 82, pp. 64–71).

22 For variety purposes, Mr. Wright selected 2 sites from Bayer Healthcare, 2 sites from Bayer CropScience and 1 site from Bayer MaterialScience. In addition, Mr. Wright selected these sites because “it also seemed to make sense to pick the bigger sites. Sites that no matter what decision is made, we're going to need documents from these five sites.” (Hearing Tr. 7/18/11, Docket No. 81, pp. 64–65).

23 During the cross-examination of Mr. Wright by Government counsel, counsel for Bayer objected to exhibits prepared by the Government to summarize a 300–page spreadsheet submitted to the IRS by Bayer in November 2010 in support of the QRE credits claimed in this case. The massive spreadsheet showed the amount of QREs claimed to have been incurred in each of the 49 research sites, as well as the number of cost centers in each of the 49 research sites that included expenses for qualified research. To eliminate Bayer's objection, the parties agreed that after Mr. Wright's review of the spreadsheet, Government counsel would be permitted to ask Mr. Wright additional questions regarding (a) the percentage of claimed QRE credits in each of the 5 research sites selected as the starting point for document collection and (b) the percentage of the 1,300 cost centers at issue in this case in each of the 5 selected research sites. (Hearing Tr. 7/18/11, Docket No. 81, pp. 112–15, 162). When Mr. Wright was called to testify in this regard the next day, another dispute arose between the parties regarding the Government's obvious inadvertent failure to specifically identify Kansas City as a site to be reviewed by Mr. Wright for purposes of his additional testimony. Ultimately, the parties resolved this dispute and stipulated to the percentage of the total QRE credits claimed in this case that were incurred at Kansas City (6%). If the parties also stipulated to Kansas City's percentage of total of cost centers generating QREs during the credit years, the stipulation was not read into the record. (Hearing Tr. 7/19/11, Docket No. 82, pp. 63–68, 71).

38. Mr. Rodriguez led the document collection teams during the site visits. The following periods of time were devoted to each site by Mr. Rodriguez and his teams: Berkeley–5 months; Kansas City–4M months; RTP–5 months; Pittsburgh–3 months (although document retrieval at the Pittsburgh site had not been completed at the time of the hearing); and West Haven–3 months.²⁴ (Hearing Tr. 7/19/11, Docket No. 82, pp. 144–46).

24 At this rate, Mr. Wright rendered the opinion that a reasonable estimation of the time it would take to collect the documents necessary to substantiate all of the QRE credits claimed by Bayer for the credit years is “decades.” (Hearing Tr. 7/18/11, Docket No. 81, p. 69). In contrast, Bayer contends that the entire case can be resolved in 2 to 3 years if a sampling methodology is utilized. (Docket No. 72, p. 3).

39. Of the 13,000 hours devoted to document retrieval for this litigation at the time of the hearing, approximately 6,000 hours related to site visits and 7,000 hours related to reviewing and preserving data on Bayer's last remaining main frame which was decommissioned in June 2011. (Hearing Tr. 7/19/11, Docket No. 82, p. 163).

40. Bayer offered testimony and/or documentary evidence to establish that at the time of the hearing, $3,632,790 had been incurred to collect documents for this litigation. (Hearing Tr. 7/18/11, Docket No. 81, p. 66). This sum consists of the following items: (1) payroll costs of approximately $1.6 million for work performed by employees of BBTS since July 2009; ²⁵ (2) $388,500 in hardware/software costs; (3) $499,935 for imaging services performed by Iron Mountain, a storage facility at which Bayer has stored 122,019 unindexed boxes of documents; ²⁶ (4) $370,355 for imaging services performed by Crivella West in connection with the West Haven site that was closed in 2006; ²⁷ and (5) outside legal fees of $694,000. (Hearing Tr. 7/19/11, Docket No. 82, pp. 135–41).

25 The hours devoted to document retrieval for this litigation by employees of BBTS is tracked by a system within the SAP Enterprise Accounting System called CATS. The exact figure for the cost of work performed by BBTS employees at the time of the hearing is $1,682,791. (Hearing, Bayer Exh. 12). This figure does not include the cost for work performed by employees of Bayer's tax and legal departments in connection with this litigation or the cost for Bayer researchers who have participated in document retrieval activities. (Hearing Tr. 7/19/11, Docket No. 82, p. 136).

26 Bayer Hearing Exh. 9.

27 Bayer Hearing Exh. 8.

41. Of the 49 research sites generating QREs during the credit years, the following 17 sites have been sold by Bayer: Addyston, OH; Akron, OH; Branchburg, NJ; Bushy Park, SC; Chicago, IL; Clayton, NC; Columbus, OH; Indian Orchard, MA; Medfield, MA; Middletown, VA; Orange, TX; Orangeburg, NY; Tarrytown, NY; Wilmington, MA; Newtown Square, PA; Trenton, NJ; and Walpole, MA.²⁸ (Hearing Tr. 7/18/11, Docket No. 81, pp. 116–18).

28 Although the exact number was not clear from the testimony and evidence admitted during the hearing, a number of the 49 research sites involved in this litigation also have been closed by Bayer.

42. When a site is sold by Bayer, financial and tax records are retained. In addition, the sales agreement includes a provision requiring the purchaser to retain records and cooperate with Bayer regarding tax issues and litigation. Mr. Wright concedes that it is more difficult to collect documents relating to this litigation from sold sites. However, he “still expect[s]” that Bayer will be able to do so. (Hearing Tr. 7/18/11, Docket No. 81, pp. 61–62, 75–77, Bayer Hearing Exhs. 4 and 5).

Bayer's Sampling Proposal and the Government's Response Thereto

43. Sampling is a scientific method developed over a hundred years ago that is designed to allow for the examination of a subset of units comprising a population in order to make inferences about the entire population through the use of probability. (Hearing Tr. 7/18/11, Docket No. 81, p. 184, Gov't Hearing Exh. 117 (Tab 17)).

44. Stephen E. Fienberg, a Maurice Falk University Professor of Statistics and Social Sciences at Carnegie Mellon University, was retained by Bayer to design a sampling method to assess the accuracy of the QREs claimed by Bayer for the credit years.²⁹ Professor Fienberg chose a statistically valid sampling methodology known as sampling with probability proportional to size (“PPS”) with replacement.³⁰ (Hearing Tr. 7/18/11, Docket No. 81, pp. 189–91). Using sampling with PPS, the likelihood of a particular research site being selected for analysis is directly proportional to the amount of QREs claimed at the research site.³¹ (Gov't Hearing Exh. 117 (Tab 17)). As to PPS “with replacement,” Professor Fienberg explained the concept as follows: “... each time I draw a site, I draw it independently of whatever happens in any other draw, and it's like picking a site and putting it back into the population of sites so that it could come up a second time.” (Hearing Tr. 7/19/11, Docket No. 82, pp. 9–10). PPS “with replacement” is utilized to ensure that sample selections are independent and that selection probabilities are constant for all draws. (Docket No. 72, p. 11).

29 In the 1960's, Professor Fienberg received a masters' degree and Ph.D. in statistics at Harvard University. He is a member of the National Academy of Sciences, an honorary society founded in 1863 to advise the Government on scientific issues. (Hearing Tr. 7/18/11, Docket No. 81, pp. 178, 181, Gov't Hearing Exh. 117 (Tab 17)). Professor Fienberg prepared an expert report and testified on Bayer's behalf at the hearing on the Amended Sampling Motion. (Hearing 7/18–19/11, Docket No. 81, pp. 177–97, No. 82, pp. 1–62).

30 Professor Fienberg based his proposed sampling methodology on lengthy discussions with Mr. Wright and Bayer's counsel concerning the problems presented by the scope of this litigation, a review of the Deloitte study and a visit to Bayer's Pittsburgh site where he met with the managers of several units who participated in the Deloitte study. (Hearing Tr. 7/18–19/11, Docket No. 81, p. 189, No. 82, pp. 30, 39).

31 For example, a research site with $100 million in QREs would be ten times as likely to be chosen for analysis as a research site with $10 million in QREs. (Docket No. 72, p. 11).

45. In summary, under Professor Fienberg's sampling methodology, 8 of the 49 research sites generating QREs during the credit years would be selected using sampling with PPS with replacement. Next, two of the credit years, i.e., 1990–2006, would be randomly selected for each of the 8 sample research sites using PPS with replacement. Next, 50 cost centers from the 8 sample research sites for the 2 sample years would be randomly selected using PPS with replacement for extensive analysis.³² (Gov't Hearing Exh. 117 (Tab 17)).

32 The 50 sample cost centers, each with 2 associated sample years (for a total of 100 sample units), would constitute the sample population and define the scope of discovery and evidence to be presented at trial. (Docket No. 72, p. 12).

46. Following completion of the sampling process and Bayer's presentation of evidence to support the QREs claimed in the 100 sample cost centers, the Court would determine (a) Bayer's entitlement to the QREs claimed in each of the 100 sample cost centers and (b) calculate, in percentage terms, the extent to which Bayer had proved the QREs claimed for the sample cost centers (“the observed percentage”).³³ The observed percentage would then be applied to the remaining cost centers at issue, i.e., the cost centers not included in the sample, without the submission of any further evidence for a final resolution of this dispute. (Docket No. 72, pp. 14–15).

33 This calculation would be performed by dividing the amount of QREs proven by Bayer for the sample cost centers by the total amount of claimed QREs for those cost centers. In effect, the calculation is a measure of the accuracy of Bayer's refund claim in this case.

47. To support its opposition to Bayer's proposed sampling methodology, the Government retained Joseph B. Kadane, a Leonard J. Savage University Professor of Statistics and Social Sciences, Emeritus at Carnegie Mellon University. ³⁴ Professor Kadane agrees with Professor Fienberg that (a) there is nothing about QREs that would preclude sampling in this case, and (b) sampling can produce more accurate information about a population than direct measurement of the entire population in certain circumstances.³⁵ However, if sampling were imposed on the Government in this case as requested by Bayer, Professor Kadane would support the use of a sampling plan that utilized unrestricted random selection of cost centers and credit years with PPS with replacement, rather than restricted as proposed by Professor Fienberg. (Hearing Tr. 7/20/11, Docket No. 83, pp. 6, 31, 41–42, 49).

34 Professor Kadane received a B.A. degree in mathematics cum laude from Harvard College in 1962 and a Ph.D. in Statistics from Stanford University in 1966. He has been elected a Fellow of the American Statistical Association and the Institute for Mathematical Statistics and a member of the International Statistical Institute and the American Academy of Arts and Sciences. Professor Kadane prepared an expert report and testified on the Government's behalf at the hearing. (Hearing Tr. 7/20/11, Docket No. 83, pp. 5–58, Gov't Hearing Exh. 110 (Tab 10)). Professors Fienberg and Kadane are colleagues at Carnegie Mellon University. In fact, Professor Kadane hired Professor Fienberg. (Hearing Tr. 7/20/11, Docket No. 83, p. 7).

35 Professors Fienberg and Kadane also agree that the accuracy of a sampling plan cannot be assessed until the sampling plan is completed. (Hearing Tr. 7/20/11, Docket No. 83, pp. 6, 31, 41–42, 49). Bayer proposes that the final determination of its total QREs for the credit years be determined from the sample mean which Professor Fienberg testified is “unbiased” and “the most likely estimate based on repeated sampling.” (Hearing Tr. 7/19/11, Docket No. 82, p. 36). On the other hand, Professor Kadane opines that if a sampling plan is imposed on the Government, the final determination of Bayer's total QREs for the credit years should be based on the 95% one-sided confidence limit that is least advantageous to Bayer because the use of sampling to establish the QRE credits at issue in this case is being proposed by Bayer. (Hearing Tr. 7/20/11, Docket No. 83, p. 14–15). In light of the Court's conclusion that Bayer's Amended Sampling Motion must be denied, it is not necessary to address this dispute.

48. Professor Kadane recommends the use of a pilot sample which he describes as a “dress rehearsal” for a full sampling plan that may facilitate a settlement of this dispute by “confronting” the parties with facts. (Hearing Tr. 7/20/11, Docket No. 83, pp. 31–33). For a pilot sample, Professor Kadane recommended the random selection of 5 to 10 cost center/years utilizing PPS with replacement. Bayer would then collect the evidence necessary to establish the QREs in those cost center/years and the Court would determine the extent to which Bayer proved the QREs. After the Court's determination, the parties would know what evidence was necessary to prove the claimed QREs and the parties could decide on the manner in which they wished to proceed. Professor Kadane could not estimate the time or cost of performing a pilot sample which would resolve only a small portion of Bayer's refund claim in this case.³⁶ (Hearing Tr. 7/20/11, Docket No. 83, pp. 36–37, 51).

36 Prior to rendering his opinions concerning Bayer's proposed sampling plan and the benefits of a pilot sample, Professor Kadane did not review Bayer's accounting system, the Deloitte study or the manner in which Bayer maintains records; he did not visit any of Bayer's research sites; and he never spoke with any employee of Bayer about the claims in this case. (Hearing Tr. 7/20/11, Docket No. 83, p. 53).

CONCLUSIONS OF LAW

1. Whether and to what extent tax deductions are allowed are a matter of legislative grace, and a taxpayer claiming a deduction must be able to point to an applicable statute and show that he meets all of the requirements. New Colonial Ice Co. v. Helvering, 292 U.S. 435, 440, 54 S.Ct. 788, 78 L.Ed. 1348 (1934). See also National Starch and Chemical Co. v. Commissioner of Internal Revenue, 918 F.2d 426 (3d Cir.1990), citing, E.I. du Pont de Nemours and Co. v. United States, 432 F.2d 1052, 1059 (3d Cir.1970) (The burden is on the taxpayer to show that expenses are deductible). Moreover, provisions granting special tax exemptions are to be strictly construed. Helvering v. Northwest Steel Rolling Mills, 311 U.S. 46, 49, 61 S.Ct. 109, 85 L.Ed. 29 (1940). See also Galloway v. United States, 492 F.3d 219, 223 (3d Cir.2007) (Courts are admonished to strictly construe tax deductions and allow such deductions only as there is a clear provision therefor).

2. Despite the foregoing well established principles of tax law, Bayer's Amended Sampling Motion seeks an Order from the Court that would eliminate entirely its burden of proof with regard to all QREs claimed at 41 of the 49 research sites during the credit years, as well as the QREs not selected for analysis in the 8 sample research sites, over the Government's objection. Simply put, the Court can find no authority for the extraordinary relief sought by Bayer.

3. The Court acknowledges that Rule 1 of the Federal Rules of Civil Procedure provides that the Rules “should be construed and administered to secure the just, speedy, and inexpensive determination of every action and proceeding,” and that the Supreme Court has stated that “the discovery provisions, like all of the Federal Rules of Civil Procedure, are subject to the injunction of Rule 1.” See Herbert v. Lando, 441 U.S. 153, 177, 99 S.Ct. 1635, 60 L.Ed.2d 115 (1979). The Court further acknowledges that “Rule 26(c) grants federal judges the discretion to issue protective orders that impose restrictions on the extent and manner of discovery where necessary ‘to protect a party or person from annoyance, embarrassment, oppression, or undue burden or expense.’ ” Pearson v. Miller, 211 F.3d 57, 65 (3d Cir.2000). Finally, the Court acknowledges that “sampling has long, been considered an acceptable method of determining the characteristics of a large universe.” Rosado v. Wyman, 322 F.Supp. 1173, 1180 (E.D.N.Y.1970), aff'd, 437 F.2d 619 (2d Cir.), aff'd, 402 U.S. 991, 91 S.Ct. 2169, 29 L.Ed.2d 157 (1971). Nevertheless, the Court concludes that Bayer is not entitled to the relief requested in the Amended Sampling Motion.

4. First, as noted in Finding of Fact No. 5, the recordkeeping requirements applicable to Bayer's claim for QRE credits during the credit years mandate that Bayer “retain records in sufficiently usable form and detail to substantiate that the expenditures claimed are eligible for the credit.” Bayer's estimate that it will take “decades” ³⁷ to gather the evidence necessary to support the QREs claimed for the credit years establishes Bayer's failure to comply with its recordkeeping obligation. The Court agrees with the Government that granting the relief sought in Bayer's Amended Sampling Motion would, in effect, constitute a reward to Bayer for failing to keep evidence regarding research expenses in “sufficiently usable form and detail.” (Docket No. 73, p. 18).

37 Hearing Tr. 7/18/11, Docket No. 81, p. 69.

5. In Kozlowski v. Sears, Roebuck & Co., 73 F.R.D. 73 (D.C.Mass.1976), a products liability suit, plaintiff filed a request for production of documents relating to complaints concerning personal injuries or death allegedly caused by the burning of children's nightwear which had been manufactured by defendant. Defendant filed a motion to quash and plaintiff filed a motion to compel. Defendant's motion was denied; plaintiff's motion was granted; and Defendant was ordered to produce the documents within 30 days. When defendant failed to produce the documents as directed, plaintiff moved for, and was granted, default judgment on the issue of liability. The district court conditioned removal of the default judgment on defendant's full compliance with the discovery order. Defendant moved to vacate the judgment by default; however, the motion was denied. The following section of the district court's opinion in Kozlowski is applicable in this case:

* * *

Under Rule 34, Fed.R.Civ.P., the party from whom discovery is sought has the burden of showing some sufficient reason why discovery should not be allowed, once it has been determined that the items sought are properly within the scope of Rule 26(b), Fed.R.Civ.P. See Wright & Miller, Federal Practice & Procedure: Civil § 2214, p. 644 (1970). Merely because compliance with a “Request for Production” would be costly or time-consuming is not ordinarily sufficient reason to grant a protective order where the requested material is relevant and necessary to the discovery of evidence. Luey v. Sterling Drug, Inc., 240 F.Supp. 632, 634–5 (W.D.Mich.1965).

In the instant case, the requested documents are clearly within the scope of Rule 26(b), Fed.R.Civ.P., the plaintiff has a demonstrable need for the documents, the defendant undisputedly has possession of them, and the plaintiff has no other access to them. Thus, the defendant has a duty pursuant to Rule 34, Fed.R.Civ.P., to produce its records of similar suits. The defendant seeks to absolve itself of this responsibility by alleging the herculean effort which would be necessary to locate the documents. The defendant may not excuse itself from compliance with Rule 34, Fed.R.Civ.P., by utilizing a system of recordkeeping which conceals rather than discloses relevant records, or makes it unduly difficult to identify or locate them, thus rendering the production of the documents an excessively burdensome and costly expedition. To allow a defendant whose business generates massive records to frustrate discovery by creating an inadequate filing system, and then claiming undue burden, would defeat the purposes of the discovery rules. See Hickman v. Taylor, 329 U.S. 495, 500, 67 S.Ct. 385, 91 L.Ed. 451 (1947); Holtzoff, Instruments of Discovery Under Federal Rules of Civil Procedure, 41 Mich.L.Rev. 205, 224 (1942).

* * *

73 F.R.D. at 76.

Similarly, in the present case, the discovery sought by the Government is within the scope of Fed.R.Civ.P. 26(b); the Government has a demonstrable need for the discovery; the discovery is in Bayer's possession; the Government has no other means to obtain the discovery; and responsibility for the burden presented by producing the discovery falls squarely on Bayer.

6. Second, although Bayer established that its employees have spent considerable time and it has incurred significant expenses in searching for evidence to support the QREs at issue in this case, the Court is not persuaded that the time spent and expenses incurred are “undue.” Of the 13,000 hours spent on document retrieval efforts at the time of the hearing, 7,000 hours were devoted to the one-time project of reviewing and preserving data from Bayer's last remaining main frame at the time it was decommissioned.³⁸ As to the 6,000 hours devoted to evidence retrieval efforts from 5 of the 49 research sites generating QREs that are relevant in this case, the 5 research sites are, or were, among Bayer's largest research sites, comprising 49% of the claimed QREs and at least 42% of the cost centers at issue. See Finding of Fact No. 37. It is simply unbelievable that the amount of time devoted to these research sites will be necessary at each of the remaining research sites many of which are, or were, much smaller.

38 Regarding the review and retention of data on Bayer's last remaining main frame, Mr. Rodriguez testified that 20 individuals worked on the project full-time. (Hearing Tr. 7/19/11, Docket No. 82, p. 98). According to the Court's calculation, assuming a 40–hour week, this project took approximately 9 weeks to complete-not an overly burdensome period of time for a one-time project.

7. Third, the court must limit the frequency or extent of discovery otherwise allowed by the Federal Rules of Civil Procedure if it determines, among other things, that “the burden or expense of the proposed discovery outweighs its likely benefit, considering the needs of the case, the amount in controversy, the parties' resources, the importance of the issues at stake in the action, and the importance of the discovery in resolving the issues.” Fed.R.Civ.P. 26(b)(C)(iii). In the present case, Bayer estimates that $175,000,000 is at stake; Bayer has not claimed, and there is no evidence of, insufficient resources to retrieve the evidence from which the business components underlying all of the QRE credits claimed for the credit years can be identified; the issues at stake clearly are important; and Bayer's identification of all business components generating QREs during the credit years is critical to proving its refund claim. Under the circumstances, the Court cannot conclude that the burden of the discovery sought by the Government outweighs its benefit.

8. Bayer notes, and the Court acknowledges, that sampling methodologies have been utilized in tax cases. However, the tax cases cited by Bayer in support of the Amended Sampling Motion are distinguishable from the present case.

9. The first type of tax case cited by Bayer involves the IRS's use of a formula based on seized evidence to reconstruct a taxpayer's unreported income from illegal activity. In Gerardo v. Commissioner, 552 F.2d 549 (3d Cir.1977), the IRS levied an assessment against the taxpayer for unreported gambling income in 1966 and 1967. To compute the assessment, the IRS calculated the average daily gross receipts from the betting slips comprising the daily tallies for 3 days' operation and projected that amount over the period of time covered by the assessment. The taxpayer appealed the Tax Court's determination of deficiencies in his income tax due for 1966 and 1967 based on the unreported income from illegal gambling activity. The taxpayer contended, among other things, that the formula used by the IRS to calculate the deficiencies in income was arbitrary and without foundation. The Court of Appeals for the Third Circuit rejected this argument stating:

“Since appellant kept no records, ... ‘it was proper and indeed necessary to devise some substitute method for reconstructing income.’ See Agnellino v. Commissioner, 302 F.2d 797, 799 (3d Cir.1962). Where unreported income from gambling is at issue, the projection of average daily gross receipts over a period of time to calculate gross income is an acceptable method of reconstruction. (citations omitted). Furthermore, Gerardo failed to sustain his burden of producing evidence which would have obviated the use of the Commissioner's formula. Therefore, we have concluded that the Tax Court did not err in approving the Commissioner's method of reconstruction.”
552 F.2d at 552, fn. 6.

A decision upholding the IRS's use of formula to reconstruct unreported illegal income due to the taxpayer's failure to meet his burden of producing evidence to establish his income simply does not support Bayer's request to be relieved from its burden of proving a substantial portion of the QREs claimed for the credit years due to its failure to maintain voluminous research expense records in “sufficiently usable form.” Further, there is no indication in Gerardo that the taxpayer would have been precluded from presenting evidence to rebut the IRS's projection of his income which is precisely what Bayer is seeking in this case.

10. Turning to the Supreme Court's decision in United States v. Janis, 428 U.S. 433, 96 S.Ct. 3021, 49 L.Ed.2d 1046 (1976), the taxpayer was arrested for bookmaking activity after the seizure of wagering records and $4,940 in cash during the search of an apartment pursuant to a warrant. Based on the arresting officer's representation that he had conducted surveillance of the taxpayer's activities which indicated the taxpayer had been involved in bookmaking activity during the 77–day period preceding his arrest, the IRS assessed the taxpayer for wagering taxes. The amount of the assessment was calculated by determining the taxpayer's average daily gross proceeds for the 5–day period covered by the seized wagering records, and then multiplying that amount by 77, the period of the police officer's surveillance of the taxpayer's bookmaking activities. Subsequently, the IRS levied on the $4,940 in cash that had been seized pursuant to the search warrant in partial satisfaction of the assessment against the taxpayer. After the search warrant was determined to be defective and quashed, the taxpayer filed a civil claim for refund of the $4,940. The taxpayer did not challenge the method used by the IRS to compute the assessment for wagering taxes, i.e., extrapolation of the taxpayer's average daily gross proceeds to the total number of days the police officer observed bookmaking activities.Rather, the taxpayer challenged the legitimacy of an assessment based solely on documents seized during an illegal search.³⁹ Thus, Janis provides even less support for the relief sought in Bayer's Amended Sampling Motion.

39 The Supreme Court held in Janis that the rule excluding evidence obtained in violation of Fourth Amendment rights would not be extended to exclude from a federal civil tax proceeding, in the absence of any proof of federal participation in the illegality, evidence obtained by a state criminal law enforcement officer in good faith reliance on a warrant later proved to be defective.

11. In the second type of tax case cited by Bayer in support of the Amended Sampling Motion, estimation of tax credits was approved. However, a close reading of those cases shows that the taxpayers were not relieved of their burden of providing evidence to support the tax credits prior to estimation. See United States v. McFerrin, 570 F.3d 672 (5th Cir.2009) (“If McFerrin can show activities that were “qualified research,” then the court should estimate the expenses associated with those activities.”); Cohan v. Commissioner of Internal Revenue, 39 F.2d 540 (2d Cir.1930) (Tax Board erred in failing to estimate entertainment expenses, even though amount may be trivial and unsatisfactory, in light of its finding that taxpayer had spent much and the sums were allowable expenses); Fudim v. Commissioner of Internal Revenue, T.C. Memo. 1994–235 (U.S.Tax Ct.1994) (After determining based on evidence presented by petitioner that he “no doubt” engaged in qualified research during 1986, 1987 and 1988, tax court estimated time worked on such research by petitioner and his wife based on petitioner's testimony and other evidence in the record. Due to insufficient information concerning the work performed by his daughter, however, petitioner was not entitled to any research credit based on the wages he paid her).

12. In the present case, the Government disputes Bayer's ability to meet its burden of identifying the business components for which the QREs claimed for the credit years were incurred, a requirement clearly set forth in Section 41 of the Internal Revenue Code. Until this burden is satisfied, quantifying the amount of QRE credits to which Bayer is entitled is premature. See also Oates v. Commissioner of Internal Revenue, 316 F.2d 56 (8th Cir.1963) (“We believe that considerable discretion exists in the application of the Cohan rule, and that such rule should be applied only in cases where the taxpayer has clearly shown that he is entitled to some deduction and that uncertainty exists only as to the exact amount thereof.”).⁴⁰

40 The Court also notes a recent tax refund suit in which the district court declined to estimate the QRE credits to which the taxpayer was entitled, despite the fact that the taxpayer provided evidence of qualified research activities. The district court in Trinity Industries, Inc. v. United States of America, 691 F.Supp.2d 688 (N.D.Tex.2010), stated in relevant part:
“The Court is aware of case law instructing it to estimate the amount of QRE if it determines that the taxpayer has made some qualified expenditure: ‘If the taxpayer can establish that qualified expenses occurred, however, then the court should estimate the allowable tax credit.’ United States v. McFerrin, 570 F.3d 672, 675 (5th Cir.2009) (citing Cohan v. Comm'r, 39 F.2d 540, 544 (2d Cir.1930)). In this case, however, Trinity did not offer any evidence from which the Court could make a meaningful estimate. The Court, therefore, finds there is no evidence from which it can estimate QREs relating to any business component smaller than an entire vessel.”

691 F.Supp.2d at 693.

See also Williams v. United States, 245 F.2d 559 (5th Cir.1957) (“That the trier, ..., might have considerable latitude in making estimates of amounts probably spent [for entertainment expenses] in light of accepted practice amongst law-abiding businessmen of moral standing considering the nature and kind of records which might reasonably be kept for such expenditures, Cohan v. Commissioner, 2 Cir., 39 F.2d 540, 543, certainly does not require that such latitude be employed. The District Court may not be compelled to guess, or estimate ... For the basic requirement is that there be sufficient evidence to satisfy the trier that at least the amount allowed in the estimate was spent or incurred for the stated purpose. Until the trier has that assurance from the record, relief to the taxpayer would be unguided largesse.”).

13. In the third type of tax case cited by Bayer in support of the Amended Sampling Motion, the Government agreed to a trial of less than all of the projects for which QRE credits were claimed. See Union Carbide Corp. and Subsidiaries v. Commissioner of Internal Revenue, T.C. Memo. 2009–50 (U.S.Tax Ct.2009) (Petitioner claimed additional QRE credits based on 106 projects conducted in various units within 6 manufacturing plants during 1994 and 1995. To resolve the action expeditiously, the parties agreed to try 5 of the largest projects underlying petitioner's additional QRE claims). Unlike the situation presented in Union Carbide, the Government has not agreed to consider less than all of the research sites generating the QRE credits claimed for the credit years to resolve Bayer's refund claim. In fact, the Government vigorously opposes Bayer's proposed sampling plan, and the Court can find no authority for compelling it to accept such a procedure. Further, there is no indication in the Union Carbide case that the taxpayer was excused from proving the QRE credits claimed for the projects not addressed at trial. Specifically, the Tax Court instructed the parties to resolve any issues regarding the QRE credits claimed for the other projects in a manner consistent with the Court's opinion.

14. As noted by the Government, attempts by taxpayers to impose sampling on the Government in tax cases have been rejected in the past.⁴¹ See, e.g., United States v. Helms, No. 08 CV 151, 2010 WL 3384997 (S.D.Cal.8/26/2010) (Tax assessments are entitled to a presumption of correctness, and burden is on taxpayer to rebut the presumption by producing countervailing evidence that assessments are in error. HELD: Taxpayer could not defeat summary judgment as to a whole category of deductions by using “samples.”); Kikalos v. Commissioner of Internal Revenue, No. 11486–01, 2004 WL 565161 (U.S. Tax Ct. 3/23/2004) (Taxpayers received unreported income in coupon and buy-down payments from cigarette company. HELD: Taxpayer could not use sampling to demonstrate that coupons and buy-downs were completely recorded); Lee v. Commissioner of Internal Revenue, Nos. 4498–94, 4503–94, 1995 WL 750122 (U.S. Tax Ct. 12/19/1995) (In taxpayer challenge to federal income tax deficiency determinations, taxpayer's use of sampling to estimate gross profit percentage was rejected where the Government's expert examined all underlying sales and purchase journals).

41 Docket No. 73, p. 19.

15. Bayer also cites the decision by the Court of Appeals for the Ninth Circuit in Ratanasen v. State of California, Dept. of Health Servs., 11 F.3d 1467 (9th Cir.1993), in support of the Amended Sampling Motion. Again, the Court finds the case distinguishable from the present case.

16. In Ratanasen, a Chapter 11 debtor-doctor filed an objection to the allowance of a claim by the State of California in the bankruptcy proceeding because the State's claim was based on an audit using sampling that revealed the debtor-doctor had overbilled the Medi–Cal program. The bankruptcy court concluded that the State's use of a random sample to prove the amount overbilled was valid and the debtor-doctor appealed. The district court affirmed the decision of the bankruptcy court, and the debtor-doctor filed a further appeal. In affirming the decision of the district court, the Ninth Circuit stated in relevant part: “We now join other circuits in approving the use of sampling and extrapolation as part of audits in connection with Medicare and other similar programs, provided the aggrieved party has an opportunity to rebut such evidence.” 11 F.3d at 1471 (emphasis added). In the Amended Sampling Motion, Bayer specifically seeks to preclude the Government from offering at a trial of this matter any evidence not produced as part of the sampling plan proposed by Professor Fienberg. Ratanasen does not provide support for such preclusion.⁴²

42 Bayer cites another distinguishable auditing case in support of its Amended Sampling Motion. Policy of the Department of Health and Human Services (“HHS”) allowed post-payment sampling audits of suspected Medicare overpayments based on the Secretary's conclusion that sampling provided the only feasible means for protecting the Medicare Trust Fund in situations where a provider is suspected of overbilling and the number of claims involved is large. In Chaves County Home Health Service, Inc. v. Sullivan, 931 F.2d 914 (D.C.Cir.1991), the amount of the plaintiff-provider's overpayment liability was calculated from the percentage of claims denied in a sample. A challenge to the sampling procedure by the plaintiff-provider was rejected by the Court of Appeals for the District of Columbia which held that the sampling audit procedure did not violate the Medicare Act or procedural due process. In so holding, the Court of Appeals noted that providers were given the same opportunity to challenge determinations regarding sample claims as that provided on pre-payment review, and, in case of any incorrect determinations, the overcharge projection would be correspondingly reduced. In contrast, Bayer's proposed sampling plan would preclude the Government from challenging any QREs that were not generated by the sample cost centers.

17. In support of the Amended Sampling Motion, Bayer also cites class action cases in which sampling methods were utilized. Again, the Court concludes that the cases are distinguishable from the present case. For example, in Long v. Trans World Airlines, Inc., 761 F.Supp. 1320 (N.D.Ill.1991), airline employees who had been replaced while on strike filed a class action to challenge the airline's failure to provide them with “designated rights” letters showing that they were eligible for first-hire rights under the employee protection provisions of the Airline Deregulation Act. Following the entry of summary judgment in favor of the employees on the issue of liability, a dispute arose concerning the propriety of using a sampling method to limit discovery on damages. Plaintiffs argued that full-blown damages discovery was unnecessary and unduly burdensome, while defendant argued that the presence of individual issues entitled them to discovery from each plaintiff. After reviewing the law concerning discovery and the law governing class actions, the district court agreed with plaintiffs that a sampling method should be used. Unlike Long in which defendant's liability was established and only the determination of damages remained, Bayer has not established its entitlement to the claimed QRE credits because it has failed to identify the business components generating QREs during the credit years. In fact, although Bayer represented in its objection to the Government's Interrogatory No. 26 that more than 100,000 business components were produced during the credit years and Mr. Wright contends that identification of business components is “common sense,” Bayer has refused for some unknown reason to identify a single business component supporting its refund claim in the almost two years this civil action has been pending.

18. Similarly, in Hilao v. Estate of Ferdinand Marcos, 103 F.3d 767 (9th Cir.1996), a class action was brought by Philippine nationals against the estate of the former Philippines president seeking damages for human rights abuses committed against them or their decedents. The issues of liability and damages were bifurcated for trial. After a jury reached a verdict in favor of the class and against the estate on the issue of liability, the district court allowed the use of a statistical sample of class claims to determine compensatory damages. Subsequently, the Court of Appeals for the Ninth Circuit rejected a due process challenge by the estate to the sampling methodology utilized by the district court to determine compensatory damages.⁴³ As in Long, statistical sampling was not utilized to establish liability which is essentially what Bayer is seeking in the Amended Sampling Motion with regard to a substantial portion of the QRE credits claimed for the credit years.

43 As noted by the Government, the Third Circuit also has noted in the class action context that liability must be established before a sampling method to establish damages can be considered. See Steamfitters Local Union No. 420 Welfare Fund v. Philip Morris, Inc., 171 F.3d 912, 929–930 (3d Cir.1999) (“What is more, in proposing a solution to the speculative nature of their damages—i.e., using aggregation and statistical modeling to measure damages—the Funds focus too far down the road to assist their case for standing: The task of accurately measuring damages can be approached only after a plaintiff has met the requirements for standing and has proven liability.”).

19. The Court agrees with the Government that this case is more analogous to the situation presented in United States ex rel. Fry v. Guidant Corp., Civ. No. 3:03–0842, 2009 WL 3103836 (M.D.Tenn.9/24/2009). Fry was a qui tam action brought pursuant to the False Claims Act, 31 U.S.C. § 3730. The United States claimed that defendant systemically and intentionally concealed from hospitals and clinics the availability of certain warranty and replacement credits due upon the replacement of implantable cardiac devices manufactured and sold by defendant, which caused hospitals and clinics to submit to the United States false and overstated claims for reimbursement following replacement procedures. Defendant filed a motion to compel the United States to produce, among other things, the cost reports submitted to the United States that contained false or inflated claims. In response, the United States sought a protective order permitting it to respond to defendant's discovery request by providing a statistically valid random sample of the hospital cost reports at issue based on a claim of undue and unnecessary burden. In granting defendant's motion to compel production of the cost reports and denying the Government's motion for a protective order, the court stated:

* * *

... None of the cited cases directly addresses the issue presented here-whether production in discovery of a sample of the allegedly false claims is sufficient in a case in which these claims are “critical to assessing both liability and damages.”

Because these cost reports form the basis of a False Claims Act case and are clearly relevant to a determination of liability and damages, the undersigned Magistrate Judge finds that the government should produce in discovery all hospital cost reports containing allegedly false claims for which it seeks to recover in this case....

* * *

2009 WL 3103836, at *2.

Similarly, because identification of business components is critical to establishing Bayer's entitlement to QRE credits for the credit years, a sampling plan that would not identify all of the business components underlying the claimed QRE credits is not acceptable in the absence of agreement by the Government.

20. In sum, the Court is not persuaded by Bayer's arguments that denial of the Amended Sampling Motion would render the credit for QREs set forth in Section 41 of the Internal Revenue Code “illusory” and “defeat the clear legislative purpose of the research credit.” (Docket No. 72, p. 5). As noted by the Government, Bayer's arguments regarding the burden presented to large research corporations by the recordkeeping requirements applicable to Section 41 of the Internal Revenue Code should be directed to the Legislative Branch, not this Court. (Docket No. 84, p. 28, ¶ 47).

* * *