Thinking MacHines Corp. v. New Mexico Taxation & Revenue Department

MEMORANDUM AND ORDER YOUNG, District Judge. This bankruptcy appeal presents a question of law which has yet to be addressed by the First Circuit and has split the other federal circuit courts of appeals: which party bears the ultimate burden of proof when a tax claim is asserted in the context of a bankruptcy proceeding? I. BACKGROUND Thinking Machines Corporation (“Thinking Machines”) appeals the decision of the Bankruptcy Court awarding the New Mexico Taxation and Revenue Department (the “Department”) a pre-petition tax claim on a tax base of $2,158,042.27. See In Re Thinking Machines Corp., 203 B.R. 1 (Bankr.D.Mass.1996) (Hillman, J.). 1 Thinking Machines argues that the Bankruptcy Court erred 1) by placing the burden of proof on the taxpayer to disprove the validity of the Department’s tax claim; 2) in concluding that Thinking Machines’ computer maintenance agreements with New Mexico residents were subject to New Mexico’s “gross receipts” tax; and 3) in calculating what percentage of services under those contracts was performed out-of-state, and thus exempt from the gross receipts tax. *428 Thinking Machines is a Massachusetts corporation engaged in the manufacture, sale, and servicing of computers. Between 1988 and 1991, it sold computer products and services to a number of government laboratories located in New Mexico pursuant to contracts referred to as “Maintenance Agreements.” These agreements called for Thinking Machines to provide full-service maintenance and support for its computers, including 1) hardware support (i.e., provision and installation of replacement parts, as well as repair of existing parts), 2) software licensing and support (i.e., debugging, provision of upgrades), and 3) an on-site applications engineer to troubleshoot and help customers get their own applications running in the Thinking Machines environment. In May of 1992, the Department conducted an audit of Thinking Machines for the period of January 1, 1988 through March 31, 1992, and issued an assessment for outstanding “gross receipts” taxes, see N.M.Stat.Ann. § 7-9-4(A), interest, and penalties on a tax base of $3,259,291.00. On August 17, 1994, Thinking Machines filed a Chapter 11 bankruptcy petition, and on September 21, 1994, the Department filed a proof of claim pursuant to 11 U.S.C. § 504 for the amount of the assessment. Thinking Machines filed a timely objection to the Department’s claim on November 27, 1995, arguing that the Maintenance Agreements predominantly involved sales of tangible personal property to organizations possessing certain nontaxable transaction certificates, and were therefore exempt from New Mexico’s gross receipts tax. See N.M.Stat.Ann. § 7-9-60. The Department, in contrast, contends that the sale of services was predominant throughout the Maintenance Agreements, thus rendering these Agreements fully taxable. Under New Mexico law, the “predominant ingredient” test is used to determine whether an activity constitutes the provision of a service or a sale of tangible property for purposes of the gross receipts tax. See EG & G, Inc. v. Director, Revenue Division Taxation and Revenue Dep’t, 94 N.M. 143, 607 P.2d 1161, 1163-64 (App.), cert. denied, 94 N.M. 628, 614 P.2d 545 (1979). After an evidentiary hearing, the Bankruptcy Court overruled Thinking Machines’ objection to the Department’s tax claim. The Bankruptcy Court held that Thinking Machines 1) had the ultimate burden of proving that the Department’s tax assessment was incorrect, 2) failed to satisfy its burden of proving that the revenues generated from its Maintenance Agreements were derived predominantly from sales of tangible property, and 3) failed to introduce adequate evidence of the proper allocation of out-of-state service costs. Relying upon one of the allocation models submitted by the Department, the Bankruptcy Court awarded the Department a pre-petition tax claim on a tax base of $2,158,042.27. II. ANALYSIS A. Burden of Proof The Bankruptcy Court’s ruling that, even in bankruptcy, the ultimate burden of proof rests with the taxpayer to disprove the validity of a tax claim, is a legal conclusion that this Court must review de novo. See Fed. R.Bankr.P. 8013; In Re Winthrop, 50 F.3d 72, 73 (1st Cir.1995). Outside of the bankruptcy forum, “[i]t is settled law that taxpayers bear the burden of proving that a tax deficiency assessment is erroneous.” Delaney v. Commissioner, 99 F.3d 20, 23 (1st Cir.1996) (citing United States v. Rexach, 482 F.2d 10, 16 [1st Cir.], cert. denied, 414 U.S. 1039, 94 S.Ct. 540, 38 L.Ed.2d 330 [1973]). “This rule is supported by ‘the presumption of administrative regularity; the likelihood that the taxpayer will have access to the relevant information; and the desirability of bolstering the record-keeping requirements of the Code.” ’ Thinking Machines, 203 B.R. at 3 (quoting Rexach, 482 F.2d at 16). Subject to certain exceptions, the United States Bankruptcy Code confers jurisdiction upon the bankruptcy courts to “determine the amount or legality of any tax” assessed against the debtor. 11 U.S.C. § 505. The Bankruptcy Code and the Federal Bankruptcy Rules of Procedure are silent, however, with respect to the proper allocation of the ultimate burden of proof *429 when a tax claim is asserted in bankruptcy, see In re MacFarlane, 83 F.3d 1041, 1045 (9th Cir.1996), cert. denied, — U.S. -, 117 S.Ct. 1243, 137 L.Ed.2d 326 (1997); In re Landbank Equity Corp., 973 F.2d 265, 269 (4th Cir.1992), and indeed with respect to the ultimate burden of proof for any claim asserted against the debtor. See 11 U.S.C. § 502; Fed.R.Bankr.P. 3001(f). 2 Although courts generally assume that this ultimate burden rests with the claimant, see In re Hemingway Transp., Inc., 993 F.2d 915, 925 (1st Cir.), cert. denied, 510 U.S. 914, 114 S.Ct. 303, 126 L.Ed.2d 251 (1993), the federal circuit courts of appeals are split on the question of whether the ultimate burden falls upon the taxpayer or the taxing authority when a tax claim is asserted in the context of a bankruptcy proceeding. The Third, Fourth, and Seventh Circuits have held that the taxpayer-debtor bears the ultimate burden of proof, just as it does outside of bankruptcy, see Resyn Corp. v. United States, 851 F.2d 660, 663 (3d Cir.1988); In re Landbank Equity Corp., 973 F.2d 265, 268-271 (4th Cir.1992); United States Internal Revenue Serv. v. Charlton, 2 F.3d 237, 239-40 (7th Cir.1993), while the Fifth, Eighth, Ninth, and Tenth Circuits place the ultimate burden on the taxing authority, see In re Placid Oil Co., 988 F.2d 554, 557 (5th Cir.1993) (citing In re Fidelity Holding Co., 837 F.2d 696, 698 [5th Cir.1988]); In re Brown, 82 F.3d 801, 805 (8th Cir.1996); In re MacFarlane, 83 F.3d 1041, 1045 (9th Cir.1996), cert. denied, — U.S. -, 117 S.Ct. 1243, 137 L.Ed.2d 326 (1997); In re Fullmer, 962 F.2d 1463, 1466 (10th Cir.1992). 3 The First Circuit has yet to address this issue. 4 As this case involves the determination of a state tax claim, this Court must look to state law “[ujnless some federal interest requires a different result.” Butner v. United States, 440 U.S. 48, 55, 99 S.Ct. 914, 918, 59 L.Ed.2d 136 (1979); see 28 U.S.C. § 1652; Vanston Bondholders Protective Comm. v. Green, 329 U.S. 156, 161, 67 S.Ct. 237, 239, 91 L.Ed. 162 (1946) (“What claims of creditors are valid and subsisting obli *430 gations against the [debtor] at the time a petition in bankruptcy is filed, is a question which, in the absence of overruling federal law, is to be determined by reference to state law.”) Under New Mexico law, a tax assessment made by the Department is presumptively correct, see N.M.StatAnn. § 7-1-17(C); Stohr v. New Mexico Bureau of Revenue, 90 N.M. 43, 559 P.2d 420, 423 (App.1976), ce rt. denied, 90 N.M. 254, 561 P.2d 1347 (1977) (presumption that an assessment of gross receipts taxes is correct), and the ultimate burden of proof that the assessment is erroneous rests with the taxpayer. See Torridge Corp. v. Commissioner of Revenue, 84 N.M. 610, 506 P.2d 354, 356 (App.1972), cert. denied, 84 N.M. 592, 506 P.2d 336 (1973). The United States Supreme Court has long recognized that burdens of proof constitute part of the substantive law of a state. See Dick v. New York Life Ins. Co., 359 U.S. 437, 446, 79 S.Ct. 921, 927, 3 L.Ed.2d 935 (1959). Courts of appeals on both sides of this issue have acknowledged that the Bankruptcy Code provides no express rules of decision for the determination of tax claims. See MacFarlane, 83 F.3d at 1045; Landbank, 973 F.2d at 270-71. 5 In MacFarlane, however, the Tenth Circuit held that implicit in the Bankruptcy Code is the principle that “taxing authorities are to be treated the same as other claimants,” and that, accordingly, bankruptcy courts must disregard the substantive tax law rule that the burden of proof rests with the taxpayer. MacFarlane, 83 F.3d at 1045; see also Fidelity, 837 F.2d at 698 (“The Bankruptcy Code ... does not differentiate between government and private claimants when proofs of claim are filed”). The MacFarlane court further explained that “the policy goals of the bankruptcy system are put at risk when one class of creditors is given the benefit of a favorable presumption which has its origins outside of bankruptcy law.” MacFarlane, 83 F.3d at 1045 (quoting In re Wilhelm, 173 B.R. 398, 402 [Bankr.E.D.Wis.1994]). This Court finds MacFarlane and Fidelity unpersuasive. By its express terms, the Bankruptcy Code does not treat taxing authorities the same as other claimants. See 11 U.S.C. § 507(a)(8) (affording priority to state and federal tax claims); 11 U.S.C. § 523(a)(1) (tax-related debts may not be discharged in bankruptcy); see also 15 Collier on Bankruptcy HTX5.03[5] (15th ed. 1996) (“Despite the egalitarian appeal of [the] position [expressed in MacFarlane], it overlooks the frequent disparate treatment of the government as tax-creditor found in the Bankruptcy Code regarding such matters as the priority and dischargeability of claims.”); Frances R. Hill, Toward a Theory of Bankruptcy Tax: A Statutory Coordination Approach, 50 Tax Law. 103, 162 (1996) (“In the absence of a provision in the Bankruptcy Code, it is far from clear that formal equality is a principle of fairness in disputes among creditors.”). Although certain provisions in the Bankruptcy Code do indeed modify the rights of taxing authorities in order to foster the objective of equality of distribution among creditors, “in those instances in which Congress has decided that the administration of the Bankruptcy Code requires alteration in the substantive rights of the parties, it has not hesitated to make such intentions clear.” Landbank, 973 F.2d at 270 (citing 11 U.S.C. §§ 544-551, 553 [authorizing the trustee to avoid otherwise legal transfers of property for the benefit of the bankruptcy estate]; 11 U.S.C. § 507 [establishing priorities for satisfaction of claims *431 against the bankrupt estate]). 6 This Court agrees with the Landbank court that, except for those provisions containing a clear expression of Congressional intent to the contrary, the Bankruptcy Code, “while providing a forum and procedures for an expedient dispute resolution, does not endeavor to supplant the substantive law under which the claim against the estate ... arose.” Id. (citing Report of the Commission on the Bankruptcy Laws of the United States, H.R. Doc. No. 93-137, pt. I, at 68-71 [1973] ). 7 This Court’s interpretation of the Bankruptcy Code is further reinforced by the “background presumption that federal law generally will not interfere with [the] administration of state taxes.” National Private Truck Council, Inc. v. Oklahoma Tax Comm’n, 515 U.S. 582, 588, 115 S.Ct. 2351, 2355, 132 L.Ed.2d 509 (1995) (citing principles of comity and federalism). Outside of the bankruptcy forum, the law places the burden of proof upon the taxpayer because 1) in all likelihood, the taxpayer is the party with access to the relevant evidence, Rexach, 482 F.2d at 16, and 2) if the taxpayer knows that the burden of proof falls upon the taxing authority, it would have little incentive to maintain the necessary records, “thereby making the government’s ease more difficult if not impossible to establish.” Id. at 17. As these reasons are just as applicable in the bankruptcy context, see In re Abel, 200 B.R. 816, 820 (E.D.Pa.1996), 8 in the view of this Court, a rule shifting the burden of proof to the taxing authority in bankruptcy proceedings would render the bankruptcy forum more favorable to taxpayers than the non-bankruptcy forum, funnel people with tax problems into the bankruptcy courts, and thus undermine the enforcement of state tax laws. 9 As the Supreme Court noted in Butner, “[u]niform treatment of property interests by both state and federal courts within a State serves to reduce uncertainty, to discourage forum shopping, and to prevent a party from receiving a ‘windfall merely by *432 reason of the happenstance of bankruptcy.’ ” Butner, 440 U.S. at 55, 99 S.Ct. at 918 (quoting Lewis v. Manufacturers Nat’l Bank, 364 U.S. 603, 609, 81 S.Ct. 347, 350, 5 L.Ed.2d 323 [1961]); see also In re Cobb, 135 B.R. 640, 641 (Bankr.D.Neb.1992) (“To rule otherwise would permit a tax litigant to shift the burden of proof to the [taxing authority] by filing bankruptcy.”). For the foregoing reasons, this Court affirms the Bankruptcy Court’s ruling that the burden of proof rests with the taxpayer, Thinking Machines. B. The Application of the “Predominant Ingredient” Standard New Mexico uses the “predominant ingredient” test to determine whether an activity constitutes the provision of a service or a sale of tangible property for purposes of the gross receipts tax. See EG & G, 607 P.2d at 1163-64. In applying this standard, courts look to the seller’s costs and compare “the relative inputs of services and tangible property.” Id. at 1164 (finding a sale of services because the seller’s costs predominantly related to the provision of services); see Phillips Mercantile Co. v. New Mexico Taxation and Revenue Dep’t, 109 N.M. 487, 786 P.2d 1221, 1224 (App.1990) (finding a sale of tangible property because the seller’s costs predominantly related to the production and sale of tangible property). Thinking Machines contends that, even if the burden of proof falls upon the taxpayer, it was clearly erroneous for the Bankruptcy Court to conclude that Thinking Machines failed to prove by a preponderance of the evidence that sales of tangible property predominated the disputed Maintenance Agreements. Thinking Machines first asserts that the Bankruptcy Court committed clear error by disregarding the “uncontroverted” testimony of its vice president of finance, Robert LaBossiere (“LaBossiere”). On the stand, LaBossiere stated that he believed the provision of replacement hardware constituted eighty to ninety percent of Thinking Machines’ total costs under the Agreements. See Joint Appendix (“App.”) at 125. As the Bankruptcy Court noted in its ruling, however, Thinking Machines “never offered any documentation that would have substantiated such a conclusion.” Thinking Machines, 203 B.R. at 7. Under New Mexico law, a taxpayer must have specific records if it wants to claim a deduction or exemption. See N.M.Stat.Ann. § 7-l-10(A); N.M. Taxation and Revenue Department Rule TA 90-10:1; Wing Pawn Shop v. Taxation and Revenue Dept., Ill N.M. 735, 809 P.2d 649, 655-56 (App.1991); see also N.M.Ann. 7-9-5 (default presumption that all receipts of a person engaging in business are subject to the gross receipts tax). Furthermore, LaBossiere did not testify that he had ever actually calculated the costs of these hardware items, and acknowledged that he did not know the average materials costs or the rates at which replacement parts were needed. See Thinking Machines, 203 B.R. at 7; App. at 111-112. 10 Thinking Machines did attempt to substantiate its cost figures indirectly, but to no avail. First, LaBossiere stated that such a cost breakdown is the “standard” in the industry, but the Bankruptcy Court ruled that he was not qualified to testify as an expert. Second, Thinking Machines submitted documentation showing how it priced items provided under the Maintenance Agreements, but the Bankruptcy Court was unpersuaded that the pricing information accurately reflected actual costs. Third, Thinking Machines produced its replacement parts log for two of its customers to establish the sheer quantity of parts it supplied under the Maintenance Agreements. The Bankruptcy Court concluded, however, that even if the logs were accurate, they failed to indicate the costs of these parts to the debtor. See Thinking Machines, 203 B.R. at 6-7. None *433 of these rulings was in error. The “predominant ingredient” test looks to the individual seller’s actual costs, rather than quantity, price, or the industry standard. See EG & G, 607 P.2d at 1163-64. Thinking Machines complains that the Bankruptcy Court applied the “predominant ingredient” test too constrictively and expected it to produce minutia (i.e., the precise material costs for the hardware). As the Department points out, however, the Bankruptcy Court did not expect minutia, but rather something other than the sweeping generalizations about costs expressed in terms of ranges in the industry. In arguing that the Bankruptcy Court should have looked no further than the relative inputs of software, hardware, and applications consulting services that comprise the Maintenance Agreements, Thinking Machines misses the point entirely. Under the Agreements, Thinking Machines did not simply furnish replacement hardware and software; it also 1) conducted repairs on existing parts; 2) installed the replacement hardware and software; and 3) helped customers determine when replacements or upgrades were necessary. The precise breakdown of Thinking-Machines’ labor and material costs is therefore very much relevant to the question of whether the Maintenance Agreements are predominantly sales of tangible property or services. 11 Accordingly, this Court affirms the Bankruptcy Court’s rulings that 1) Thinking Machines failed to prove by a preponderance of the evidence that the Maintenance Agreements predominantly involved sales of tangible property, and 2) that the Maintenance Agreements are therefore taxable as sales of services. C. Allocation of Out-of-State Services Given the Bankruptcy Court’s determination that the Maintenance Agreements predominantly involved sales of services, all of the proceeds from these agreements— including those attributable to the “less than predominant” sales of tangible property — are subject to New Mexico’s gross receipts tax, see EG & G, 607 P.2d at 1164, with one important exception. New Mexico law provides that “receipts from selling services performed outside New Mexico the product of which is initially used in New Mexico” are exempt from the gross receipts tax. N.M.Stat.Ann. § 7-9-13.1(A); see also New Mexico Taxation and Revenue Department, Publication GR-90 Gross Receipts and Compensating Tax Act Regulations at 3(F):33 (Jan.1992) (hereinafter “GR-90”) (“[rjeeeipts from performing services, except research and development services, outside New Mexico are not subject to the gross receipts tax”). Determining which sales of services qualify for this out-of-state exemption is by no means a simple proposition. GR-90 at 3(K):3 states that: When a prime contractor performs services both within and without New Mexico, cost accounting records which reasonably allocate all costs to the location of the performance of the service shall be used to determine the amount of services performed in New Mexico. If adequate cost accounting records are not kept for the allocation of costs to specific locations, the receipts from performing such services *434 shall be prorated based on the percentage of service actually performed within New Mexico. The percentage shall be calculated by dividing the time spent ... performing such services in New Mexico by the total contract time spent performing services everywhere. 12 Here, the Bankruptcy Court relied upon one of the models submitted by the Department in its post-trial reply brief (“Exhibit D”) to calculate the percentage of services provided under the Maintenance Agreements that were performed outside of New Mexico. This Court holds that the Bankruptcy Court’s adoption of this model was clearly erroneous. As set forth below, Exhibit D is premised upon assumptions which are incompatible with the Bankruptcy Court’s prior conclusion that costs attributable to sales of services predominated the Maintenance Agreements. Exhibits C and D of the Department’s post-trial reply brief and page two of Trial Exhibit 7 (“Exhibit 7”) are all models — albeit ones with different assumptions — submitted to the Bankruptcy Court by the Department’s auditor, Michael Giles, on how to allocate out-of-state service costs. Exhibit 7 was the model that Giles explained to the Bankruptcy Court while testifying, and Exhibit D is the model actually adopted by the Bankruptcy Court. All three of these models are attached hereto as appendices, as is Trial Exhibit 10 mentioned herein. Column A in these models represents the total revenues earned by Thinking Machines under the Maintenance Agreements, as derived from the yearly audits of the company. Column B reflects the “total costs” under the Maintenance Agreements. 13 In Exhibit D, the total costs are derived by taking the total revenues from the yearly audits, and dividing them by the yearly “mark-up” figures submitted by Thinking Machines in Exhibit 10. 14 Column C constitutes the payroll costs for the applications engineers that Thinking Machines furnished to its customers under the Maintenance Agreements. It is undisputed that the services furnished by these applications engineers were performed in New Mexico, and thus qualify as “in-state service costs.” Column D “grosses” up these New Mexico payroll costs by 30% to reflect additional costs such as employment taxes, FICA, worker’s compensation, health insurance, travel expenses, and other employee benefits. Column E, entitled “Other Costs,” represents all of the costs incurred by Thinking Machines under the Agreement except for *435 the payroll costs for the New Mexico engineers. (Column B minus Column D). It is undisputed that all of these “other costs” were actually incurred in Massachusetts, where Thinking Machines’ headquarters are located. At the Bankruptcy Court hearing, auditor Giles testified that due to the failure of Thinking Machines to produce any records breaking down its labor and materials costs, he had no reliable way to calculate what percentage of these “other costs” related to sales of tangible goods (e.g., the provision of new hardware) and what percentage constituted out-of-state service costs (e.g., hardware repair services), so he arbitrarily picked a 50-50 split between the two. App. at 42. In Exhibit 7, Column F reflects this 50-50 breakdown of “other costs” (Column E times 50%). In its ruling, the Bankruptcy Court “conclude[d] that the 50-50 split [of ‘other costs’] is appropriate.” Thinking Machines, 203 B.R. at 10. Nevertheless, the Bankruptcy Court then proceeded to adopt Exhibit D, which instead allocates only 15% of “other costs” to out-of-state service costs (Column F), 15 and even more importantly, only 35% of the total costs under the Maintenance Agreements to service costs (Column G divided by Column B). 16 This 35% ratio of service costs to total costs directly contradicts the Bankruptcy Court’s prior (and proper) ruling that costs relating to sales of services predominate the Maintenance Agreements. Due to the failure of Thinking Machines to submit adequate cost accounting records, it was necessary, for the Bankruptcy Court to approximate when determining what percentage of “other costs” in Column E constitute service costs. But it was clearly erroneous for the Bankruptcy Court to select a percentage which belies the proposition that the Maintenance Agreements involved predominantly sales of services. In its post-trial reply brief containing Exhibit D, the Department explains that it selected the 15% allocation figure for out-of-state service costs because it falls between 10 and 20%, which was the range given by LaBossiere as standard for the industry. The Department goes on to note, however, that Exhibit D “do[es] not purport to be helpful in determining whether the maintenance agreements constitute predominately the sale of tangibles or services.” App. at 392. In fact, this argument only reinforces Exhibit D’s central flaw: it effectively permits the Department to have its cake and eat it too. Recognizing that the tax burden will be highest when 1) costs relating to sales of services predominate the Agreements, but 2) of the costs incurred out of state, as few as possible are deemed costs relating to sales of services, see supra note 12, the Department rejects LaBossiere’s testimony regarding the breakdown between services and tangible goods when determining which predominates the Agreements, but then, in Exhibit D, wholeheartedly embraces his “unsubstantiated” statements for the purpose of calculating what percentage of the total costs incurred . out of state are fairly attributable to sales of services. The Department neglects to mention that if LaBossiere’s assertion — that 85% of the costs incurred in Massachusetts (i.e., “other costs”) relate to sales of tangible goods — is taken as true, it necessarily follows that 65% of total costs (in New Mexico and Massachusetts) relate to sales of tangible goods, thereby rendering the Maintenance Agreements exempt from the gross receipts tax. 17 *436 In contrast, the 50% allocation figure for out-of-state service costs used in Exhibit 7 yields a ratio of total service costs to total costs of 72.9%, 18 a result consistent with the Bankruptcy Court’s conclusion that costs relating to sales of services predominate the Maintenance Agreements. Nevertheless, on remand, the Bankruptcy Court is under no obligation to adopt Exhibit 7; this Court merely notes that it is but one example of an acceptable allocation model. 19 In light of the shortage of objective evidence in the record with respect to Thinking Machines’ costs, the Bankruptcy Court has a broad range of discretion in selecting the allocation and markup percentages to be used, provided that the model selected is compatible with the Bankruptcy Court’s conclusion that sales of services predominate the Maintenance Agreements. III. CONCLUSION For the foregoing reasons, the ruling of the Bankruptcy Court is AFFIRMED in part and REVERSED in part. The Bankruptcy Court properly ruled that the burden of proof rests with the taxpayer to disprove the validity of the Department’s claim. The Bankruptcy Court’s conclusion that Thinking Machines failed to satisfy this burden was not clearly erroneous. The allocation model used by the Bankruptcy Court to approximate what percentage of services were performed out-of-state, however, was incompatible with the Bankruptcy Court’s holding that the Maintenance Agreements predominantly involved sales of services, and is therefore clearly erroneous. Accordingly, this ease is hereby REMANDED to the Bankruptcy Court for a recalculation of the appropriate tax base in a manner consistent with this opinion. SO ORDERED. APPENDIX (adopted by the Bankruptcy Court) Submitted with the Department’s Post-Trial Reply Brief COMPUTATION OF RECEIPTS FROM THE PERFORMANCE OF SERVICES, INCLUDING SOFTWARE SERVICES, IN NEW MEXICO BY CALCULATING ESTIMATED PERCENTAGES OF THE COST OF SERVICES PERFORMED IN NEW MEXICO AND OUTSIDE OF NEW MEXICO. THE COST OF IN-STATE SERVICES PERFORMED BASED ONLY UPON INSTATE PAYROLL OF ENGINEERS GROSSED UP 30%(SEE NOTE # 1) FOR PAYROLL BURDEN THAT WOULD BE EXPENSED IN THE COST OF REVENUE. SOURCES: EXHIBIT NUMBER 10, AUDIT WORKPAPERS, DATABASE RUN OF NM PAYROLL AND TRIAL TRANSCRIPT. AUTHORITY:REGULATION GR 3(K):3, SECTION 7-1-10 NMSA 1978, REGULATION TA 10:1. *437 MARKUP PERCENTAGES ARE FROM! EXHIBIT NUMBER 10: 1988 IS 100% 1989 IS 22% 1990 IS 12% 1991 IS -6% 1992 IS -10% [[Image here]] NOTE # 1: THE 20% PROPOSED BY THINKING MACHINES IS REJECTED, SINCE THINKING MACHINES DOES NOT ACCOUNT FOR NON-NEW MEXICO EMPLOYEES WHO OCCASIONALLY WORKED IN NEW MEXICO. NOTE # 2: ASSUMPTION THAT 85% OF OTHER COSTS(COLUMN E) IS OVERHEAD EXPENSES RELATING TO HARDWARE. SEE TRANSCRIPT PAGE 80. (described during the testimony of the Department’s auditor, Michael Giles) COMPUTATION OF RECEIPTS FROM THE PERFORMANCE OF SERVICES, INCLUDING SOFTWARE SERVICES, IN NEW MEXICO BY CALCULATING ESTIMATED PERCENTAGES OF THE COST OF SERVICES PERFORMED IN NEW MEXICO AND OUTSIDE OF NEW MEXICO. THE COST OF IN-STATE SERVICES PERFORMED BASED ONLY UPON INSTATE PAYROLL OF ENGINEERS GROSSED UP 30% FOR PAYROLL BURDEN THAT WOULD BE EXPENSED IN THE COST OF REVENUE SOURCES: DISCLOSURE STATEMENT, AUDIT WORKPAPERS, DATABASE RUN OF NM PAYROLL AUTHORITY:REGULATION GR 3(K):3, SECTION 7-1-10 NMSA 1978, REGULATION TA 10:1 COMPUTATION OF MARKUP FACTOR BASED UPON THE 1995 FORCASTED INCOME STATEMENT IN THE DISCLOSURE STATEMENT FOR THE FIRST AMENDED JOINT PLAN OF REORGANIZATION: SOFTWARE & SERVICE REVENUE $17,058.000 SOFTWARE & SERVICE COST $8.466,000 REVENUE DIVIDED BY COST = 2.014883= MARKUP FACTOR *438 YEAR 1988 1989 1990 1991 1992 TOTAL SOFTWARE & MAINT. REV. PER AUDIT $69,230.00 $521,662.00 $1,068,713.00 $1,309,925.00 $289,761.00 $3,259,291.00 B TOTAL SERVICE COST (A/MARKUP) $34,359.32 $258,904.36 $530,409.46 $650,124.60 $143,810.34 $1,617,608.08 NEW MEXICO PAYROLL FOR ENGINEERS $0.00 $79,084.68 $195,220.90 $224,798.57 $71,360.70 $570,464.85 NEW MEXICO SERVICE COST (C*130%) $0.00 $102,810.08 $253,787.17 $292,238.14 $92,768.91 $741,604.30 OUT OF STATE SERVICE & MATERIALS (B-D) $34,359.32 $156,094.28 $276,622.29 $357,886.46 $51,041.43 $876,003.78 YEAR 1988 1989 1990 1991 1992 TOTAL OUT OF STATE SERVICE COSTS (E*50%) $17,179.66 $78,047.14 $138,311.15 $178,943.23 $25,520.71 $438,001.89 G TOTAL SERV. COSTS (D + F) $17,179.66 $180,857.22 $392,098.32 $471,181.37 $118,289.62 $1,179,606.19 H INSTATE SERVICE PERCT. (D/G) 0.0000% 56.8460% 64.7254% 62.0224% 78.4252% OUT OF STATE SERVICE PERCT. (F/G) 100.0000% 43.1540% 35.2746% 37.9776% 21.5748% J INSTATE REVENUE (H*A) $0.00 $296,543.98 $691,728.76 $812,446.92 $227,245.64 $2,027,965.30 NOTE: ATTRIBUTED 50% OF COSTS IN COLUMN F TO MATERIALS AND 50% TO SERVICES BECAUSE OF INABILITY TO QUANTIFY AN AMOUNT FOR EITHER COMPONENT NOR A RELATIONSHIP BETWEEN THEM. Exhibit C Submitted with the Department’s Post-Trial Reply Brief COMPUTATION OF RECEIPTS FROM THE PERFORMANCE OF SERVICES, INCLUDING SOFTWARE SERVICES, IN NEW MEXICO BY CALCULATING ESTIMATED PERCENTAGES OF THE COST OF SERVICES PERFORMED IN NEW MEXICO AND OUTSIDE OF NEW MEXICO. THE COST OF IN-STATE SERVICES PERFORMED BASED ONLY UPON INSTATE PAYROLL OF ENGINEERS GROSSED UP 30%(SEE NOTE # 1) FOR PAYROLL BURDEN THAT WOULD BE EXPENSED IN THE COST OF REVENUE. SOURCES: DISCLOSURE STATEMENT, AUDIT WORKPAPERS, DATABASE RUN OF NM PAYROLL AND TRIAL TRANSCRIPT. AUTHORITY:REGULATION GR 3(K):3, SECTION 7-1-10 NMSA 1978, REGULATION TA 10:1. COMPUTATION OF MARKUP PERCENTAGE BASED UPON THE 1995 FORCASTED INCOME STATEMENT IN THE DISCLOSURE STATEMENT FOR THE FIRST AMENDED JOINT PLAN OF REORGANIZATION: SOFTWARE & SERVICE REVENUE $17,058,000 SOFTWARE & SERVICE COST $8,466,000 COST DIVIDED BY REVENUE = 0.496307= MARKUP PERCENTAGE *439 [[Image here]] NOTE # 1: THE 20% PROPOSED BY THINKING MACHINES IS REJECTED, SINCE THINKING MACHINES DOES NOT ACCOUNT FOR NON-NEW MEXICO EMPLOYEES WHO OCCASIONALLY WORKED IN NEW MEXICO. NOTE # 2: ASSUMPTION THAT 85% OF OTHER COSTS(COLUMN E) IS OVERHEAD EXPENSES RELATING TO HARDWARE. SEE TRANSCRIPT PAGE 80. Trial Exhibit 10 1988-92 Service Margins 1988 1989 1990 1991 1992 Service Revenue 1,836 4,886 7,727 9,209 11,103 Service Costs 0 3,813 6,820 9,797 12,194 Gross Margin 1,836 1,073 907 (588) (1,091) Margin % 100% 22% 12% -10% Footnotes [1] It is undisputed that this Court has jurisdiction to hear this appeal pursuant to 28 U.S.C. §§ 158(a), 158(c)(1). [2] It is necessary to briefly clarify what is meant by the term "ultimate” burden of proof in the bankruptcy context. Prepetition claims are asserted by means of a proof of claim. See 11 U.S.C. § 501. A properly filed proof of claim is prima facie evidence of the validity and amount of the claim. See 11 U.S.C. § 502(a); Fed. R.Bankr.P. 3001(f). "The interposition of an objection [by the debtor] does not deprive the proof of claim of presumptive validity unless the objection is supported by substantial evidence.” In Re Hemingway Transp., Inc., 993 F.2d 915, 925 (1st Cir.), cert. denied, 510 U.S. 914, 114 S.Ct. 303, 126 L.Ed.2d 251 (1993) (citations omitted). Only once the debtor meets this initial burden of producing substantial evidence does the question of which party bears the ultimate burden of proof come into play. See In Re Fullmer, 962 F.2d 1463, 1466 (10th Cir.1992). Other than stating that the bankruptcy court should determine such a claim after notice and a hearing, 11 U.S.C. § 502(b), neither the Bankruptcy Code nor Rule 3001(f) provide any further guidance with respect to the ultimate burden of proof. In this case, Judge Hillman simply mentioned these prima facie standards, and then moved right into a discussion of which party bears the ultimate burden of proof when a tax claim is asserted in bankruptcy. As a result, the Department argues that 1) the Bankruptcy Court properly decided that the ultimate burden of proof rests with the taxpayer, or 2) in the alternative, this Court should affirm because Thinking Machines never even rebutted the prima facie validity of the Department's claim by producing substantial evidence. As this Court agrees with Judge Hillman’s analysis that the ultimate burden lies with the taxpayer, it is not necessary to address the Department’s second argument. [3] The Supreme Court recently declined an opportunity to resolve this conflict by denying certiorari in MacFarlane, - U.S. -, 117 S.Ct. 1243, 137 L.Ed.2d 326 (1997), but only after the United States Solicitor General pointed out that the Ninth Circuit lacked jurisdiction over the appeal in that case. See Brief for the United States as Amicus Curiae at 12-14, California Franchise Tax Bd. v. MacFarlane, - U.S. -, 117 S.Ct. 1243, 137 L.Ed.2d 326 (1997). [4] In Hemingway, the First Circuit stated that once the debtor meets its initial burden under Fed.R.Bankr.P. 3001(f) of producing substantial evidence, "the ultimate risk of nonpersuasion as to the allowability of the claim resides with the party asserting the claim." Hemingway, 993 F.2d at 925. That case, however, involved a claim by a private creditor who would have borne the burden of proof outside of bankruptcy, see id. at 932, and the court dealt with the burden of proof issue in the context of interpreting 11 U.S.C. § 502(e)(1)(B), which disallows certain contribution and indemnification claims. See id. at 925, 933. There is no indication that the First Circuit contemplated the application of its statement regarding the ultimate burden of proof to a claim asserted by a taxing authority. [5] Some of the courts placing the burden of proof upon the taxing authority have simply stated, without any further analysis, that under Fed. R.Bank.P. 3001(f), the ultimate burden of proof always rests with the claimant. See Fullmer, 962 F.2d at 1466; Brown, 82 F.3d at 804-05 (citing Hemingway, 993 F.2d at 925). As discussed above, however, the plain language of Rule 3001(f) does not speak to the ultimate burden of proof for any claim asserted against the debtor. Although courts typically assume that the ultimate burden rests with the claimant, this Court agrees with the Bankruptcy Court that the rationale for this general principle is to place the claimant in the same position it would have been in as a civil plaintiff outside of the bankruptcy context. Thinking Machines, 203 B.R. at 2 (citing In Re Premo, 116 B.R. 515, 518 [Bank.E.D.Mich.1990] ). The blind application of this general principle is therefore problematic when the claimant is a taxing authority that does not bear the burden of proof as a civil plaintiff. [6] Other sections of the Bankruptcy Code and Federal Bankruptcy Rules of Procedure indicate that when Congress wishes to assign the burden of proof to a particular party, it does so expressly. See 11 U.S.C. §§ 362(g), 364(d)(2), 547(g), and 1129(d); Fed.R.Bankr.P. 4003(c), 4005. [7] The MacFarlane decision is, moreover, internally inconsistent. The court first restates the holding of Fidelity that the Bankruptcy Code requires equal treatment of claims filed by government and private claimants, but then affirms the district court's conclusion that “since tax claims already receive a statutory priority over other creditor’s claims,” placing the burden of proof on the taxpayer would grant the taxing-authority "a double benefit not authorized by statute [or the case law].” MacFarlane, 83 F.3d at 1045. As recognized by the Bankruptcy Court below, to the extent that the Bankruptcy Code’s priority scheme is at all relevant to the question of which party bears the burden of proof, the priority status afforded to the taxing authority only reinforces the rule that the burden of proof rests with the taxpayer. Thinking Machines, 203 B.R. at 4. In fact, the stated desire of the MacFarlane court to prevent taxing authorities from receiving a “double benefit” is more accurately characterized as an effort to "create equitable impediments for a certain class of creditors based on the notion that Congress has given them too much.” Brief of Appellee New Mexico Taxation and Revenue Department at 10. Such use of purported equitable principles to override legislatively created rights is an unwarranted invasion by the courts upon the legislative function. See United States v. Noland, - U.S. -, -- -, 116 S.Ct. 1524, 1527-28, 134 L.Ed.2d 748 (1996). [8] In Premo, the Bankruptcy Court noted that although it is true that the taxpayer is generally more likely to have access to the relevant information, "[tjhere is usually no reason to presume that another creditor, or even a bankruptcy trustee, has greater access to the debtor's documents than does [the taxing authority]." Premo, 116 B.R. at 523. A taxing authority differs from most private creditors, however, in that 1) it is a nonconsensual creditor, and 2) any losses borne by it are passed on the general public in the form of higher taxes. Hill, Toward a Theory of Bankruptcy Tax, 50 Tax Law. at 112. [9] See Brief for the States of Arizona, Connecticut, Hawaii, Idaho, Iowa, Massachusetts, Michigan, Minnesota, Montana, Nevada, New Hampshire, New York, North Carolina, North Dakota, Ohio, Oklahoma, Oregon, South Dakota, Texas, Utah, Vermont, Virginia, Wisconsin, and New Mexico as Amici-Curiae in Support of Petitioner at 13-14, California Franchise Tax Bd. v. MacFarlane, - U.S. -, 117 S.Ct. 1243, 137 L.Ed.2d 326 (1997). The bankruptcy court in Premo also considered this possibility, but concluded that there was little risk that taxpayers would allow themselves to fall into bankruptcy to obtain a more favorable burden of proof. Premo, 116 B.R. at 523. This Court respectfully but adamantly disagrees. [10] LaBossicrc’s testimony was also internally inconsistent. He first testified that costs relating to replacement hardware constituted 80 to 90% of the total out-of-state costs incurred under the Maintenance Agreements, App. at 86, and only later stated that this 80-90% figure also applied to Thinking Machines' overall costs, App. at 125. As it is undisputed that the only costs actually incurred by Thinking Machines in New Mexico were costs relating to the provision of services by the on-site applications engineers, it is mathematically impossible for both of these statements to be accurate. [11] This is not to suggest that all labor costs relate to the provision of services. The labor involved in building new hardware, for example, would constitute an input to the cost of producing tangible property. In contrast, the labor involved in repairing an existing part, or installing a new part, would constitute an input to the provision of services. Nevertheless, without breaking down the inputs in this manner, it is impossible to discern with any degree of accuracy what percentage of costs incurred under the Maintenance Agreements related to sales of tangible property. In further support of its argument that the Bankruptcy Court applied the predominant ingredient test too restrictively, Thinking Machines refers to a sentence in Judge Hillman’s ruling noting that Thinking Machines failed to prove by a preponderance of the evidence that the replacement hardware parts constituted 80-90% of the costs under the Maintenance Agreements. See Thinking Machines, 203 B.R. at 7. This statement is taken out of context; the Bankruptcy Court was not suggesting 80-90% as a measure of predominance, but rather noting that Thinking Machines had failed to provide any evidence to corroborate LaBossiere's testimony. The Bankruptcy Court repeatedly stated that Thinking Machines need prove only that sales of tangible property constituted the predominant ingredient of the Maintenance Agreements. See id. at 7, 9. [12] The following hypothetical illustrates the ramifications of these provisions. Assume that a Court concludes that of the $1,000 in costs incurred under an Agreement, $800 were attributable to sales of services, while $200 were attributable to sales of tangible property. Under the predominant ingredient test, the entire $1,000 is subject to the gross receipts tax. If the Court were to find, however, that 30% of the service costs incurred under the Agreement were attributable to out-of-state sales of services, then only $700 of the revenues would be taxable in New Mexico. Note that it is only services performed out of state that influence this percentage; the percentage of costs relating to sales of tangible goods that are incurred out of state is irrelevant. At the same time, however, in applying the percentage (i.e. multiplying it by the total contract proceeds), the sales of tangible property are affected. Thus, the tax burden will be highest when 1) costs relating to sales of services predominate the Agreement, but 2) of the costs incurred out of state, as few as possible are deemed to be costs relating to the provision of services. [13] In all three models, Column B and Column G are both labeled "Total Service Costs.” After reviewing the testimony of auditor Giles, it is clear to this Court that Column B actually reflects "Total Costs” under the Agreements, while Column G reflects "Total Service Costs.” [14] The fact that Exhibit D relies upon mark-up figures derived from Trial Exhibit 10, entitled "1988-92 Service Margins,” is further evidence of the internal contradictions contained in the Bankruptcy Court's analysis. In its opinion, the Bankruptcy Court stated: [T]he Debtor entered into evidence a document entitled "1988-92 Service Margins.”... I cannot accord this document great weight. I do not know who prepared the document, I do not know its purpose and I do not know upon what documents it is based. Thinking Machines, 203 B.R. at 10 n. 8. Despite this assessment, the Bankruptcy Court then proceeded to select Exhibit D over the other models submitted by the Department (Exhibit C and Exhibit 7), even though none of the other models relied upon Trial Exhibit 10 to calculate the mark-up percentages. [15] The end of the Bankruptcy Court's opinion does state, without any explanation, that the 15% allocation for out-of-state service costs in Exhibit D "is a fair resolution of the controversy presented in this section.” Thinking Machines, 203 B.R. at 10. Only four sentences earlier, however, the Bankruptcy Court concluded that a 50% allocation is "appropriate.” Id. This suggests that the Bankruptcy Court did not fully comprehend the various assumptions underlying Exhibit D. [16] The ratio of total service costs (Column G) to total costs (Column B) under Exhibit D is $1,094, 957.71/$3,097,293.74 = 35.352%. [17] The Department also attempts to justify the 15% allocation for out-of-state service costs used in Exhibit D by noting that some of these "other costs" incurred in Massachusetts are actually overhead expenses that are fairly attributable to services performed in New Mexico. See GR-90 at 3(F):65; United States v. New Mexico, 581 F.2d 803, 811 (10th Cir.1978) (citing Mountain States Advertising, Inc. v. Bureau of Revenue, 89 N.M. 331, 552 P.2d 233 [App.1976]). Although the Department is likely correct that some percentage of "other costs” may be allocated in this manner (e.g., plant, accounting, management, *436 and insurance costs), this Court notes that many of the costs that Thinking Machines incurred in Massachusetts — such as the costs associated with debugging software and repairing computer parts — do not fall within New Mexico’s definition of overhead or general administrative expenses. See GR-90 at 3(F):65; United States v. New Mexico, 581 F.2d at 805. Moreover, overhead expense is an input — just like direct materials and labor — that must be allocated either to sales of services or sales of tangible property; it does not simply vanish into thin air. As Exhibit D does not increase either total in-state or out-of-state service costs to reflect overhead costs, this Court must assume that any such overhead costs were allocated to the provision of tangible goods. See also Exhibit D, note 2 ("assumption that 85% of other costs (Column E) is overhead relating to hardware”). Thus, the Department’s characterization of part [or all] of this 85% of "other costs” as overhead does not alter the fact that under Exhibit D, only 35% of Thinking Machines’ total costs under the Maintenance Agreements constitute costs relating to sales of services. [18] The ratio of total service costs (Column G) to total costs (Column B) under Exhibit 7 is $1,179, 606.19/$1,617,608.08 = 72.92%. [19] This Court notes that Exhibit 7 uses different mark-up percentages than Exhibit D.