Opinion
No. 2004AP3239.
Oral argument November 17, 2005.
Decided January 25, 2007.
APPEAL from an order of the circuit court for Dane County: STEVEN D. EBERT, Judge. Reversed.
On behalf of the respondent-appellant, the cause was submitted on the briefs of Leonard S. Sosnowski and Andrew L. Nelson of Foley Lardner LLP, Madison, and Maureen A. McGinnity of Foley Lardner LLP, Milwaukee. There was oral argument by Maureen A. McGinnity.
On behalf of the petitioner-respondent, the cause was submitted on the briefs of and oral argument by F. Thomas Creeron III, assistant attorney general, and Peggy A. Lautenschlager, attorney general.
A nonparty brief was filed by Patricia J. Kaeding and Brady Williamson of La Follette Godfrey Kahn, Madison, for Wisconsin Manufacturers and Commerce.
Before Vergeront, Deininger and Higginbotham, JJ.
Menasha Corporation appeals a circuit court order reversing a decision of the Wisconsin Tax Appeals Commission. The commission ruled that a computer system purchased by Menasha was a customized computer program within the meaning of WIS. ADMIN. CODE § Tax 11.71(1)(e), and therefore exempt from sales and use tax under WIS. STAT. § 77.51(20) (2003-04). The circuit court reversed the commission's decision and reinstated the Wisconsin Department of Revenue's (DOR) determination that the program was non-custom software and therefore taxable as tangible property. We conclude that the commission's decision that the computer software Menasha purchased was customized software and not prewritten, and therefore not taxable, rests on a reasonable interpretation and application of § Tax 11.71. We also conclude that the DOR's construction of the rule is not more reasonable than the commission's. We therefore reverse the circuit court's order reversing the commission's decision, and affirm the commission's decision in favor of Menasha.
All references to the Wisconsin Statutes are to the 2003-04 version unless otherwise noted.
Menasha also argues that the circuit court's ruling violated equal protection guarantees. Because we decide this case in Menasha's favor, we need not address this issue.
BACKGROUND
¶ 2. The following undisputed material facts are taken from the commission's decision. Menasha is a corporation, based in Wisconsin, with sixty-three business locations in twenty states and eight countries. Through its subsidiaries, Menasha provides products to a variety of industries, including paperboard, packaging, plastics, material handling, promotions, and printing. SAP, a German company, is one of the world's largest designers of integrated business application computer software, including the software at issue, the R/3 System.
The commission required the parties to submit proposed stipulated facts. The commission considered submissions from both parties and issued what was labeled as "Undisputed Material Facts." However, the DOR disputed the characterization or accuracy of some of the facts ultimately adopted by the commission. On appeal, the DOR disputes two of the commission's Undisputed Material Facts: #29 and #47. Because, as we explain in Section II of this opinion, we conclude that the commission properly concluded that any factual disputes in this case were not material or did not raise conflicting inferences, we treat the facts as undisputed in the background section of this opinion.
¶ 3. The R/3 System consists of more than seventy software modules. Each module can provide a rudimentary business and accounting software system for a different segment of a client's business. The system is not usable to a client as sold; it must be modified to fit a client's business operations. It becomes usable for serving a client's business and accounting needs only after the modifications are completed. ABAP (an acronym for "Advanced Business Application Programming") is the programming language SAP developers use to customize the R/3 System to fit a particular customer's business. SAP customers who license the R/3 System "almost always retain either SAP or SAP's designated consultants" to help customize the R/3 System modules to their businesses.
¶ 4. In 1993, Menasha hired a consulting company to conduct a feasibility study evaluating available computer-based business systems that could accommodate its special processing needs, e.g., integrating its subsidiaries' systems and accommodating their unique and diverse requirements, while maximizing uniformity through a common database. The consulting company concluded that it would be feasible for Menasha to transition to a global application software system, provided that the new system allowed custom modification to meet Menasha's unique business needs. Menasha consequently selected SAP as the vendor for its new system, due to the flexibility and customization which Menasha required and which SAP could provide.
Licensing, Installing and Customizing the R/3 System
¶ 5. During initial discussions with SAP, Menasha emphasized that the software system it selected must be capable of being customized to fit its business needs. SAP conducted several demonstrations of the R/3 System at Menasha's office to determine the extent of the modifications needed to satisfy Menasha's business needs. Based on the data collected from these demonstrations, SAP affirmed the need to "Customize with Minimal Retrofitting." SAP continued to be "involved in the overall implementation strategy, including training and project planning of the system." At Menasha's request, SAP provided additional information about the modification techniques and tools and other techniques available within the R/3 System, and demonstrated how to make modifications to the system through ABAP programming.
¶ 6. In 1995, Menasha's Board of Directors approved the license of the R/3 System. SAP projected that the cost to implement the system was $46,575,000. Menasha then entered into an agreement to license the R/3 System. The agreement contained no provision for SAP to customize the R/3 System.
¶ 7. When SAP was demonstrating the R/3 System to Menasha, it advised Menasha that, due to the system's complexity and the substantial customization required to make it usable for Menasha's unique business operations, Menasha would have to retain either SAP consultants or consultants listed by SAP as SAP certified consulting partners or "logo" partners to implement the R/3 System with the custom features Menasha required. SAP also advised Menasha that since it could not supply all the consultants needed to install and customize Menasha's R/3 System, Menasha would have to work with one of SAP's logo partners. As advised, Menasha hired one of SAP's logo partners, ICS Deloitte.
¶ 8. From September 1995 to March 1996, Menasha worked with ICS Deloitte and SAP to analyze its systems, prepare Menasha's hardware for the R/3 installation, and begin introducing Menasha's technology team to the customization process. During this period, SAP representatives provided training and other support services to Menasha's information systems staff as they prepared for the R/3 installation. Menasha completed downloading the R/3 System's basic modules in March 1996.
¶ 9. The R/3 System was not ready for use as delivered. Implementation teams were created to determine the needs of each Menasha subsidiary and to configure and customize the R/3 System modules to meet their needs. The implementation teams were comprised of Menasha employees, SAP representatives, ICS Deloitte representatives, and third-party consultants.
¶ 10. When the implementation teams were unable to achieve the desired results, they were assisted by ABAP programming teams, which were directed by SAP and ICS Deloitte. The ABAP programming teams included Menasha information support staff and third-party consultants. Menasha also contracted with SAP to provide an on-site programmer, James Kammerer, who became a member of the Poly Hi Solidur ABAP programming team and helped provide programming fixes to Menasha's R/3 System.
¶ 11. The implementation and ABAP programming teams worked from March 27, 1996 to January 1, 1997, to customize the R/3 System to meet Menasha's needs, closing functional gaps through custom interface (sometimes referred to as "user exit"), by creating new subsystems to ran parallel with the R/3 System, or by employing other remedies. The implementation and programming teams made more than 3000 modifications to Menasha's R/3 System.
¶ 12. SAP also provided Menasha with patches to correct functional gaps identified during implementation that could not be fixed by the ABAP programming team. Some of the patches included new source code written specifically for Menasha's system to remedy shortfalls of the R/3 System as it applied to Menasha's business, replacing the source code of the original system, modules.
Testing the R/3 System; Training and Bringing it Online
¶ 13. Once the customization of Menasha's R/3 System was complete, Menasha worked with SAP and ICS Deloitte to test the system to determine whether it was operational in accordance with Menasha's required specifications. During this process, SAP representatives provided off-site and on-site technical support to Menasha's information support staff, and helped trouble-shoot problems that arose during the testing.
¶ 14. Menasha's information support staff and ICS Deloitte prepared written materials to help train Menasha's employees to use the R/3 System. The commission found that "[b]ecause of the significant customization of [Menasha's] R/3 System, SAP was unable to provide these materials [but] SAP did provide petitioner and ICS Deloitte with extensive materials focused on the running and maintaining of the R/3 System."
¶ 15. Finally, during the four-week process of bringing the R/3 System online, which required converting Menasha's old data to the new system, SAP representatives and ICS Deloitte provided support to Menasha's employees and helped remedy operational problems. The customization and installation of Menasha's R/3 System cost the company more than $23 million, of which only $5.2 million was for the core R/3 System. Menasha made payments for customization costs as follows: $2.5 million to SAP; $13 million to ICS; and approximately $775,000 to third-party consultants.
¶ 16. At least twenty-three SAP consultants were involved in the on-site implementation and installation of Menasha's R/3 System. The SAP consultants helped Menasha with programming, configuration, training, and testing. Menasha pays SAP a continuing maintenance fee for technical assistance, new releases, upgrades and patches to its customized R/3 System at a cost averaging $975,000 per year. Additionally, "[d]ue to the inherent complexity of the R/3 System and the significant modifications made to customize [Menasha's] R/3 System," SAP continues to provide Menasha with technical assistance two to three times a week through online or telephone support.
Audit of SAP
¶ 17. The commission's decision included the following undisputed facts submitted by the DOR regarding the DOR's audit of SAP in 1998:
65. During 1998, respondent audited the American subsidiary of SAP for liability under Wisconsin's sales and use taxes for the period 1991 through 1997.
66. In the course of the audit, SAP and respondent agreed that SAP's sales of R/3 software in Wisconsin were subject to Wisconsin's sales tax as sales of noncustom software. Respondent reached this determination for [reasons including R/3 sales being (1) "off-the-shelf standardized software, i.e., written before the sale and intended for a wide variety of customers; (2) lack of pre-shipment modification to R/3 modules; (3) the delivery of existing R/3 software modules to all customers regardless of individual contracts; (4) SAP's delivery of "keys" to customers that would access particular modules delivered; (5) that development tools like ABAP/4 were only enhancements which did not change the R/3 source code; (6)-(7) that the implementation of original base modules is a complex, lengthy process with many stages and processes; (8)-(9) that it is rare for customers to implement R/3 on their own, so customers have the option of hiring SPA consultants or logo partners to help with implementation; (10) that "SAP keeps the licensing and maintenance agreement separate for many contracts and/or agreements for consulting services;" (11) that SAP's transactions with Menasha were similar to other R/3 sales, "and in each case, SAP agreed that the sale was of noncustom software"; (12) that SAP's release of software upgrades, routine maintenance updates, and patches, are general in nature; and (13) changes to customer-specific functionality and user exit type programs are the customer's responsibility].
67. As a consequence of the agreement by SAP that sales of the R/3 program modules were taxable, SAP paid respondent more than $1.9 million in tax and interest for sales to Wisconsin customers and agreed to collect sales and use tax thereafter. This figure did not include sales to petitioner here at issue, in that petitioner provided SAP with a statement that petitioner would directly pay the sales tax, which petitioner did. . . .
Procedural History
¶ 18. In 1998, Menasha filed a refund claim for $342,614.45 in use tax paid on its acquisition of the R/3 System from SAP and for payment of maintenance fees. The DOR denied Menasha's refund claim and its petition for redetermination. Menasha appealed to the tax appeals commission. Menasha and the DOR filed cross-motions for summary judgment. The commission granted Menasha's motion, concluding that the R/3 software is exempt custom software under WIS. STAT. § 77.51(20) as defined by WIS. ADMIN. CODE § Tax 11.71(1)(e) and (k).
The parties have since entered into a settlement agreement regarding the taxation of the maintenance fees.
DISCUSSION
I. Standard of Review
¶ 19. This case involves the interpretation and application of an administrative rule, WIS. ADMIN. CODE § Tax 11.71 and related subsections, which define "custom software" and "prewritten software" for purposes of determining whether computer software is tangible property and therefore taxable under WIS. STAT. § 77.51(20). This appeal is taken from a circuit court decision reversing a decision of the Wisconsin Tax Appeals Commission. The scope of our review is the same as the circuit court's. Target Stores v. LIRC, 217 Wis. 2d 1, 11, 576 N.W.2d 545 (Ct.App. 1998). We review the commission's decision, not the circuit court's. Advance Pipe Supply Co. v. DOR, 128 Wis. 2d 431, 484, 383 N.W.2d 502 (Ct.App. 1986).
WISCONSIN STAT. § 77.51(20) provides:
"Tangible personal property" means all tangible personal property of every kind and description and includes electricity, natural gas, steam and water and also leased property affixed to realty if the lessor has the right to remove the property upon breach or termination of the lease agreement, unless the lessor of the property is also the lessor of the realty to which the property is affixed. "Tangible personal property" also includes coins and stamps of the United States sold or traded as collectors' items above their face value and computer programs except custom computer programs.
The text of WIS. ADMIN. CODE § Tax 11.71 is provided and addressed at length in the next section.
¶ 20. The appropriate standard of review is a subject of dispute in this case. The DOR argues that, because the commission lacks expertise in determining whether software is prewritten or custom under the recently enacted amendment to WIS. STAT. § 77.51(20), the commission's interpretation of that statute and WIS. ADMIN. CODE § Tax 11.71(1)(e) is not entitled to any deference. The DOR further argues that, because it has extensive experience construing this tax rule, its interpretation of the rule is entitled to controlling weight. In a separate section of its brief, the DOR argues that its "prior construction of the `significant modification by the vendor' language in its own rule is controlling." Menasha argues that the DOR's interpretation is entitled to no weight and that the commission's interpretation of the tax rules is entitled to controlling weight.
¶ 21. We conclude that the DOR's interpretation and application of its tax rules are not entitled to any deference. We also conclude that the commission's interpretation and application of the tax rules are entitled to due weight deference.
¶ 22. On review of an administrative agency's interpretation of its own rules, we generally accord controlling weight, unless the interpretation is clearly erroneous or inconsistent with the language of the regulation. Pfeiffer v. Board of Regents of Univ. of Wis. Sys., 110 Wis. 2d 146, 154-55, 328 N.W.2d 279 (1983) (citations omitted); see also Orion Flight Servs., Inc. v. Basler Flight Sew., 2006 WI 51, ¶ 18, 290 Wis. 2d 421, 714 N.W.2d 130; Hillhaven Corp. v. DHFS, 2000 WI App 20, ¶ 12, 232 Wis. 2d 400, 606 N.W.2d 572. WISCONSIN STAT. § 73.01(4), which defines the powers and duties of the tax appeals commission, states that "the commission shall be the final authority for the hearing and determination of all questions of law and fact [on appeals to the commission]." WIS. STAT. § 73.01(4)(a).
¶ 23. Judicial decisions have reiterated the scope of the commission's powers and duties. In Kamps v. DOR, we explained that the tax appeals commission "is charged by the legislature with being the final authority for all questions of law involving taxes. . . ." Kamps v. DOR, 2003 WI App 106, ¶ 14, 264 Wis. 2d 794, 663 N.W.2d 306; see also Bender v. DOR, 2005 WI App 31, ¶ 13, 278 Wis. 2d 731, 693 N.W.2d 311. We further explained in Kamps "that in construing [a tax exemption statute] the commission has employed the expertise gained from discharging that duty over the years." Kamps, 264 Wis. 2d 794, ¶ 14. In DOR v. Caterpillar, Inc., we reviewed a decision by the commission involving the construction of a statute. DOR v. Caterpillar, Inc., 2001 WI App 35, ¶ 6, 241 Wis. 2d 282, 625 N.W.2d 338. The DOR and the commission rendered conflicting interpretations of that statute. Id., 6 n. 3. In clarifying the standard of review in this context, we explained that "[b]ecause the commission is the final administrative authority that reviews the decisions of the DOR, any deference that might be due to the decision of an administrative agency is due to the commission, not to the DOR." Id. (citation omitted). Although this standard ordinarily applies in the context where there is a conflict between the DOR's and the commission's construction of a tax statute, we see no principled reason why this standard should not apply where there is a conflict between the DOR's and the commission's construction of a tax rule.
¶ 24. Consequently, both WIS. STAT. § 73.01(4) and the above case law plainly establish that the commission is the final authority on all the facts and questions of law regarding the tax code; the DOR is not. Any deference due to an administrative agency's decision in this context is given to the commission, not the DOR. Our deference is based on the commission's extensive expertise achieved by discharging its legislative duty as the final authority on the facts and questions of law in resolving tax disputes. Furthermore, under statutory provisions of WIS. STAT. § 73.015 and ch. 227 regarding judicial review, the "agency" decision we review is the commission's, not the DOR's. Therefore, it is the commission's interpretation and application of an administrative tax rule we review and give deference to. The next question we must answer, then, is what degree of deference we should give to the commission's interpretation and application of WIS. ADMIN. CODE § 11.71.
¶ 25. When reviewing an administrative agency's determination, we may accord great weight deference, due weight deference, or de novo review. Caterpillar, Inc., 241 WIS. 2d 282, ¶ 6. We accord great weight deference when the legislature charges the agency with the duty to administer a statute, the agency's interpretation is longstanding, "the agency employed its expertise or specialized knowledge in forming the interpretation," and "the agency's interpretation will provide uniformity and consistency in the application of the statute." Bender, 278 Wis. 2d 731, ¶ 11. "Under the great weight standard, we uphold an agency's reasonable interpretation of the statute if it is not contrary to the clear meaning of the statute, even if we conclude another interpretation is more reasonable." Id., ¶ 12.
¶ 26. We apply due weight deference "when the agency has some experience in the area but has not developed the expertise that necessarily places it in a better position than the court to make judgments regarding the interpretation of the statute." Id., ¶ 11. Under this standard, "we uphold the agency's reasonable interpretation if it comports with the purpose of the statute and we conclude there is not a more reasonable interpretation." Id., ¶ 12.
¶ 27. The third standard we may apply in reviewing an agency's interpretation and application of a statute is de novo deference. Id., ¶ 10. In other words, we give no deference to the agency's determination. We apply this standard when the issue before the agency is one of first impression, the agency lacks expertise, or the agency's determination is so inconsistent as to provide no guidance on the topic. Caterpillar, Inc., 241 Wis. 2d 282, ¶ 7.
¶ 28. We recognize that the commission is the final authority on questions of law relating to tax rules. As we have explained, an administrative agency's interpretation of its own rules is ordinarily controlling unless the interpretation is clearly erroneous or is inconsistent with the language of the rule. Pfeiffer, 110 Wis. 2d at 154-55; Orion Flight Servs. Inc., 290 Wis. 2d 421, ¶ 18; Hillhaven Corp., 232 Wis. 2d 400, ¶ 12. Thus, ordinarily we would accord controlling or great weight to the commission's interpretation and application of the tax rules. However, the commission did not promulgate the rules at issue here. Moreover, even though the commission has had some experience in construing WIS. ADMIN. CODE § 11.71(1)(e), its experience is not of longstanding. Consequently, it is not appropriate to give controlling weight to the commission's interpretation of the tax rules.
¶ 29. We conclude that the commission's decision is entitled to due weight deference. In addition to being designated the final authority on all questions of law involving taxes, the commission has generated and employed its substantial experience discharging its duty in construing the rules governing the taxability of tangible property since WIS. STAT. § 77.51(20) was enacted. See, e.g., DOR v. River City Refuse Removal, Inc., 2006 WI App 34, ¶¶ 12, 14, 289 Wis. 2d 628, 712 N.W.2d 351, review granted, 2006 WI 108, 292 Wis. 2d 409, 718 N.W.2d 723 (No. 2004AP2468); All City Commc'n Co. v. DOR, 2003 WI App 77, ¶¶ 6, 11, 263 Wis. 2d 394, 661 N.W.2d 845.
¶ 30. While the commission may not yet have interpreted the 1992 amendment to WIS. STAT. § 77.51(20), which specifies that custom computer programs are included in the statute's exemptions, the commission has accrued significant experience interpreting and applying the definition of "custom computer programs" under WIS. ADMIN. CODE § Tax 11.71. See, e.g., IBM Corp. v. DOR, Wis. St. Tax Rep. (CCH) 202-854 (W.T.A.C. Mar. 23, 1987), aff'd, 145 Wis. 2d 903, 428 N.W.2d 646 (Ct.App. 1988); Health Micro Data Sys., Inc. v. DOR, Wis. St. Tax Rep. (CCH) 203-062 (W.T.A.C. May 23, 1989). In addition, in All City Communication Co., we explained that because of the commission's extensive experience administering § 77.51(20) and other tax statutes, even where the commission has not previously decided a case under the particular facts at issue, we should accord at least due weight deference. All City Commc'n Co., 263 Wis. 2d 394, ¶¶ 6, 11; see also Telemark Dev. Inc. v. DOR, 218 Wis. 2d 809, 820, 581 N.W.2d 585 (Ct.App. 1998) ("we have recognized in a series of cases that an agency's experience and expertise need not have been exercised on the precise — or even substantially similar — facts in order for its decisions to be entitled to judicial deference") (citation omitted). Consequently, while the commission may not have extensively addressed the text of the § 77.51(20) amendment in its past decisions, its extensive experience interpreting the sales and use tax exemption in general, and the definition of custom software under § Tax 11.71 specifically, supports our according due weight deference in this case.
II. DOR's Objections to the Commission's Undisputed Material Facts
¶ 31. We address first the DOR's argument that there are genuine issues of material fact and, therefore, that the commission improperly granted summary judgment. Both parties moved for summary judgment. As part of its scheduling order, the commission ordered each party to submit proposed undisputed facts along with that party's initial summary judgment brief. Both parties in their respective summary judgment submissions included proposed undisputed facts. The DOR accepted the majority of Menasha's proposals, but objected to several others. The DOR proposed a number of changes to the factual descriptions with which it disagreed; the DOR also submitted documents in support of its proposed changes. In deciding which proposed undisputed facts to accept, the commission described its method this way:
Where the parties agree, there is an effective stipulation of facts. Where there is a material disagreement, the commission has reviewed the submission of the parties and has made the alterations it deemed appropriate. The resolution of these disagreements is described in detail in the Ruling. The Commission, however, has been careful not to weigh the competing submissions of the parties. Rather, the Commission has sought to determine whether the submission of the parties present competing inferences.
(Footnote omitted.) The commission then addressed each of the DOR's alleged factual disputes, explaining when it did not adopt the DOR's proposed changes to Menasha's proposed undisputed facts and why it did not accept those proposed changes. The commission ultimately rendered what it has labeled as a "Summary of Undisputed Material Facts."
¶ 32. On appeal, Menasha argues that by filing its own motion for summary judgment, the DOR in effect agreed that there were no genuine issues of material fact. Menasha further argues that there were in fact no genuine issues of material fact. Menasha asserts that where DOR disagreed with Menasha's proposed undisputed facts, the commission examined whether the evidence presented competing inferences and carefully examined the evidence for each of the DOR's proposed changes before concluding that those proposed changes were either immaterial or unsupported by competent record evidence.
¶ 33. The DOR counters that, because each respective party's motion for summary judgment was supported by a different set of facts, it is inaccurate to assume that DOR was stipulating that Menasha's proposed undisputed facts created no dispute of material fact. We agree with the DOR. We therefore consider whether DOR's objections to the commission's Undisputed Material Facts create a genuine issue of material fact, which would warrant remand to the commission for an evidentiary hearing. We conclude that the commission properly examined each of the DOR's objections to Menasha's proposed undisputed facts and provided sufficiently reasonable explanations as to why those proposed changes by the DOR that the commission rejected did not rise to the level of genuine issues of material facts.
¶ 34. On appeal, the DOR's only objections relate to Undisputed Material Facts #29 and #47. We thus turn our attention to those two Facts.
Undisputed Material Fact #29
¶ 35. Undisputed Material Fact #29 states that
[Menasha] understood, prior to licensing the R/3 System, that the customization process could take years to complete and would cost tens of millions of dollars, and petitioner's budget for the purchase of the R/3 System included the amounts it expected to pay to SAP and SAP's designated consultants for the configuration, modification, and customization of the system. [Menasha] also understood prior to its licensing of the R/3 System that, without this customization, the system would be of no value to its operations. The customization was necessary to justify any amount spent on the licensing of the basic R/3 System modules.
¶ 36. The DOR proposed that the commission revise Fact #29 to state that any amounts in Menasha's 1995 budget for payments to SAP were for training and consultation, not customization. The DOR continues to maintain on appeal that Fact #29 should be reworded as suggested because otherwise it would conflict with Undisputed Material Facts #31 and #32. However, in making this argument, the DOR misquotes Fact #31 as stating "that SAP had advised Menasha . . . that SAP could not perform the modifications that Menasha desired. . . ." This quote misstates the text of Fact #31. What Fact #31 actually states is that SAP advised Menasha "it was unable to supply all the consultants necessary for the installation and customization. . . ." (Emphasis added.) The commission's text of Fact #31 does not support, the DOR's concerns.
¶ 37. Regarding Fact #29, the DOR relied on an affidavit from the DOR's auditor, Abba Nof. In that affidavit Nof averred that "SAP invoices and contracts were overwhelmingly for training and information sharing., not for performing implementation, configuration or integration of the system." The commission, however, concluded that Nof "has not demonstrated that he has any personal knowledge of the specifics of the arrangement between petitioner and SAP" The commission concluded that the DOR's argument did not apply to the actual wording of Fact #29, which refers to Menasha's understanding and its budget, not to what was actually paid to SAP The commission consequently rejected the DOR's proposed change to Fact #29 after concluding that there was no genuine factual dispute.
Undisputed Material Fact #47
¶ 38. Undisputed Material Fact #47 states that "[t]hroughout the installation and customization of [Menasha]'s R/3 System, SAP representatives provided both off-site and on-site technical and functional support to petitioner directly and through ICS Deloitte. SAP consultants served on the ABAP programming team and performed ABAP programming to customize the R/3 System to [Menasha's] business."
¶ 39. The DOR argues that Nof's affidavit renders infirm Fact #47's statement that SAP consultants performed ABAP programming to customize the R/3 system. We first observe that this is an inaccurate characterization of Nof's affidavit. Nof actually averred that there is "no evidence . . . that SAP was to oversee or to take responsibility for any phase of the implementation," and that there is a difference between modifying R/3 and creating "user exits." Nof also averred that the "only person specifically identified by the Petitioner . . . as an SAP consultant that did programming during the implementation of the R/3 software at Petitioner's office — James Kammerer — did not join the implementation team until December 2, 1996, . . . three months after the claimed `customization' was completed. . . ."
¶ 40. In addition, the commission explained that it did not adopt the DOR's suggested language that "there is no evidence that any SAP consultant performed" ABAP programming to customize the R/3 system to Menasha's business because the affidavit the DOR gave in support of its change was not on point or reflective of Nof's direct knowledge. On appeal, the DOR continues to maintain that Nof had enough direct and expert knowledge of Menasha's operations to provide the information at issue. However, we need not resolve this issue because we conclude that the language of Nof's actual affidavit does not conflict with the language of Undisputed Material Fact #47, and the commission's conclusion that the suggested language was not on point is reasonable.
¶ 41. In sum, we agree with the commission that the arguments the DOR raises in relation to Undisputed Material Facts #29 and #47 are either unsupported by the evidence or do not otherwise create a genuine issue of material fact.
III. Interpretation of WIS. ADMIN. CODE § Tax 11.71
¶ 42. Menasha argues that the commission's construction and application of WIS. ADMIN. CODE § Tax 11.71 are reasonable. Menasha argues that the commission reasonably construed the introduction to § Tax 11.71(1)(e) as requiring the commission to consider "all facts and circumstances" — i.e., the nonexhaustive factors of § Tax 11.71(1)(e)1.-5. along with subsections (1)(e)6. and 7. and (k) — in determining whether a program is a custom program.
¶ 43. The DOR argues that the commission erred in concluding that the R/3 System was customized software. More specifically, the DOR contends that the taxability of a sale under Wisconsin's tax laws is determined solely by reference to a taxpayer's obligations under the particular transaction. It asserts that the commission erred by failing to recognize that the purchase of computer software by Menasha was a standalone transaction and that future transactions between Menasha and other vendors for installation and other services are separate transactions, subject to possible taxation. The DOR also argues that WIS. ADMIN. CODE § Tax 11.71(1)(e)6. and 7. control the determination of whether a computer program is prewritten or customized, in isolation from the other factors, and that under subsections (1)(e)6. and 7., the R/3 System is prewritten and therefore taxable.
¶ 44. We conclude, applying due weight deference, that the commission's interpretation and application of WIS. ADMIN. CODE § 11.71(1)(e) and (k) are reasonable and consistent with the plain language of the rule. We also conclude that the DOR fails to offer a more reasonable interpretation of this rule.
¶ 45. All sales of tangible personal property are taxed unless specifically exempted. WIS. STAT. §§ 77.52(1) and 77.53(1). Exemptions to taxation statutes are to be strictly construed against the taxpayer. Wolter v. DOR, 231 Wis. 2d 651, ¶ 18, 605 N.W.2d 283 (Ct.App. 1999). Computer programs are tangible personal property and therefore taxable under WIS. STAT. § 77.51(20), except for custom computer programs. Custom computer programs are not taxable under § 77.51(20) if the programs meet the definition of "custom computer programs" in WIS. ADMIN. CODE § Tax 11.71(1)(e).
¶ 46. WISCONSIN ADMIN. CODE § Tax 11.71(1) provides in pertinent part:
(e) "Custom programs" mean utility and application software which accommodate the special processing needs of the customer. The determination of whether a program is a custom program shall be based upon all the facts and circumstances, including the following:
1. The extent to which the vendor or independent consultant engages in significant presale consultation and analysis of the user's requirements and system.
2. Whether the program is loaded into the customer's computer by the vendor and the extent to which the installed program must be tested against the program's specifications.
3. The extent to which the use of the software requires substantial training of the customer's personnel and substantial written documentation.
4. The extent to which the enhancement and maintenance support by the vendor is needed for continued usefulness.
5. There is a rebuttable presumption that any program with a cost of $10,000 or less is not a custom program.
6. Custom programs do not include basic operational programs or prewritten programs.
7. If an existing program is selected for modification, there must be a significant modification of that program by the vendor so that it may be used in the customer's specific hardware and software environment.
. . . .
(k) "Prewritten programs," often referred to as "canned programs," means programs prepared, held or existing for general use normally for more than one customer, including programs developed for in-house use or custom program use which are subsequently held or offered for sale or lease.
¶ 47. In its decision, the commission provided the analytical framework for interpreting and applying WIS. ADMIN. CODE § Tax 11.71(1)(e) in the following way:
Section Tax 11.71(1)(e) . . . defines custom programs as "utility and application software which accommodate the special processing needs of the customer." In determining whether a program meets the definition of a custom program, we are to consider "all the facts and circumstances," including seven factors listed in the rule. The factors set forth in the rule are not elements, each of which must be met for a program to be considered custom. Rather, the factors in the rule are to be weighed along with any other facts and circumstances. . . .
¶ 48. The DOR appears to ignore the first five factors in WIS. ADMIN. CODE § Tax 11.71(1)(e) and instead focuses solely on § Tax 11.71(1)(e)6. and 7. The DOR contends that, if the R/3 System meets the definition of prewritten as set forth in § Tax 11.71(1)(k), "then it is taxable regardless of any other provision in the rule." The DOR argues that § Tax 11.71(1)(e)6. and 7. are mandatory and provide no exceptions, and therefore trump the other factors set forth in § Tax 11.71(1)(e)1.-5. At oral arguments and in its brief on appeal, the DOR contends that § Tax 11.71(1)(e)6. "has meaning only if its application is controlling over the preceding criteria in the rule." We conclude that the DOR's construction of § Tax 11.71(1)(e) is not more reasonable than the commission's.
¶ 49. As the commission pointed out, WIS. ADMIN. CODE § Tax 11.71(1)(e) requires that the commission consider "all the facts and circumstances" in determining whether a computer program comports with the definition of a custom program, including factors 1.-7. of the tax rule. We observe that the plain language of § Tax 11.71(1)(e) imposes this requirement. We agree that § Tax 11.71(1)(e)6. unambiguously states that "[c]ustom programs do not include basic operational programs or prewritten programs." We also agree that § Tax 11.71(1)(e)7. is mandatory in its requirement that "[i]f an existing program is selected for modification, there must be a significant modification of that program by the vendor so that it may be used in the customer's specific hardware and software environment." However, the commission determined that, because it had concluded that the R/3 System was a custom program, § Tax 11.71(1)(e)7. did not apply in this case. Aside from its narrow and unreasonable interpretation of § Tax 11.71(1)(e), the DOR offers no reason for why we should ignore the commission's determination that the tax rule shall be considered in its entirety in determining whether a computer program is customized. We now turn to the commission's analysis of each of the tax rule's factors.
¶ 50. The first factor to consider in determining whether a computer program is a custom program is "[t]he extent to which the vendor or independent consultant engages in significant presale consultation and analysis of the user's requirements and system." WIS. ADMIN. CODE § Tax 11.71(1)(e)1. The commission found that DOR conceded that significant presale consultation and analysis occurred. The DOR does not contest this finding.
¶ 51. The second factor to consider is "[w]hether the program is loaded into the customer's computer by the vendor and the extent to which the installed program must be tested against the program's specifications." WIS. ADMIN. CODE § Tax 11.71(1)(e)2. The commission concluded first that "the fact that a former SAP employee loaded the [R/3] software weighs in favor of a finding that the software at issue is custom software." The commission also rejected the DOR's contention that Menasha "did not provide documentation that SAP tested the R/3 System against program specifications." In the commission's view, the second factor is not concerned with who tested the software program against program specifications, "but the degree to which the installed program must be tested." The commission noted that the DOR conceded "that, once customization of the R/3 System for [Menasha's] business was complete, the system was tested for three to four months to insure it met petitioner's operational requirements." Based on the above, the commission concluded that the second factor supports the conclusion that the R/3 System is custom software. The DOR does not dispute this conclusion.
On appeal, the DOR argues that the "by the vendor" language in WIS. ADMIN. CODE § Tax 11.71(1)(e)7. should be construed as requiring that in order for computer software to be non-taxable as customized software, the software must be modified and installed by the vendor who sold the software. This argument has implications for how § Tax 11.71(1)(e)2. should be construed, but the DOR does not argue this point in relation to that subsection. The DOR makes this argument in relation to § Tax 11.71(1)(e)7. We therefore limit our consideration of this argument in relation to § Tax 11.71(1)(e)7.
¶ 52. Addressing the third and fourth factors, WIS. ADMIN. CODE § Tax 11.71(1)(e)3.-4., the commission concluded that the DOR conceded that the R/3 System required substantial training and written documentation, as well as enhancement and maintenance support. The DOR does not contest this conclusion.
¶ 53. The fifth factor creates a "rebuttable presumption that any program with a cost of $10,000 or less is not a custom program." WIS. ADMIN. CODE § Tax 11.71(1)(e)5. The commission concluded that this section does not apply, "because it is undisputed that the cost of the R/3 System greatly exceeded this threshold." The DOR does not dispute this conclusion.
However, the commission also concluded, in light of all the facts and circumstances of the transaction, that "[i]f there is a rebuttable presumption that a program costing less than $10,000 is not a custom program, then since the cost of the R/3 System to [Menasha] was $5.2 million, the significant price tag must argue in favor of a conclusion that the R/3 System is a custom program." The DOR does not address this conclusion. However, we disagree with the logic underlying the commission's conclusion. The plain language of WIS. ADMIN. CODE § Tax 11.71(1)(e)5. does not support the commission's logic that the cost of the program must be a factor merely because the factor establishes a rebuttable presumption that a program costing less than $10,000 is not a custom program. There is nothing in the text of this factor, or any reasonable inference arising from the factor's language, that indicates the agency intended for the cost of the computer program to otherwise be a factor in determining whether a computer program is custom software. Although we disagree with the commission's conclusion, we observe that its conclusion is dicta because it is not central to resolving the dispute at hand. Therefore, our conclusion that this part of the commission's interpretation of § Tax 11.71(1)(e)5. is unreasonable does not disturb our holding that the commission's decision is reasonable.
¶ 54. The real dispute in this case centers on the interpretation and application of WIS. ADMIN. CODE § Tax 11.71(1)(e)6. and (k). Section Tax 11.71(1)(e)6. states that "[c]ustom programs do not include basic operational programs or prewritten programs." Section Tax 11.71(1)(k) states:
"Prewritten programs," often referred to as "canned programs," means programs prepared, held or existing for general use normally for more than one customer, including programs developed for in-house use or custom program use which are subsequently held or offered for sale or lease.
¶ 55. The commission concluded that the R/3 System was custom software, not prewritten, within the meaning of § Tax 11.71(e)(1)6. Because the commission concluded that the software was custom, it further concluded that the § Tax 11.71(e)(1)7. exception for prewritten software later modified by the vendor did not apply. The DOR challenges these conclusions.
¶ 56. In determining whether the R/3 System was a custom or prewritten program under WIS. ADMIN. CODE § Tax 11.71(1)(e)6., the commission considered the factors set forth in § Tax 11.71(1)(e)1.-5., as well as the definition of "prewritten programs" in § Tax 11.71(1)(k). The commission first observed that the DOR improperly relied on the "plain English" meaning of the term "prewritten." The commission concluded that the only applicable definition of "prewritten" is the one provided in § Tax 11.71(1)(k), i.e., "prepared, held or existing for general use normally for more than one customer."
¶ 57. The commission next observed that the critical factor in determining whether a computer program is customized is the degree of difficulty in making the software ready for use. Put another way, each of the factors in determining whether a program is customized "hinges on the degree to which the software is ready for use off-the-shelf." The commission explained, "[t]he more pre-sale planning, the more testing, the more training, the more written documentation, the more enhancement, and the more maintenance, then the more likely the software is custom software." The commission concluded that "[t]he distinction between custom and prewritten programs hinges on the amount of effort necessary to get the software operational for a particular customer's needs. Given the substantial amount of resources, time, and effort needed to bring the R/3 System online, we cannot conclude that the software at issue is prewritten."
¶ 58. We conclude, applying the due weight standard, that the commission's construction of WIS. ADMIN. CODE § Tax 11.71(1)(e) and its conclusion are reasonable and that the DOR does not provide a more reasonable interpretation of the tax rule. As we have explained, § Tax 11.71(1)(e) requires the commission to consider "all the facts and circumstances" to determine whether it is customized for tax purposes. The commission did just that. We agree with the commission's conclusion that § Tax 11.71(1)(e)6. should not be considered in isolation of the other factors in the rule. When the factors are read together, the most reasonable construction of the rule is that the definitions of "custom" and "prewritten" software in § Tax 11.71(e) and (k) are informed in large part by consideration of the first five factors of § Tax 11.71(1)(e). The commission reasonably concluded that the more effort required making the software usable, the more likely it is custom software. This conclusion is consistent with the plain language of the rule. The first five factors of § Tax 11.71(1)(e) emphasize the degree to which certain effort is made to make a computer program usable by the buyer.
¶ 59. Here, although the R/3 System is a modular system, which is "prewritten" in the sense that it is made up of "standard software modules" and has been mass-marketed to thousands of different businesses, the record indicates that over 3000 modifications were made to the software before Menasha could use it for its intended manner. It is also undisputed that at the time of delivery, Menasha could not use the R/3 System in its current form. The commission therefore concluded that because "a program like the R/3 system is useful only after a significant investment of resources . . . then the software cannot be said to be prepared, held or existing for general use." (Emphasis added.) The commission pointed to other undisputed facts demonstrating that the R/3 System was not "prewritten," but customized:
1. The basic modules of the R/3 System must be subject to a certain degree of customization, and it is only after this customization process is complete that the client has a usable software system;
2. As delivered, the R/3 System was inadequate for petitioner's use;
3. Members of petitioner's implementation team working with SAP and ICS Deloitte determined the operational and functional needs of each subsidiary in order to configure and customize the system;
4. The implementation team worked to configure and modify the R/3 System to adapt the system to each subsidiary's identified needs;
5. The implementation and ABAP programming teams worked to customize the R/3 System to meet petitioner's functional needs;
6. The ABAP programming teams created codes for hundreds of user exits to integrate external programs with the R/3 System, so that petitioner was able to realize the functionality needed for its unique business while preserving the functional efficiencies of the R/3 System;
7. The ABAP programming teams created new subsystems to run parallel to the R/3 System for operations not available within the R/3 System more critical to petitioner's business;
8. The ABAP programming teams customized fields and reports within the R/3 System to insure it produced output to be useful to petitioner's business;
9. SAP provided petitioner with patches to correct functional gaps identified during implementation, some of which included new source code written to address shortfalls of the R/3 System;
10. In total, more than 3,000 modifications were made to petitioner's R/3 System.
We are satisfied that the commission reasonably interpreted WIS. ADMIN. CODE § Tax 11.71(1)(e) and (k) and concluded that the R/3 System was custom software and therefore not taxable as tangible property under WIS. STAT. § 77.51(20).
¶ 60. The DOR makes several arguments in support of its position that the commission erred by concluding that the R/3 System is custom software for tax purposes. It first makes the unsurprising argument that where there are two separately taxable transactions, each transaction is analyzed separately to determine the taxability of that transaction. The DOR contends that the commission ignored this fundamental principle of sales tax law. As an example, the DOR points to the situation where a vendor sells prewritten software which requires customization. It explains that, if the vendor later enters into a transaction agreeing to customize the software, the two transactions are analyzed separately. By contrast, the DOR explains, "if the vendor enters into one sales transaction providing for both the sale of prewritten software and its subsequent customization, that transaction cannot be divided into separate parts." According to the DOR, WIS. ADMIN. CODE § Tax 11.71(1)(e)7. treats the sale and customization of the prewritten software as a single transaction, "but only on the express condition that `there must be a significant modification of that [existing] program by the vendor.'" The DOR summarizes its argument this way:
The vendor must collect the tax when the sale occurs. The vendor's obligations at that time determine whether tax should be collected. If there is only one contract at that time, those contractual obligations are controlling. There is no other workable framework under which sales tax law or the rule can be administered.
¶ 61. We fail to see how this construction of WIS. ADMIN. CODE § Tax 11.71(1)(e)7. resolves the dispute at hand. As we have observed, the commission reasonably determined that the R/3 System was custom software within the meaning of the rule and therefore did not need to apply § Tax 11.71(1)(e)7. In addition, as we have explained, the commission reasonably concluded that the entire rule must be considered, not just one or two subsections, to determine whether a computer program is customized. The DOR's construction of the rule fails to consider § Tax 11.71(1)(e)7. in the context of the entire rule. More importantly, the agency has as a matter of public policy established the criteria for determining whether computer software is customized by creating the first five factors of § Tax 11.71(1)(e). There is nothing in the text of this rule and its subsections evincing the requirement that the commission determine whether a computer program is custom by interpreting and applying only the last two provisions of § Tax 11.71(1)(e) in isolation of the other factors, particularly when § Tax 11.71(1)(e)7. only applies once software has been determined to be existing or prewritten software.
¶ 62. Having rejected the DOR's construction of the rule that a vendor's obligation at the time of the sales transaction determines the taxability of that transaction, we also reject the DOR's argument that Menasha never established that the software was significantly modified by SAP as part of the 1995 sales transaction. We agree with the DOR that, if WIS. ADMIN. CODE § Tax 11.71(1)(e)7. applied to the present facts, its reading of the rule might reasonably cause the R/3 System to be taxed as tangible property because the evidence of whether SAP, as the vendor, modified the system is ambiguous. However, the commission reasonably concluded that the software is not prewritten; thus we do not address the question of whether SAP modified the system.
The DOR argues that its "prior construction of the `significant modification by the vendor' language in its own rule is controlling." We reject this argument. As we have explained, the DOR's construction of its own rules is entitled to no deference. Because the commission is the final authority whose decision is subject to judicial review, we give deference to the commission's construction of the DOR's rules, not the DOR's.
¶ 63. The DOR next argues that Menasha failed to prove that the R/3 software was not prewritten at the time it entered into the initial license agreement with SAP This argument rests on the DOR's assertion that, if the R/3 System meets the definition of "prewritten" as defined in WIS. ADMIN. CODE § Tax 11.71(1)(k), "then it is taxable regardless of any other provision in the rule." The DOR contends that because § Tax 11.71(1)(e)6. is dispositive, it "has meaning only if its application is controlling over the preceding criteria in the rule." We reject this construction of the tax rule. Again, as we have concluded, the commission reasonably concluded that the proper method for determining whether a computer program is custom is to consider the entire rule, not just parts of it. The DOR's construction of § Tax 11.71(1)(e) ignores the rules of statutory interpretation. The DOR offers no reason for deviating from these rules.
As with statutory interpretation, we interpret the language of a regulation in the context in which it is used, "not in isolation but as part of a whole; in relation to the language of surrounding or closely-related [regulations]; and reasonably, [so as] to avoid absurd or unreasonable results." State ex rel. Kalal v. Circuit Court for Dane County, 2004 WI 58, ¶ 46, 271 Wis. 2d 633, 681 N.W.2d 110; see also Brown v. Brown, 177 Wis. 2d 512, 516, 503 N.W.2d 280 (Ct.App. 1993) (we apply the same rules in interpreting administrative rules as we do in interpreting statutes).
¶ 64. In arguing that the R/3 System is prewritten software, the DOR construes the phrase "general use" in WIS. ADMIN. CODE § Tax 11.71(1)(k) to mean "general purpose." It also contends that because the software was "existing" and "mass-marketed," the R/3 System was in "general use" and therefore "prewritten." In its view, any software that may be used for various purposes and is mass-marketed is prewritten software. The DOR asserts that "the essence of the transaction involving development of software from scratch is the sale of a pure computer programming service. Programming services are not taxable." In other words, according to the DOR, computer programs not created from scratch, regardless of the number and extent of modifications made to the software to render it usable, are prewritten and therefore taxable. We are not persuaded that this is a more reasonable interpretation of § Tax 11.71(1)(k) than the commission's.
¶ 65. As we have explained, according to the commission, whether a computer program exists for general use rests on the extent to which "significant investment of resources in planning, testing, training, enhancement, and maintenance" is required. This construction of the tax rule recognizes that computer programs such as the R/3 System may contain standardized components, but if the program is not usable as sold and requires substantial effort to render it usable, then the software is not prewritten. The DOR has conceded that the software was not usable as purchased; it also conceded that as delivered, the R/3 System was inadequate for Menasha's needs and required extensive modification and customization before Menasha could use the system. These facts support our conclusion that the commission's decision that the R/3 System was customized was reasonable and that the DOR's is not more reasonable.
¶ 66. The DOR appears to equate general use with mass-marketing. In other words, because the R/3 System, has been mass-marketed, in the DOR's view that means the software is prepared, held or existing for general use. We agree that whether a computer program is mass-marketed may be considered a factor under the rule that all the facts and circumstances should be considered in determining whether a computer program is customized. We emphasize, however, that simply because a computer program is mass-marketed, it does not necessarily follow that the program is prewritten. Here, there is no dispute that the R/3 System has been mass-marketed, yet the commission apparently did not give much weight to that factor. The commission gave greater weight to other factors, such as the need to make significant modifications to the software to render it usable to Menasha, in determining whether the software was prewritten. We cannot say this was less reasonable than DOR's proposed approach. As we have explained, the commission properly determined that the entire rule is to be considered in determining whether a computer program is prewritten or customized. As we have also explained, the commission reasonably determined that the first five factors of WIS. ADMIN. CODE § Tax 11.71(1)(e) inform the decision of whether a computer program is prewritten or customized.
CONCLUSION
¶ 67. We conclude that the commission reasonably interpreted and applied WIS. ADMIN. CODE § Tax 11.71(1)(e) and (k) in determining that the R/3 System sold by SAP to Menasha was customized software, and that the DOR's construction of this tax rule is not more reasonable than the commission's. We also conclude that the commission properly granted summary judgment to Menasha. We therefore reverse the circuit court's decision reversing the commission's decision, and affirm the commission's decision granting a refund to Menasha for taxes paid on the R/3 System.
By the Court. — Order reversed.