SPRINT CORPORATION AND SUBSIDIARIES, F.K.A. UNITED TELECOMMUNICATIONS, INC., PETITIONER v. COMMISSIONER OF INTERNAL REVENUE, RESPONDENT, United States Tax Court, 108 T.C. 384, Docket No. 13159-94., Filed April 30, 1997
SPRINT CORPORATION AND SUBSIDIARIES, F.K.A. UNITED TELECOMMUNICATIONS, INC., PETITIONER v. COMMISSIONER OF INTERNAL REVENUE, RESPONDENT
United States Tax Court
108 T.C. 384
Docket No. 13159-94.
Filed April 30, 1997.
P, a telephone company, purchased certain telecommunications equipment, digital switches, that required computer software to operate. P claimed investment tax credits (ITC) and depreciation deductions under the accelerated cost recovery system (ACRS) with respect to the total cost of each digital switch, which included the cost of the software used in each switch. R determined that P's expenditures allocable to the software did not qualify for the ITC or depreciation under the ACRS. P treated property known as "drop and block" as 5-year property, as defined in sec. 168(c)(2)(B), I.R.C. R determined that the property was 15-year public utility property. For the years in issue, the property was depreciated under the ACRS.
1. Held, P's expenditures allocable to the software qualify for the ITC and depreciation under the ACRS. Norwest Corp. & Subs. v. Commissioner, 108 T.C. 358 (1997), is followed.
2. Held, further, drop and block is 5-year property under sec. 168(c)(2)(B), I.R.C.
Jay H. Zimbler, Michael A. Clark, Michael R. Schlessinger, and William J. McKenna, Jr., for petitioner.
Alan M. Jacobson, John W. Duncan, and Patricia Pierce Davis, for respondent.
HALPERN, Judge: Respondent determined deficiencies in petitioner's Federal income taxes for the years and in the amounts as follows:
Sprint Corp. (Sprint or petitioner) is a Kansas corporation with its principal office in Westwood, Kansas. Formerly known as United Telecommunications, Inc., petitioner officially changed its name to Sprint Corp. as of February 26, 1992. Unless otherwise noted, references herein to Sprint and petitioner will include the period during which petitioner was known as United Telecommunications, Inc. Petitioner filed consolidated Federal income tax returns for itself and its eligible subsidiaries for the 1982, 1983, 1984, and 1985 taxable years.
After concessions by the parties, the issues remaining for decision are (1) whether certain expenditures made by petitioner during the years in issue that are allocable to the cost of computer software used in central office equipment (COE or digital switches) qualify for the investment tax credit (ITC) and depreciation under the accelerated cost recovery system (ACRS) and (2) the proper classification as recovery property of certain telecommunications equipment known as "drop and block". Unless otherwise noted, all section references are to the Internal Revenue Code in effect for the years in issue, and all Rule references are to the Tax Court Rules of Practice and Procedure.
FINDINGS OF FACT 1
1 The stipulation of facts and accompanying exhibits are incorporated herein by this reference. The trial Judge made the following Findings of Fact, which we adopt.
During the taxable years in issue, petitioner and its subsidiaries were engaged in the business of providing local and long-distance telephone service. Petitioner generally operated its business of providing telephone service through separately incorporated, wholly owned subsidiaries. The local companies are generally known by names indicating the parent company and their geographic location (e.g., United of Iowa or UT of Florida) and will be so referred to herein where reference to a specific subsidiary is necessary. In all other instances, references to Sprint and petitioner shall be deemed to include petitioner's subsidiaries.
A. The Digital Switch Issue
The equipment that provides the switching function that enables one telephone subscriber to connect to another has undergone an evolution from its earliest form, when the switching function was performed by human operators sitting at manual switchboards. Manual switching was automated with the advent of electromagnetic relay switches. Since 1980, switching in the United States has increasingly been done by digital switches consisting of a number of integrated circuits, clocks, processors, and central processing units (hardware). Digital switches operate in accordance with programmed instructions initially encoded on magnetic tape (software). The central processing units are specially designed computers that are used, and can only be used, to control the switch function.
During the years in issue, petitioner purchased from different vendors several digital switches (hardware and software) for use in its telephone business, generally to replace existing electromechanical switches. Investment tax credits and depreciation deductions under the ACRS were claimed with respect to the total cost of each switch. In the notice of deficiency, respondent disallowed that portion of the claimed investment tax credits and accelerated depreciation deductions relating to the costs of the software.
Broadly, the three basic components of a telecommunications system are station equipment, transmission facilities, and switches. Station equipment is generally located on the customer's premises and includes such items as the telephone at a residential customer's house. Transmission facilities provide the paths over which information is transmitted between customers (whether the medium is used for local network transmission or transmission over trunks, which are lines between different networks) and consist of transmission media such as copper and fiber optic cable, as well as the equipment used to amplify and regenerate the transmitted signals. Switches connect transmission facilities at key locations and route incoming and outgoing calls.
The basic objective of the telephone switch is to connect any calling outlet with any wanted inlet, a process that can be visualized by picturing an operator sitting at an old manual switchboard. As a call is made, the operator pulls a flexible cord connected to the caller's line and physically plugs it into a receptacle connected to another line in the same network or to a trunk line if the recipient is in a different network.
The process of switching actually involves four sequential phases, each consisting of certain activities or functions. The first phase, preselection, encompasses activities related to recognizing a new call request and determining how to route it. The second phase of switching is call completion, which entails the actual connection of the requesting outlet to the wanted inlet utilizing the determined routing, and initiation of the charging process. The third phase of switching is conversation. The fourth and final phase is release, which is the disconnection of the call, completion of the charge record, and restoration of the network to the normal (idle) state.
In addition to switching activities, modern switching equipment must also perform certain management functions (such as automatically detecting and isolating system and component malfunctions), as well as provide certain customer services (such as call forwarding or coin return at a pay telephone when the call is not completed).
The implementation of the various activities and functions of the switching process is complicated by certain system requirements, including availability, reliability, privacy, and economy, which at times may conflict with one another. The first requirement, availability, refers to the need to have sufficient paths so that a connection can be made on demand. The reliability requirement refers to the need to assure that the system as a whole, or a particular connection, does not go down (fail). The privacy requirement reflects the need to switch correctly (to the desired customer exclusively) or not at all. Intentional misconnections, such as those caused by attempts to avoid proper charging, as well as accidental misconnections, must be prevented. The economy requirement refers to the need to provide service at a competitive price. To that end, it is necessary to avoid overbuilding a system despite the desire for availability, reliability, and privacy. Availability and privacy, which generally are presumed requirements, voice quality, price, and reliability are the bases by which systems compete, with reliability being the key differentiating basis. The reliability standards for telephone switches are far more stringent than for most modern computer systems.
An automated telephone switch comprises (1) the switching network, (2) various interfaces, and (3) control mechanisms. The heart of the switch is the network, which consists of individual devices designed to connect (and disconnect) communication paths. Before integrated circuit switching, automated switching was accomplished through a series of electromagnetic relays. When a call was placed, a physical connection path would be formed by closing the appropriate relays. In its most simplified form, i.e., a network system consisting of only two telephone customers, the operation of the switch would involve nothing more than closing and opening the relay switch on the line running between the two customers. When the relay was closed, the lines of the two customers would be connected and the switch would have functioned, enabling conversation.
A local telephone system may consist of hundreds of thousands of customers. It would be cost prohibitive either to connect each subscriber directly to every other subscriber or to have a centralized switch within a network with the same number of direct lines as there are combinations of customers. Rather, to satisfy the economic considerations of a telephone system, the switch function in a given area is centralized at an office (the central office) which receives calls placed by customers and then routes those calls through one of a number of outlets to other subscribers within the local network or to other local networks via long-distance trunks.
The earliest automated switching systems used direct progressive control to operate the switch, whereby relay switches along the path connecting the calling and the called parties would be closed as each digit in a telephone number was dialed until a complete connection was made.
To address certain disadvantages of progressive control systems, telephone designers in the 1940's began incorporating registers, devices which store and release dialed telephone numbers into telephone switches. Using registers, systems could be devised for looking ahead to ascertain the best possible routing. Moreover, with such use, common control, or control of various telephone functions by a centralized mechanism shared by separate lines, was possible. By the 1950's, switches used electromechanical switches and relays to accomplish the key control functions of a switch on a common basis, including determining routing, seizing trunk lines, and ringing the call user. Even identifying, measuring, and recording a long-distance toll call was accomplished through common control mechanical devices.
With the advent of solid-state electronics, some or all of the common control functions were accomplished by utilizing integrated circuits on which processing (control) instructions were encoded. Instead of electromechanical devices opening and shutting in a predetermined (programmed) fashion to respond to various alternative situations, electronic circuits would be opened or closed in accordance with the embedded logic.
In the mid-1960's, new telephone switches began to use specially designed processors called stored program control (SPC), which execute programs encoded on magnetic tape or other media, rather than wired-logic, to control certain switch functions. At first, the tendency in digitalized switch design was to centralize most switch control functions in one central processing unit. By the end of 1985, it was deemed more beneficial if certain control functions were performed by decentralized processors controlled by the central processing unit.
The advantages of SPC were that, because its program was loaded by tape, rather than in electromechanical devices or hard-wired integrated circuits, the program could be more easily maintained (or changed), and the speed of electronics could be more easily harnessed to allow a large network to be controlled by a single high-speed processor.
Design and Architecture of the Modern Digital Switch
Modern digital switches are designed from the ground up by a handful of manufacturers around the world. In the 1980's, the major North American manufacturers were ITT Corp., Northern Telecom, Inc. (NTI), American Telephone & Telegraph Co., GTE Corp., TRW Vidar, Inc., and Stromberg Carlson, Inc.
Beginning in the late 1970's, engineers designing a particular model of a switch did not merely arrange standard electronic components into a standard structure. Rather, a particular overall structure was conceived, and then the components of each of the interfaces, the network devices, and the control mechanisms (including the encoded program) were specially engineered in the context of that concept. As a result of design choices and the proprietary nature of certain custom-designed components (many of which are patented), the architecture of a modern digital switch produced by one manufacturer differed (and still differs) significantly from the architecture of a switch produced by another.
Because each switch was designed for the particular parameters (number of subscribers, usage patterns, potential for growth) of a given central office location, there could be differences in architecture between switches produced by the same manufacturer but installed in two different locations.
In addition to choices between the actual logic, engineers designing the encoded programs for the control mechanisms of a modern digital switch also had the option to place various parts of the encoded logic in software (i.e., the encoded medium that will be loaded into the central processing unit) or firmware (permanently encoded chips located in the central processing unit or in microprocessors located throughout the switch). The choices made were determined to a large extent by the switch's architecture. A switch manufacturer writes the programs and decides how they would be incorporated into the processors in the context of the design requirements of a particular model. Just as the architecture of the switch affects the programming, the programming limitations and requirements affect the architecture of the switch. The programming is also affected by the given location. Due to the unique parameters of a given location, programs are invariably location specific.
Because of the interrelationship of the architecture and programming and the unwillingness of manufacturers to sell the switch hardware without the switch software, programming for a digital switch is written by the manufacturer of the switch and is not available from third parties. Absent the manufacturer's programming, the manufacturer's hardware cannot operate. Because the design of the programming is specific to the architecture and even the location of the digital switch, not only can the programming of one manufacturer not be used on the equipment of another, but the programming of a switch in one location generally cannot be used on a switch of the same type in another location.
Digital Switch Acquisitions (1982-85) ≈ Generally
The particular design of the digital switches acquired by Sprint during the years in issue varied from manufacturer to manufacturer and from model to model. However, the facts relating to the DMS-100, manufactured by NTI, including its acquisition and software, are representative of the purchases in issue.
Although there were many design configurations and software/firmware combinations, none of that concerned petitioner, which was interested in a turnkey switch; that is, Sprint desired that the vendor would engineer, furnish, install, and prepare a complete switch for service. Sprint's primary purpose in the procurement process was to determine which manufacturer could satisfy petitioner's requirements for the lowest price. Based on those criteria, Sprint selected a manufacturer and negotiated a final price.
Once an agreement was reached, a software load for the switch was developed by the manufacturer, NTI in this instance. First, NTI technicians reviewed Sprint's order to ascertain the number of lines and trunks and the types of desired features and prepared a written plan or blueprint to be used in compiling the software load. The software load was then made by downloading a base load module from a software library onto a blank magnetic tape, pulling down other existing modules from the software library to meet design specifications, and manually installing various "data information" and translation codes onto the tape. Three copies of the software load were made, two being sent to Sprint and one being retained solely for safety reasons by NTI. NTI technicians then installed the custom software load on the particular switch for which it was designed. Once installed, the custom software load was tested and validated by NTI technicians. An invoice from NTI indicates that, on a particular switch, of 532 total hours to develop the software load, 500 hours were spent writing the software blueprint and testing and validating the installed software. There was little or no time actually spent writing new software for the digital switch, and there was no evidence as to how many preexisting modules were used.
Once the equipment became operational, software load updates were periodically provided by the manufacturers. If there was a new feature to be added, it was done at the time of the periodic update. The time required to effect those modifications varied between 100 and 160 man-hours. If the hardware of the switch was moved to a different geographic location, a new custom software load would be prepared utilizing the same steps.
Nature of Sprint's Possession of the Software
The direct sales agreement between petitioner and NTI provided a warranty as to the software loads. Provided that Sprint adequately maintained the equipment, NTI warranted that the software would function in accordance with the specifications applicable on the shipment date, and, upon its failure to do so, NTI would correct the failure and act to ensure that the software was operating as specified.
Section 8 of the sales agreement, labeled "Software License", provided that Sprint was granted a nonexclusive paid-up license to use the software for its intended purpose, as long as the switch was in use. Software is defined as computer programs contained on a magnetic tape, disk, semiconductor device, or other memory device or system memory. Sprint agreed that the software provided by NTI was to be treated as the exclusive property of NTI. To that end, Sprint promised to hold the software in confidence for the benefit of NTI; not provide or make the software available to any person except to its employees on a "need to know" basis; not modify the software; not reproduce or copy the software in whole or in part; and return to NTI any magnetic tape, disk, semiconductor device or other memory device or system, and documentation or other material, which had been replaced, modified, or updated. Sprint was not required to protect NTI's interest in any data or information that became available to the general public, commonly known as "public domain" software. In the event that NTI modified or changed the software to permit additional features or services, the updated software was to be made available to Sprint at NTI's then-current price for those features or services.
Although there were significant limitations on Sprint's rights in the software, under section 8.6 of the sales agreement, Sprint was given the right, upon transfer of title to the digital switch, to assign the license for the software or, upon a lease or other nonpermanent transfer, sublicense the software, provided the assignee or sublicensee agreed in writing to the above terms.
Around 1988 or 1989, Sprint effected a trade of title of approximately 30 DMS-100 type digital switches with ConTel Co., another telephone company. None of the switches were actually moved, and service to the customers was not interrupted. Although neither party specifically notified NTI, the trade received sufficient attention in the industry to give NTI at least constructive notice of the trade. NTI did not object to the trade.
Respondent determined in the notice of deficiency that the costs of the software are not eligible for either the ITC or accelerated depreciation under the ACRS because (1) Sprint received a license to use the software rather than owning the software and (2) in any event, the software is not tangible property.
Sprint acknowledges that the substantial value of encoded programming on software, in general, relates to the programming's conceptual or intangible value, but nonetheless contends that (1) Sprint owned the software in issue and (2) the software qualifies as tangible personal property.
Respondent concedes that if Sprint in fact owned the software and the software constitutes tangible property, Sprint is entitled to the ITC and accelerated depreciation, as claimed. If Sprint did not own the software in issue, or if the software is not tangible personal property, respondent contends that Sprint is entitled to amortize the cost of the software on a straight-line basis over an 18-year period, while Sprint contends that the costs should be amortized in accordance with Rev. Proc. 69-21, sec. 4.01(2), 1969-2 C.B. 303. The 18-year period is the class life asset depreciation range (CLADR) midpoint life of the switch hardware with which the software is associated. Rev. Proc. 69-21, sec. 4.01(2), supra, is the procedure pursuant to which, during 1982 through 1985, Sprint capitalized and amortized the cost of purchased software, other than the software purchased in connection with the digital switches (rather than claiming the ITC and accelerated depreciation under the ACRS).
B. Drop and Block Issue
A telephone network includes transmission facilities and station equipment (or station apparatus). Transmission facilities consist of the wiring and ancillary equipment used to transmit telephone signals between the telephone company's central office and the customer's (whether caller or callee) station apparatus (i.e., telephone, modern, or other device). Transmission facilities consist of three distinct segments: (1) The main cable, either buried or aerial, (2) the wire (drop wire) running from the distribution network to and including the station protector (also known as the block) located on the outside wall of the customer's premises (together the "drop and block"), and (3) the wire running from the station protector to and around the inside of the customer's premises (the inside wiring).
As a public utility, the telephone industry is regulated by the Federal Communications Commission (FCC). As part of its regulatory function, the FCC prescribes the accounting treatment of revenues earned and expenses incurred in the operation of a telephone business. The relevant rules are set forth in 47 C.F.R. part 31 (part 31), Uniform System of Accounts for Class A and Class B Telephone Companies. During the years in issue, all of petitioner's subsidiaries were telephone companies subject to part 31 rules.
As of January 1, 1981, part 31 specified that the investment in drop and block be accounted for in FCC account No. 232 (station connections). As of January 1, 1981, and in accordance with part 31, petitioner so accounted for its investment in drop and block. Beginning January 1, 1984, part 31 was changed and specified that investment in drop and block be accumulated in FCC account No. 242 (aerial and buried cable). For income tax purposes, petitioner treated property in FCC account No. 232 as depreciable over 5 years and property in FCC account No. 242 as depreciable over 15 years. In petitioner's 1984 and 1985 Federal income tax returns, the following subsidiaries of petitioner continued to account for drop and block in FCC account No. 232, treating it as 5-year property for depreciation purposes:
In petitioner's 1984 and 1985 Federal income tax returns, the following subsidiaries of petitioner accounted for drop and block in FCC account No. 242, treating it as 15-year public utility property for depreciation purposes:
Respondent concedes that, if the drop and block property placed in service during the years in issue is properly classified by reference to its pre-January 1, 1984, inclusion in FCC account No. 232 and CLADR Asset Guideline Class 48.13, then, with the exception of the depreciation described in the preceding paragraph, the depreciation claimed with respect to that drop and block per return as filed was correct; petitioner is then entitled to adjustments increasing depreciation for the companies listed in the preceding paragraph for which drop and block was treated as 15-year public utility property on petitioner's Federal income tax returns as filed for 1984 and 1985.
Petitioner concedes that, if the drop and block placed in service during the years in issue is properly classified by reference to its post-December 31, 1983, inclusion in FCC account No. 242 and CLADR Asset Guideline Class 48.14, then the depreciation claimed by petitioner in its Federal income tax returns for 1984 and 1985 with respect to such drop and block would be reduced, with the result that taxable income would be increased with respect to the companies which treated the drop and block as 5-year property.
I. Digital Switches
Petitioner purchased central office equipment (COE or digital switches) for use in its business of providing telephone service. Petitioner claimed investment tax credits (ITC) and depreciation deductions under the accelerated cost recovery system (ACRS) with respect to the total cost of each digital switch, which included the cost of the custom computer software load (software load) necessary to make each switch operable. In the notice of deficiency, respondent disallowed that portion of the claimed investment tax credits and accelerated depreciation deductions relating to the costs of the software loads.
The parties agree that, if this Court determines that petitioner owned the software loads in issue and that those software loads constitute tangible personal property, petitioner is entitled to the claimed investment tax credits and accelerated depreciation deductions. We conclude that petitioner owned the software loads in issue and that those software loads constitute tangible personal property. Therefore, we hold that petitioner is entitled to the claimed credits and deductions.
Today, in Norwest Corp. & Subs. v. Commissioner, 108 T.C. 358, 375 (1997), we decided that operating and applications software that was subject to license agreements entitling the taxpayer to use the software on a nonexclusive, nontransferable basis for an indefinite or perpetual term qualifies for the ITC as tangible personal property. In holding that the taxpayer's acquisition of the software without any associated, exclusive, intangible intellectual property rights was precisely the type of investment Congress intended to encourage in enacting the ITC, we noted that "Intangible intellectual property rights and the tangible or physical manifestations or embodiments of those rights are distinct property interests." Id. at 375 (citing 17 U.S.C. sec. 202 (1994)). We see no material distinction between the software in the Norwest case and the software loads in issue here. In this case, however, respondent has raised the question of whether petitioner acquired sufficient benefits and burdens of ownership with respect to the software loads to be considered the owner of those loads for purposes of the ITC and the ACRS.
The parties agree that the issue of ownership of the software loads is governed by the substance of the sales agreements between petitioner and the various digital switch manufacturers, not the labels used in those agreements. See, e.g., Tomerlin Trust v. Commissioner, 87 T.C. 876, 881-883 (1986); see also Leahy v. Commissioner, 87 T.C. 56, 66 (1986) (transfer of the benefits and burdens of ownership govern for Federal tax purposes, rather than the technical requirements of passage of title under State law). The parties also agree that the direct sales agreement between petitioner and Northern Telecom, Inc. (NTI), in effect from January 1, 1983, to December 31, 1985 (the Sprint/NTI agreement), is representative of all of the agreements pursuant to which petitioner acquired the digital switches in issue. Therefore, we must determine whether petitioner owned the software load transferred by NTI to petitioner (the NTI software load). Whether petitioner became the owner of the NTI software load is a question of fact to be ascertained by reference to the Sprint/NTI agreement, read in light of the attending facts and circumstances. See, e.g., Grodt & McKay Realty, Inc. v. Commissioner, 77 T.C. 1221, 1237 (1981).
Petitioner acquired from NTI magnetic tapes containing copies of the computer software necessary to make the digital switch operable, i.e., the NTI software load, and did not acquire any of the underlying, exclusive, intangible intellectual property rights. See, e.g., 17 U.S.C. sec. 101 (1994) (a nonexclusive license is not within the definition of the term "transfer of copyright ownership"). Petitioner possessed all of the significant benefits and burdens of ownership with respect to those magnetic tapes. NTI simply did not retain a residuary interest in the NTI software load commensurate with an interest typically retained by a lessor of property. See Crooks v. Commissioner, 92 T.C. 816, 819 (1989) ("interest retained by the transferor is the primary distinction between a sale and a lease").
First, petitioner paid a fixed amount for the digital switch, which included the cost allocable to the NTI software load, and did not incur any obligation to make further payments for that load, contingent or otherwise. More importantly, petitioner acquired the exclusive right to use the NTI software load for the useful life of the digital switch, which was tantamount to acquiring the perpetual right to use that particular load because of the interrelationship between the switch hardware and software. In addition, petitioner had the right to transfer the software load in conjunction with a transfer of the digital switch without NTI's consent. Respondent attempts to characterize that right as a "right without substance" because any new owner of the digital switch would have to acquire a new software load, unless the switch was used in the same location. That condition, however, appears relatively unrestrictive in light of Sprint's trade of approximately 30 similar digital switches with ConTel Co., in the late 1980's, which resulted in none of the switches' actually changing location. Lastly, the Sprint/NTI agreement provided that the risk of loss with respect to the digital switch, including the NTI software load, would pass to petitioner upon delivery.
Respondent points to certain provisions in the Sprint/NTI agreement as evidence that the benefits and burdens of ownership did not pass from NTI to petitioner. In particular, respondent focuses on certain provisions in the Sprint/NTI agreement that provide that the software transferred with the digital switch manufactured by NTI was to be treated as the exclusive property and trade secret of NTI and that petitioner was under certain obligations to protect NTI's interest in the software. The provisions of the Sprint/NTI agreement cited by respondent all relate to NTI's interest in the intellectual property underlying the NTI software load. Those provisions protect, reinforce, and extend NTI's intellectual property rights in that software. NTI's retention of those rights is consistent with the conclusion that petitioner owned the NTI software load. Cf. Conde Nast Publications, Inc. v. United States, 575 F.2d 400, 407 (2d Cir. 1978) (limitations on transferee's rights in subject property that serve only to protect transferor's interest in other property do not divest transferee of ownership). Although those provisions created certain obligations with respect to petitioner's ownership of the NTI software load, such as holding that load in confidence and only making that load available to employees on a "need to know" basis, petitioner's ownership interest in that particular load remained intact. Moreover, many of the restrictions imposed on petitioner's use of that load were also imposed on the technical and proprietary information relating to the digital switch hardware; apparently, however, respondent does not question petitioner's ownership of that hardware.
In sum, petitioner acquired from NTI all of the significant benefits and burdens of ownership with respect to the NTI software load. The limitations on petitioner's use of that load served only to protect NTI's underlying intellectual property rights and did not divest petitioner of ownership in that particular load. We find that petitioner owned the NTI software load and did not purchase any exclusive, intangible intellectual property rights underlying that load. In accordance with the parties' agreement and our holding in Norwest Corp. & Subs. v. Commissioner, supra, we hold that petitioner owned all of the software loads used in all of the digital switches in issue and that those software loads constitute tangible personal property for purposes of the ITC and the ACRS. 2 Therefore, petitioner is entitled to the claimed investment tax credits and accelerated depreciation deductions.
2 In accordance with Norwest Corp. & Subs. v. Commissioner, 108 T.C. 358 (1997), we hold today that the software loads in issue constitute tangible personal property for purposes of the investment tax credit (ITC) and tangible property for purposes of the accelerated cost recovery system (ACRS). Although respondent does not assert that there exists a distinction in the meaning of the term "tangible" as it is used in the term "tangible personal property" for purposes of the ITC and the term "tangible property" for purposes of the ACRS, we believe that our reliance on the legislative history of the ITC in Norwest requires at least a brief discussion of that issue.
As a preliminary matter, sec. 168 and the regulations thereunder do not define the term "tangible property" for purposes of the ACRS. Sec. 168, which implements the ACRS, was enacted by the Economic Recovery Tax Act of 1981 (ERTA), Pub. L. 97-34, sec. 201, 95 Stat. 203. The relevant committee reports accompanying the enactment of ERTA indicate that Congress, in substantial part, considered the ITC and the ACRS to be in pari materia. See H. Rept. 97-201, at 73-74 (1981); S. Rept. 97-144, at 47 (1981), 1981-2 C.B. 412, 425; H. Conf. Rept. 97-215, at 206, 213 (1981), 1981-2 C.B. 481, 487, 490. This Court will not create a distinction unintended by Congress, and, thus, we conclude that the software loads in issue constitute tangible property for purposes of the ACRS.
II. The Drop and Block Issue
The drop and block portion of petitioner's telephone network consists of the wire running from petitioner's transmission network to a station protector located on the outside of a customer's premises. We must determine the depreciation class of the drop and block for purposes of the ACRS.
Taxpayers have long been allowed asset depreciation deductions in order to allow them to allocate their expense of using an income-producing asset to the periods that are benefited by that asset. * * * an allocation of depreciation to a given year represents that year's reduction of the underlying asset through wear and tear. * * * [ Simon v. Commissioner, 103 T.C. 247, 253 (1994), affd. 68 F.3d 41 (2d Cir. 1995).]
Such wear and tear, or "using up", can be thought of as being a gradual sale of the capital asset. United States v. Ludey, 274 U.S. 295, 300-301 (1927). The estimation of the wear and tear of the capital asset for a given period is based on the historical cost and does not take into consideration later fluctuations in valuation through market appreciation. Fribourg Navigation Co. v. Commissioner, 383 U.S. 272, 277 (1966). Originally, depreciation was calculated by apportioning the historical cost of the asset, less its salvage value, to the period the taxpayer expected to use the asset in his business. Massey Motors, Inc. v. United States, 364 U.S. 92, 107 (1960).
At one time, taxpayers were required to establish the useful life of the asset, which was the period the taxpayer expected to use the asset in his trade or business, and which did not necessarily coincide with the economic life of the asset. Id. at 104; sec. 1.167(a)-1(b), Income Tax Regs. For assets placed in service after December 31, 1970 (and before 1981), the asset depreciation range system (ADR) was the primary means of determining useful lives. Sec. 1.167(a)-11, Income Tax Regs. The ADR was meant to objectify and standardize the useful lives of assets by grouping assets into nearly 125 different asset guideline classes, each with its own guideline period. See sec. 1.167(a)-11(b)(4)(i)( b ), Income Tax Regs.; Rev. Proc. 77-10, 1977-1 C.B. 548. Taxpayers had to elect an asset depreciation period from the ADR assigned to each guideline class, which was 80 to 120 percent of the asset guideline period. Sec. 1.167(a)-11(b)(4)(i), Income Tax Regs. If no ADR was in effect, or if the taxpayer did not elect to use the ADR system, the useful life was determined with reference to the facts and circumstances surrounding the asset. Simon v. Commissioner, supra.
In 1981, Congress enacted the ACRS by the Economic Recovery Tax Act of 1981, Pub. L. 97-34, sec. 201, 95 Stat. 203. The ACRS was meant to stimulate the economy by allowing greater depreciation by taxpayers through shortened depreciation periods, as well as simplifying depreciation calculations by reducing the number of property classes from around 125 under the ADR system to 5. It was expected that the reduction in the number of property classes would help alleviate problems associated with the complexity of the ADR depreciation system. S. Rept. 97-144, at 47 (1981), 1981-2 C.B. 412, 425. The ACRS is mandatory and must be used for most depreciable property placed in service after 1980. Sec. 168(e)(1). Section 168, in relevant part, provides:
SEC. 168. ACCELERATED COST RECOVERY SYSTEM.
(a) ALLOWANCE OF DEDUCTION. ≈ There shall be allowed as a deduction for any taxable year the amount determined under this section with respect to recovery property.
* * * * * * *
(c) RECOVERY PROPERTY. ≈ For purposes of this title ≈
(1) RECOVERY PROPERTY DEFINED. ≈ * * * the term "recovery property" means tangible property of a character subject to the allowance for depreciation ≈
(A) used in a trade or business, or
(B) held for the production of income.
(2) CLASSES OF RECOVERY PROPERTY. ≈ Each item of recovery property shall be assigned to one of the following classes of property:
* * * * * * *
(B) 5-YEAR PROPERTY. ≈ The term "5-year property" means recovery property which is section 1245 class property and which is not 3-year property, 10-year property, or 15-year public utility property.
* * * * * * *
(E) 15-YEAR PUBLIC UTILITY PROPERTY. ≈ The term "15-year public utility property" means public utility property * * * with a present class life of more than 25 years.
The parties agree that drop and block is recovery property. To ascertain which class of property includes the drop and block, it is necessary to determine the present class life, defined in section 168(g)(2):
SEC. 168(g). DEFINITIONS. ≈ For purposes of this section ≈
* * * * * * *
(2) PRESENT CLASS LIFE. ≈ The term "present class life" means the class life (if any) which would be applicable with respect to any property as of January 1, 1981, under subsection (m) of section 167 (determined without regard to paragraph (4) thereof and as if the taxpayer had made an election under such subsection).
Section 167(m), as it applies to section 168(g)(2), provides that the Secretary shall prescribe the class lives for each class of property, which will reasonably reflect the anticipated useful life of the property class to the industry or group. Sec. 167(m)(1). The Commissioner issued numerous revenue procedures, pursuant to section 167(m), prescribing or modifying class lives. See Rev. Proc. 77-10, 1977-1 C.B. 548; Rev. Proc. 72-10, 1972-1 C.B. 721. As of January 1, 1981, the following pertinent asset guideline classes in Rev. Proc. 77-10, supra, were in effect:
See also Rev. Proc. 83-35, 1983-1 C.B. 745. 3 Under the regulations promulgated by the FCC for class A and class B telephone companies, 4 as of January 1, 1981, FCC account No. 232 included the original cost of drop and block wires. 47 C.F.R. sec. 31.232 (1980). Pursuant to Rev. Proc. 77-10, supra, FCC account No. 232 has an asset guideline period of 10 years, making it 5-year property under section 168(c)(2)(B), while property in FCC account No. 242 has an asset guideline period of 35 years, making it 15-year public utility property under section 168(c)(2)(E). The parties agree that prior to 1984, drop and block was properly included in FCC account No. 232 and, thus, was 5-year property. In 1984, the FCC regulations were amended so that drop and block was included in FCC account No. 242. 5 47 C.F.R. secs. 31.242:1, 3 (1984).
3 Rev. Proc. 83-35, 1983-1 C.B. 745, was meant to replace, with certain modifications, preceding revenue procedures, including Rev. Proc. 77-10, 1977-1 C.B. 548, that prescribed asset guideline classes, asset guideline depreciation periods, and ranges for the class life asset depreciation range system. This revenue procedure leaves intact the assets included and the depreciation range for asset guideline classes 48.13 and 48.14.
4 Companies having annual operating revenues exceeding $250,000 are class A telephone companies; Sprint is a class A company.
5 Drop and block is actually segregated into Federal Communications Commission (FCC) account No. 242.1, aerial cable, or 242.3, buried cable. For our purposes, because both accounts are treated in the same fashion, we need not make the distinction and will refer simply to FCC account No. 242.
Respondent argues that drop and block was FCC account No. 242 property when it was placed in service in 1984 and 1985, and that FCC account No. 242 property, on January 1, 1981, was 15-year public utility property. Respondent contends that she is still using the class lives that were in place on January 1, 1981; the FCC, by changing the accounting treatment of drop and block, has caused the property to be classified as 15-year property for the years after January 1, 1984. Respondent also cites various reports of the FCC which tend to support its accounting change.
Petitioner argues that the plain language of section 168(g)(2) requires that we apply to the property the class lives that were in effect as of January 1, 1981, thus making drop and block 5-year property. Finally, petitioner argues that Congress has twice amended section 168(g)(2), and those amendments should be read as requiring specific authorization from Congress for any departure from the January 1, 1981, class lives.
Public utilities, whose rates are often mandated by expenses, are routinely required to utilize uniform systems of accounting promulgated by regulatory agencies. See Pacific Enters. & Subs. v. Commissioner, 101 T.C. 1 (1993) (gas company); Kansas City S. Indus., Inc. v. Commissioner, 98 T.C. 242 (1992) (railroad); Oglethorpe Power Corp. v. Commissioner, T.C. Memo. 1990-505 (electric company); American Tel. & Tel. Co. v. Commissioner, T.C. Memo. 1988-35 (telephone company). For Federal tax purposes, if the generally accepted method of accounting of a taxpayer is made compulsory by a regulatory agency, and the method clearly reflects income, it is virtually presumed to be valid for Federal tax purposes. Commissioner v. Idaho Power Co., 418 U.S. 1, 15 (1974). The FCC accounted for drop and block in account No. 232 until 1984, when it was accounted for in account No. 242.1 or 242.3. Respondent looks to the administrative justifications for such changes. Although there may have been legitimate and persuasive reasons for the administrative change in accounting by the FCC, as well as some industry support, we do not need to reach that issue. It is not for us today to decide whether the action of the FCC was valid, justified, or authorized.
Instead the analysis begins with the plain meaning of the statute, and for reasons which follow, ends there. Respondent contends that the class lives have never changed; that is, account No. 242 has always been 15-year public utility property. That is only one-half of the analysis, for we must also apply those class lives to property as they would have been applied on January 1, 1981. The statute, in relevant part, provides that "`present class life' means the class life (if any) which would be applicable with respect to any property as of January 1, 1981". Sec. 168(g)(2) (emphasis added). We agree with petitioner; the language is not ambiguous, and accordingly we need not peer into the legislative history. Nevertheless, such an inquiry would support our analysis. The intent of the ACRS was to eliminate disagreement between taxpayers and the Commissioner and to stimulate economic activity. One essential theme of the ACRS was predictable depreciation periods; that was accomplished by freezing in time the property classifications as they were on January 1, 1981. Until further amendment by Congress, there were to be no changes.
Drop and block, as of January 1, 1981, was included in FCC account No. 232. That account had an asset guideline period of 10 years, making it 5-year property under section 168(c)(2)(B). We conclude that drop and block placed in service in the years in question is 5-year property for purposes of the ACRS.
Decision will be entered under Rule 155.
Reviewed by the Court.
SWIFT, PARR, WELLS, RUWE, WHALEN, BEGHE, FOLEY, VASQUEZ, and GALE, JJ., agree with this majority opinion.
CHIECHI, J., did not participate in the consideration of this opinion.
COLVIN, J., dissents.
KжRNER, J., dissenting: The majority, relying on Norwest Corp. & Subs. v. Commissioner, 108 T.C. 358 (1997), filed this date, concludes that the computer software in issue is tangible for purposes of the investment tax credit and for purposes of the accelerated cost recovery system (ACRS). I disagree with the conclusion reached in Norwest, and respectfully dissent from its application to this case.
I. Majority Opinion
The majority holds that based on Norwest, the software is tangible, and further that Sprint was the owner of the software. I did not have a vote in Norwest, and therefore was unable to voice my opposition at the time of its adoption. In Norwest, the Court offers an expansive analysis that discredits the intrinsic value test; unfortunately, its analysis of its own test is nowhere near as thorough. Indeed, one of the faults the Court found in Ronnen v. Commissioner, 90 T.C. 74 (1988), was that it lacked "rigorous analysis". Norwest Corp. & Subs. v. Commissioner, supra at 369. One would expect that the Norwest majority, in light of such murky reasoning as it perceived in Ronnen, would take the opportunity to clear the air with a definitive test, or at the very least offer some compelling reasoning to abandon the established precedent of this Court. Instead, they summarily conclude, with virtually no analysis, that the software was tangible. This conclusion is based on their interpretation of the legislative history of the investment tax credit that the tangibility requirement should be construed broadly. Their conclusion may also be based (it is not clear) on the fact that Norwest did not possess the right to distribute, sell, lease, or license the software it purchased (the "copyright rights").
A. Attack on the Intrinsic Value Test
The Norwest majority attacks this Court's implicit conclusion in Ronnen that computer programs were different from the seismic data at issue in Texas Instruments, Inc. v. United States, 551 F.2d 599 (5th Cir. 1977), and that the inextricable connection which existed between the seismic data and tapes in Texas Instruments was not present between the software and disks in Ronnen. Norwest Corp. v. Commissioner, supra at 369. I disagree with the Court's conclusion that there is no fundamental difference between seismic data and a computer program.
The distinction made in Ronnen v. Commissioner, supra, was appropriate and warranted by the facts. The seismic data consisted of the recording of a natural phenomenon. Although a recording of a natural phenomenon is the result of human exertion, it is neither the expression of an idea nor an unobvious improvement of prior technology or art. Accordingly, copyright or patent protection is not available for it. 1 Software, which is the result of human creativity (not mere exertion), can exist as source code on tapes, disks, or computer memory, or written out on paper. As the result of human creativity and design, copyright or patent protection 2 is available for it. I would therefore conclude that there exists a material difference between the sound recordings in Texas Instruments and the computer software purchased by Sprint, and at issue in Ronnen, and Norwest.
1 Although a recording of music is a recording of a natural phenomenon which can be copyrighted, it is the creative element that is copyrightable. See infra.
2 Traditionally, software, which is fundamentally a written set of instructions, was protected under copyright law, and infringement actions first were brought under copyright law. Later came a trend to allow patent protection for the design portion of computer applications. Petry, Taxation of Intellectual Property, secs. 1.08, 3.04 (1980).
B. Majority's "Traditional Approach"
The Norwest result is based upon an interpretation of the legislative history of the ITC that the term "tangible" should be construed broadly and possibly in the absence of copyright rights. I agree with Judge Jacobs and the other dissenters in Norwest that the majority's reading of the legislative history is inappropriate for the reasons stated therein. No purpose would be served to repeat those arguments. There is, however, an additional factor in Sprint not present in Norwest. Section 168 requires that property must be tangible to qualify for ACRS treatment. The majority points out that the ITC and ACRS were considered by Congress to be in pari materia, and therefore they extend their expansive construction of "tangible" property to ACRS. Because I disagree that the legislative history requires such a broad construction for purposes of the ITC, I similarly disagree with its extension to ACRS.
1. Absence of Intellectual Property Rights
The second basis of the Norwest majority's holding is that no copyright rights were passed to Norwest (or Sprint). In Norwest the Court failed to offer any analysis or cite any cases which indicate why the presence or absence of such rights should control the character of the tangible medium which, as the majority itself points out, is distinct and separate property from the copyright rights. Norwest Corp. v. Commissioner, supra at 374. The notion that copyright rights are separate and independent from a tangible embodiment is well supported. In Rev. Rul. 80-327, 1980-2 C.B. 23, the rights to manufacture and distribute books were acquired with the plates used in the printing of the books. The Service analyzed the two types of property separately and ruled that the plates were tangible, while the distribution rights were intangible. There simply is no rational basis to conclude that the presence or absence of one separate and distinct property interest, the intangible copyright right, should control the character of other separate and independent property.
Furthermore, this approach ignores the fact that the computer source code, which is intellectual property, is property separate and distinct from the copyright rights and the tangible medium. As the Court of Appeals for the Sixth Circuit indicated in Comshare, Inc. v. United States, 27 F.3d 1142, 1145 (6th Cir. 1994), computer software can consist of three types of property: The tangible computer tapes and disks, the intangible source code found on the disks, and the intangible copyright rights. Each of these types of property must be analyzed separately, unless there is some compelling reason to analyze them together. A computer program, which may be the creative expression of an idea, or an unobvious improvement on existing technology or art, can be protected by copyright and/or patent. Part of this property is the copyright rights. This intellectual property can exist in multiple forms, such as on disk, tape, computer memory, or written out on paper. An analysis must take place when a program is purchased as to what exactly was purchased. The components must be identified, and it must be determined whether there is any compelling reason to consider one or more of the components together.
The fallacy of the Norwest approach, and the majority here, is illustrated by considering that if the same property had been transferred to Norwest (or Sprint), coupled with a copyright right, then the property would become intangible. Further, consider that if software was purchased in one year, and the next year the right to reproduce and sell was acquired, where the controlling factor for character determination is the presence of that right, then the property would be tangible in year 1 and intangible in year 2, despite the fact that it was the same property.
2. Majority Opinion in Conflict With Case Law
The majority's reliance on the presence or absence of one or more intangible intellectual property rights to control character is in direct conflict with the case law. In Comshare, Inc. v. United States, supra, the taxpayer received (by purchase) the right to distribute the software. Despite the presence of this intangible property right, the court went on to conclude that the software was tangible. Thus, Norwest is in direct conflict with Comshare. The application of the rationale in Norwest to this case likewise brings this case in conflict with Comshare.
II. Conservative Approach
Rather than dispose of a hazy test which this Court adopted in Ronnen v. Commissioner, 90 T.C. 74 (1988), for another one which is just as hazy and not supported by any case law, I think we should clarify the test in Ronnen and attempt to distinguish Comshare, Inc. v. United States, supra. To do so would leave intact our own precedent, as well as its progeny that relied upon it.
A. Identify the Subject Property
The first step in applying the intrinsic value test is to identify the subject property, or in other words, determine what exactly the taxpayer has purchased or created. In Texas Instruments, Inc. v. United States, 551 F.2d 599 (5th Cir. 1977), the court found that seismic data did not exist without the tapes upon which the data was stored, and accordingly that the data and tapes were inextricably connected. The intangible information could not exist without the tangible medium. See Bank of Vermont v. United States, 61 AFTR 2d 88-788, at 88-790, 88-1 USTC par. 9169, at 83,250 (D. Vt. 1988). Thus, the subject property was the tape that contained the seismic data. In Comshare, the court found that the source code was the subject property, but like the seismic data in Texas Instruments, that it could not exist without the disks upon which it was stored. Thus, there the subject property was the unit consisting of the tapes and disks which contained the source code. In Ronnen v. Commissioner, supra, the subject property was the source code. We did not find the same "inextricable connection" that existed between the seismic data and tapes in Texas Instruments. We instead found that the disks containing the source code were but one type of conduit for the ideas contained on it. In Rev. Rul. 80-327, supra, plates to print books were purchased with the copyright rights to reproduce and sell copies of the books. The physical plates containing the creative work product were the subject property (while the copyright rights were analyzed separately).
In this case, the subject property is the source code. Sprint paid a fixed amount for the right to one working copy of that software. There was not simply one embodiment of the software (as was the case in Comshare ). Rather, the intellectual property existed in more than one locale, and Sprint purchased the right to use, or possess, a working copy of that intellectual property. Sprint possessed two copies, while the manufacturer possessed one. The record indicates that if Sprint had lost one of its copies, or if one had been destroyed, it would have been provided with another by the manufacturer. The copies were interchangeable. There was no significance as to which copy it used, where it existed, or in what form it existed. Sprint purchased the right to use the intellectual property of the manufacturer. The nexus between the intangible information and the tangible medium is far more attenuated here than in Texas Instruments, and like the software in Ronnen, is independent of the tapes upon which it was received. See Bank of Vermont v. United States, supra.
B. Characterize Property
Once the subject property is identified, it must be characterized. This is a facts and circumstances analysis, the focus of which is upon the relationship between the intangible intellectual property and any tangible medium upon which it exists. In Texas Instruments, the subject property, consisting of the intangible information (seismic data) and the tapes, was permanently embodied and inextricably bound. Therefore, the property was tangible. Although in Texas Instruments, the subject property was not software, we nevertheless borrowed the analysis and applied it to software in Ronnen. Although our analysis was less than thorough, 3 we concluded that "the intrinsic value of the [subject] software is attributable to its intangible elements rather than to its tangible embodiments." Ronnen v. Commissioner, supra. Although this phrase is somewhat ambiguous, because it appears immediately after the Texas Instruments analysis, I would interpret it to mean that in Ronnen, the integral connection between the intangible and tangible present in Texas Instruments did not exist between the software and the physical medium. The taxpayer's investment was not in an intangible which was inextricably bound to the specific tangible medium upon which it existed, but rather was in the intangible alone.
3 This discussion illustrates the hazards of adopting a standard that is less than clear. I cannot adequately emphasize how improper it would be to abandon our own precedent on the grounds that it is not fully developed only to replace it by an analysis that is equally undeveloped.
Turning to the software purchased by Sprint, an examination of what Sprint purchased, the right to a copy of source code that would operate its switches, leads to the conclusion that the property right is intangible.
Based on the foregoing, I cannot agree with the majority that software is tangible. Therefore, the software is not eligible for the ITC or ACRS treatment. I believe the following analysis regarding depreciation of the software, in light of its intangible character, is appropriate. Petitioner contends that its costs should be amortized over a period no shorter than 60 months pursuant to section 4.01(2) of Rev. Proc. 69-21, 1969-2 C.B. 303. Respondent contends that the software should be depreciated over 18 years, the asset guideline period for central office equipment (COE), for the software was an integral part of the COE.
Rev. Proc. 69-21, supra, provides in section 4.01:
(.01) With respect to costs of purchased software, the Service will not disturb the taxpayer's treatment of such costs if the following practices are consistently followed:
* * * * * * *
2. Where such costs are separately stated, and the software is treated by the taxpayer as an intangible asset the cost of which is to be recovered by amortization deductions ratably over a period of five years or such shorter period as can be established by the taxpayer as appropriate in any particular case if the useful life of the software in his hands will be less than five years.
Under this revenue procedure, if the taxpayer uses a period shorter than 5 years, he must establish that the useful life is less than 5 years; otherwise, the Commissioner will let stand the amortization period.
The software was separately stated on petitioner's books and in its purchase invoices from the assets from which it was purchased. The only issue is whether petitioner treated the software as an intangible asset. Petitioner did not originally amortize the software pursuant to this revenue procedure but rather treated the software and COE as one whole asset and depreciated that whole asset pursuant to ACRS. Sec. 168(c)(2)(B); Rev. Proc. 83-35, 1983-1 C.B. 745, 758. Petitioner amortized all other software as an intangible and amortized it pursuant to Rev. Proc. 69-21, supra, over a 5-year period.
Respondent argues that petitioner did not treat the software as an intangible and amortize it on its income tax returns for the years in issue, and that petitioner has not shown that a 5-year amortization period is appropriate. Petitioner does not need to show that a 5-year period is appropriate, for it did not claim an amortization period less than 5 years. 4
4 If I had to decide whether a shorter period was appropriate, I would take particular notice of the fact that the software loads were updated as often as every 6 months, but at least every 2 years.
Respondent looks to the treatment of the COE into which the software went. COE's belong to asset guideline class 48.12, which has an asset guideline period of 18 years for purposes of the class life asset depreciation range system. Rev. Proc. 83-35, 1983-1 C.B. 745. Under section 168(c)(2)(B), such property is treated as 5-year property and depreciated over 5 years. However, as we have held, the software is intangible and therefore does not qualify for ACRS. Accordingly, respondent has determined that ACRS treatment is not available for the software, but the software is still part of the COE, and therefore depreciable over 18 years.
COHEN, CHABOT, JACOBS, GERBER, and LARO, JJ., agree with this dissent.