Full opinion text
MEMORANDUM OPINION AND ORDER SHADUR, District Judge. American Medical Association (“AMA”) has sued the United States for a refund of federal income taxes imposed on AMA income earned in the years 1975 through 1978 from the sale of advertising in AMA periodicals. AMA contends the Internal Revenue Service (“IRS”) denied deductions for AMA’s costs of producing and distributing its periodicals that would have been allowed a commercial publisher and are therefore permitted AMA under Internal Revenue Code (“Code”) §§ 511-513, 26 U.S.C. §§ 511-513. Because IRS construed the Code-implementing Treasury-regulations to preclude those deductions, AMA urges alternatively: 1. IRS construed the regulations erroneously. 2. If IRS construed the regulations correctly, they are inconsistent with the Code and therefore invalid. To enable this Court to resolve those issues without a testimonial hearing, the parties have tendered a written evidentiary record and trial memoranda. After full consideration of those materials, in accordance with Fed.R.Civ.P. (“Rule”) 52(a) this Court sets forth the following Findings of Fact (“Findings”) and Conclusions of Law (“Conclusions”). Findings of Fact I. Background A. AMA 1. AMA is a not-for-profit and tax-exempt federation of state medical associations formed “to promote the science and art of medicine and the betterment of public health” (Stip. TITT1-2; Stip. Ex. 1-D at 1). From 1975 through 1978 about one-half the eligible physicians in the United States were members (Stip. 112). 2. In aid of its purpose, AMA publishes various medical periodicals. Deductions for two of those periodicals, JAMA and the American Medical News {“AM News”), aggregate more than 95% of AMA’s refund claim (AMA Mem. 4 n. *; Stip. 1110). Other AMA periodicals involved in the refund claim are Prism (discontinued in February 1976) and a number of specialty journals: Archives of Surgery, Archives of Dermatology, Archives of Ophthalmology, Archives of Otolaryngology and American Journal of Diseases of Children (Stip. IFlf 21(b) and 22). In 1977 the total cost of publishing AMA periodicals was $16,572,-508. Total other exempt activity costs that year were $39,185,000 (Stip. ¶ 32; Stip. Ex. 7). 3. Most AMA members pay dues to belong to the organization. In 1977 AMA members were in the following categories (Stip. ¶ 2): Category No. of members Dues Regular dues-paying 147,413 $250 Interns and residents 13,706 35 Medical students 14,303 15 Dues-exempt members 32,119 0 During the years at issue membership “dues include[d] the right to receive” without further payment JAMA, AM News, Prism (until it was discontinued) and one of the specialty journals (Stip. TÍTT10,18 and 22; Stip. Ex. 1-D at 6). Those same periodicals were sold to nonmembers for a subscription fee. 4. Dues collected from 1975 through 1978 were used in part for activity costs (including AMA’s publication-related costs), while the rest was placed in an “association equity” fund — a reserve fund of liquid assets to be used in the event of a future activity cost deficit (Sammons Aff. 11113-5). For example, of the $37,469,346 in dues collected during 1977, $22,640,728 was used to pay part of the $55,757,167 in total activity costs, with the remainder ($14,828,618) placed in the association equity account (Stip. If 4). Amounts added to the association equity account in 1975 through 1978 remained on AMA’s books as association equity or some other capital liability account through 1984 (Sammon Aff. If 6). In 1985, when AMA’s revenues were not sufficient to pay its activity costs, AMA reduced its association equity fund to reflect use of the reserves and treated (for accounting purposes) the funds so used as dues received that year (id.). B. Sale of Advertising in AMA Periodicals 5. To defray publishing costs, AMA sells advertising in its periodicals, primarily to pharmaceutical companies (Hannon Aff. 1ÍU 3 and 5). During 1977 JAMA costs and revenues were as follows (Stip. ¶¶ 11 and 13): Costs Advertising content Readership content Total $3,850,035 $4,525,873 $8,375,908 Revenues Subscriptions Miscellaneous Advertising Total $916,428 $12,680 $5,956,084 $6,885,192 6. In selling advertising in its periodicals, AMA competes with publishers of many other periodicals distributed to physicians (Hannon Aff. If 4). Pharmaceutical manufacturers have developed sophisticated methods of measuring the effectiveness of advertising in individual periodicals, using a number of widely accepted audience measurement reports (id. 115; Stip. 111123-28). Those reports permit manufacturers to determine the types of physicians most likely to prescribe their products and the periodicals that will reach the largest number of those high-prescribing groups (id.). 7. Most periodicals competing with JAMA and AM News for advertisements, and virtually all medical periodicals published by commercial publishers, distribute their periodicals free of charge to physicians falling within the categories most likely to use or prescribe the products sold by prospective advertisers (id. 117; Stip. 1110). Such free distribution to attract advertising is called “controlled circulation” (id.; Stip. 1110). 8. In each of the years at issue AMA engaged in controlled circulation by distributing JAMA (27,000 to 43,000 copies of each issue) and AM News (56,000 to 65,000 copies of each issue) free of charge to nonmember physicians in certain specialties (Hannon Aff. 116; Stip. H1Í10, 11 and 15). AMA also used controlled circulation for Prism until it was discontinued (Hannon Aff. II6). In choosing those “control group” physicians AMA relied in large part on data published by the National Disease and Therapeutic Index and the knowledge that AMA competitors were distributing their periodicals free of charge to those physicians (id.). AMA’s purpose in engaging in controlled circulation was to provide coverage to advertisers comparable to that offered by competing medical periodicals and thus to increase advertising revenues of the periodicals (id.). 9. Some AMA members belonged to the control group and thus would have received JAMA, AM News and Prism free of charge even had they not belonged to AMA (id.) Indeed, when those members later resigned from the AMA, they continued to receive the AMA periodicals free of charge (Stip. 111113 and 15). II. Relevant Tax Laws and Regulations 10. AMA has been recognized by IRS as a professional association under Code § 501(c)(6) and is thus exempt from federal income tax under Code § 501(a). Despite that tax-exempt status, AMA must pay taxes on its “unrelated business taxable income” (Code § 511), defined as (Code § 512(a)(1), emphasis added): the gross income derived by any organization from any unrelated trade or business (as defined in section 513) regularly carried on by it less the deductions allowed, by this chapter which are directly connected with the carrying on of such trade or business____ Reg. § 1.513-l(b) (emphasis added) explains the purpose of that tax: The primary objective of adoption of the unrelated business income tax was to eliminate a source of unfair competition by placing the unrelated business activities of certain exempt organizations upon the same tax basis as the nonexempt business endeavors with which they compete. See also United States v. American College of Physicians, 475 U.S. 834, 106 S.Ct. 1591, 1594, 89 L.Ed.2d 841 (1986) (Congress sought to strike “a balance between its two objectives of encouraging benevolent enterprise and restraining unfair competition....”). 11. It is undisputed for purposes of this case that the income AMA earned from 1975 through 1978 from advertisements in its periodicals was “unrelated business taxable income” and therefore taxable (AMA Mem. 4). AMA calculated that taxable income each year based on its interpretation of the Treasury regulations. It reported taxable income for JAMA and AM News for the entire four-year period in an amount equal to 12.7% of the total advertising revenues received for those two periodicals (Stip. UU 5, 34; Stip. Exs. 6-A-D; AMA Mem. Ex. A). 12. Later IRS redetermined AMA’s unrelated taxable business income, finding it exceeded 37% of advertising revenues, and concluded AMA owed additional taxes on that income (Stip. UU 6, 35; Stip. Ex. 7; AMA Mem. Ex. A). Although IRS did not disagree with the total advertising revenues and total expenses reported by AMA for its periodicals, it disagreed with AMA’s interpretation of Treasury regulations governing the deductibility of periodical costs. 13. According to the regulations, periodical costs fall into two categories: (a) “Direct advertising costs” are costs “directly connected with the sale and publication of advertising,” such as the cost of producing and distributing the advertising content of periodicals (Reg. (f)(6)(ii)). Direct advertising costs are fully deductible from gross advertising income (Reg. (f)(2)(i)). (b) “Readership costs” are costs “directly connected with the production and distribution of the readership content, of the periodical” — all costs “not allocated to direct advertising costs” (Reg. (f)(6)(iii)). Readership costs are deductible from gross advertising income only to the extent they exceed “circulation income” (Reg. (f)(2)(ii)(b)). Circulation income is defined in Reg. (f)(3)(iii) (emphasis added) as: the income attributable to the production, distribution or circulation of a periodical (other than gross advertising income) including all amounts realized from or attributable to the sale or distribution of the readership content of the periodical, such as amounts realized from charges made for reprinting or republishing articíes and special items in the periodical and amounts realized from sales of back issues. Where the right to receive an exempt organization periodical is associated with membership or similar status in such organization for which dues, fees or other charges are received (hereinafter referred to as “membership receipts”), circulation income includes the portion of such membership receipts allocable to the periodical (hereinafter referred to as “allocable membership receipts”). 14. Allocable membership receipts, in other words, represent members’ “payments” by their dues, rather than a separate subscription charge, for periodicals they receive. Thus Reg. (f)(3)(iii) explains: Allocable membership receipts is the amount which would have been charged and paid if— (a) The periodical was that of a taxable organization, (b) The periodical was published for profit, and (c) The member was an unrelated party dealing with the taxable organization at arm’s length. That provision in turn refers to Reg. (f)(4) “for a discussion of the factors to be considered in determining allocable membership receipts____” Reg. (f)(4) actually sets out three methods for determining the amount of membership receipts to be allocated to the circulation income of a particular periodical. Only two of those are relevant here: (a) Method 1 (Reg. (f)(4)(i)) applies to most of the AMA specialty journals: Subscription price charged to nonmembers. If 20 percent or more of the total circulation of a periodical consist of sales to nonmembers, the subscription price charged to such nonmembers shall determine the price of the periodical for purposes of allocating membership receipts to the periodical. (b) Method 3 (Reg. (f)(4)(iii)) applies to JAMA, AMA News and Archives of Surgery: Pro rata allocation of membership receipts. Since it may generally be assumed that membership receipts and gross advertising income are equally available for all the exempt activities (including the periodical) of the organization, the share of membership receipts allocated to the periodical, where paragraphs f(4)(i) and (ii) of this section do not apply, shall be an amount equal to the organization’s membership receipts multiplied by a fraction the numerator of which is the total periodical costs and the denominator of which is such costs plus the cost of other exempt activities of the organization. 15. In September 1971, when the proposed amendments to Reg. § 1.512(a)-l were announced in a “Notice of Proposed Rule Making” (36 Fed.Reg. 18,316), Reg. (f)(4) contained seven factors to be used in determining the allocation of membership receipts to a periodical’s circulation income (Stip. if 30; Ex. 4). But when the December 18,1975 notice announcing the adoption of final amendments to Reg. § 1.512(a)-l was published (40 Fed.Reg. 58,637), Reg. (f)(4) contained only the current three methods for determining allocable membership receipts (Stip. II31; Stip. Ex. 5). Conclusions of Law Jurisdiction over this action exists under 28 U.S.C. § 1346(a)(1). There has been no dispute on that score, and subject matter jurisdiction is clear in any event. Before this opinion turns to the Conclusions themselves, it is useful to identify AMA’s lines of attack in this litigation. AMA contends IRS committed six errors in calculating deductions from taxable advertising income to which AMA is entitled under Reg. § 1.512(a)-l. One relates to the deduction of direct advertising costs and the others relate to the deduction of readership costs: 1. Direct advertising costs: AMA argues IRS’ calculation of taxable advertising income from JAMA, AM News and Prism should have treated as fully deductible direct advertising costs (rather than partially deductible readership costs) the costs of producing and distributing the “readership content” of periodicals distributed free of charge to nonmember physicians in the control group to increase advertising revenues. 2. Readership costs: AMA argues IRS committed five errors in calculating membership receipts allocable to the circulation income of each periodical, resulting in too high an allocation of membership receipts to each periodical’s circulation income and consequently in too small a deduction of readership costs for each periodical: (a) In calculating membership receipts allocable to the circulation income of certain periodicals under method 3 in Reg. (f)(4)(iii), IRS purportedly committed three errors: 1. As to JAMA, AM News and Archives of Surgery, it should not have included among total membership receipts the AMA dues placed in the association equity account rather than used to pay activity costs. 2. As to JAMA and AM News, it should not have included among total membership receipts the dues collected from AMA members in the control group who would have received those periodicals free of charge from AMA even had they not been dues-paying members. 3. As to JAMA and AM News, it should have included as part of the “cost of other exempt activities” (in the denominator of the n. 14 formula) the total costs of all other AMA periodicals, not just the readership costs of those periodicals. (b) In calculating membership receipts allocable to each specialty journal’s circulation income under method 1 in Reg. (f)(4)(i), IRS assertedly committed two errors: 1. It should have determined the yearly “subscription price charged to such nonmembers” was $14 (one-half the two-year subscription rate) rather than $18 (the one-year subscription rate). .2. It should have used only a percentage of the “subscription price charged to such nonmembers” in determining the price of the periodical for AMA members paying reduced dues. Finally, as stated at the outset of this opinion, AMA has an alternative or fallback position: If IRS correctly construed the regulations, then those regulations are invalid because they conflict with the Code. These Conclusions will deal with each AMA contention in turn. Direct Advertising Costs This opinion has found AMA distributed copies of JAMA, AM News and Prism free of charge to nonmember physicians in the control group for the purpose of increasing advertising revenues. AMA Mem. 27-30 urges IRS therefore erred when it refused to permit AMA to deduct as direct advertising costs the costs of producing and distributing the readership content in those copies. G. Mem. 21-22 responds by attacking the predicate for that argument: It asserts AMA’s tax-exempt goal is education, so AMA cannot now claim it sends periodicals to doctors solely to generate advertising revenue. But G. Mem. 22 reaches that conclusion by misstating the evidence: If, in fact, the sole purpose of the AMA in distributing JAMA and the AM News is simply to obtain revenues from the advertisements for drugs, all of its circulation income should be taxable and the exemption of the AMA itself might come into question. Of course AMA does not contend it distributed all copies of those periodicals solely for the purpose of increasing advertising revenue. Nor does it seek to deduct as a direct advertising cost the readership content cost of all copies of the periodicals. Instead AMA contends (and has proved) it sent periodicals to certain nonmember physicians solely to generate advertising revenue. As to those copies of the periodicals, all the advertising revenue from which AMA must treat as taxable income, it seeks to deduct all (even the readership content) costs. Nothing in the Code or regulations bars AMA, just because of its tax-exempt status, from incurring such costs to generate advertising income (Rensselaer Polytechnic Institute v. Commissioner of Internal Revenue, 732 F.2d 1058, 1062 (2d Cir.1984)). Indeed the tax laws and regulations expressly provide for the taxation of non-exempt income as well as the deduction of certain costs from that income. And as Rensselaer explained (id.): Moreover, should the [tax-exempt organization] so pervert its operations that the institution no longer “engages primarily in activities which accomplish * * (its exempt purposes),” Treas.Reg. § 1.501(c)(3)-l(c)(l), the commissioner has adequate remedies available to correct any abuse or even terminate the exemption. While it may well be those nonmember physicians read and learned from the readership content in the free periodicals, that does not alter AMA’s income-generating motivation for sending them. That, however, is not the end of the analysis, though it seems the limit of the government’s line of attack. Though this Court has found AMA produced and distributed the control-group periodicals solely to generate advertising revenue, the question remains whether the cost of producing and distributing the readership content of those periodicals is deductible as a direct advertising cost. To resolve that question, this opinion looks first to the regulations. Reg. (f)(6)(ii) defines direct (fully-deductible) advertising costs as costs “directly connected with the sale and publication of advertising.” That definition would seem to cover all costs of producing and distributing periodicals (even costs attributable to readership content in those periodicals) whose sole aim was to increase advertising revenue. Nonetheless the same regulation also provides (id.): The items allowable as deductions under this [direct advertising cost] subdivision do not include any items of deduction attributable to the production or distribution of the readership content of the periodical. Linked to that provision, Reg. (f)(6)(iii) defines readership costs as items “directly connected with the production and distribution of the readership content of the periodical.” As construed (implicitly) by the government’s memorandum, those regulations create a per se rule: Costs attributable to, or directly connected with, the production and distribution of the readership content (as opposed to the advertising content) of a periodical cannot be “directly connected with the sale and publication of advertising” and hence are never deductible as direct advertising costs. Yet the government is understandably silent as to any justification for such a per se reading. After all, the whole purpose of enacting the unrelated-business-income provisions of the Code was to eliminate unfair competition between tax-exempt and taxable organizations by imposing the same tax on income of the former unrelated to a tax-exempt purpose. Just like its commercial competitors, AMA (1) distributes “control group” periodicals solely to generate advertising income, (2) gets no circulation income from that distribution and (3) pays taxes on the advertising revenue generated from the distribution. Yet IRS permits commercial publishers but not AMA to deduct the readership content costs of such periodicals. Far more reasonable, and wholly consistent with the Code’s purpose, is the construction of the regulations proposed by AMA Mem. 29 (emphasis in original): An item of deduction is “attributable to” [or “directly connected with”] the production or distribution of the readership content only if dissemination of readership content for its own sake is a primary purpose of the item. Where costs are incurred solely for the purpose of increasing advertising revenues, these costs are not “attributable to” [or “directly connected with”] the readership content, even if they necessarily include some costs of physically producing and distributing readership material. That construction comports with the Code’s objective of equal tax treatment: It permits AMA, like its commercial competitors, to deduct readership content costs when they are incurred solely to generate advertising revenue rather than to further AMA’s exempt purposes. Accordingly this Court concludes AMA should have been permitted to deduct the readership-content costs of periodicals distributed to control-group physicians to generate advertising revenue. Readership Costs: Allocable Membership Receipts Under Method 3 Readership costs of a periodical are deductible against unrelated taxable income only to the extent they exceed that periodical’s circulation income. For that purpose the circulation income of an AMA periodical distributed to members because they pay dues includes whatever portion of the total dues (allocable membership receipts) represents payment for the periodical. Because it is not possible (according to the regulations) to calculate the subscription price of JAMA, AM News and Archives of Surgery under methods 1 and 2 (Reg. (f)(4)(i) and (f)(4)(ii)), allocable membership receipts are determined under method 3 (Reg. (f)(4)(iii)) under the formula found at n. 14. That formula lumps the total periodical costs and the cost of all the organization’s other exempt activities into the denominator, on the assumption that membership receipts and gross advertising income are “equally available for all the AMA’s exempt activities (including the periodical) of the organization ...” (id.). Thus method 3 assigns to any specific periodical a portion of total membership receipts used for all exempt activities in proportion to its costs. That forms the background for AMA’s first challenge on readership-cost grounds. All its membership receipts collected in each of the years 1975 through 1978 were not used for exempt activities when collected. Instead some of those dues were placed in an “association equity” account — a reserve fund to be used if activity costs ever exceeded revenues. For example, in 1977 AMA collected $37,469,346 in dues but placed $14,828,618 in the association equity account, applying $22,640,728 to pay AMA activity costs that year. That means a portion of just that $22,640,728 (and not of the entire $37,469,346) in collected dues was available for each AMA periodical. AMA Mem. 34-36 is therefore right in urging the regulations should be construed to exclude from each year’s total membership receipts (for formula purposes) any dues collected that year but placed in the association equity account. Because such dues were not actually used for any activity costs (including periodical costs) in the year collected, there is no basis for concluding a portion of those dues were used for the costs of a particular periodical that year. Cf. TAM No. 8305002 at 3-4 (provided in AMA R. Mem. Ex. 2). Instead such dues should be included, for formula purposes, as total membership receipts in the year they are taken from the association equity fund and actually used for activity costs — as AMA did in 1985 (Sammons Aff. 116). That way the “equally-available” assumption underlying the method 3 formula holds true, and the formula does not allocate a greater percentage of membership receipts than of advertising revenue to activity costs. G. Mem. 22-24 nonetheless insists all dues must be included as membership receipts in the year collected for three reasons: 1. Reg. (f)(3)(iii) defines membership receipts as “dues, fees or other charges,” covering all dues regardless of whether they are used for activity costs. 2. Because AMA members were able to deduct the dues paid to AMA, AMA was not taxed unfairly as compared to its commercial competitors who did not have to treat capital contributions as receipts. 3. Widespread tax avoidance or endless disputes with IRS auditors would result were AMA permitted to allocate funds to a capital account. None of those arguments has merit. First, in light of the regulation’s purpose, the only way to avert a successful attack on its validity is to construe “membership receipts” in a given year — for formula purposes — to mean dues, fees or other charges used for activity costs that year. Second, the ability of AMA members to deduct their dues each year is simply and obviously irrelevant to the question of AMA’s entitlement to certain deductions from taxable advertising income. Finally this Court rejects the government’s bald assertion that widespread tax avoidance or endless disputes with IRS would result from the proposed construction of the regulation. No reason is suggested (1) why AMA and like organizations could not be required to report, and IRS could not determine from AMA’s financial records, whether and when dues initially placed in the association equity account were later used for activity costs and (2) why those dues could not be reported as membership receipts at that later time. AMA’s second attack in the readership cost area stems from the fact that many AMA members were within the physician group most likely to prescribe the products sold by prospective advertisers, so those members would have received AMA periodicals free of charge even had they not been dues-paying AMA members. AMA Mem. 37-38 thus argues the dues of those members should have been excluded from total membership receipts in the method 3 formula, reasoning: 1. Reg. (f)(3)(iii) says circulation income should approximate the amount that would be “charged and paid” by members if “the periodical was that of a taxable organization ... published for profit.” 2. Any commercial publisher of JAMA and AM News, dealing at arm’s length with the AMA control-group members, would not charge them for the periodicals. 3. Equal tax treatment would therefore be attained only if the dues of those AMA members are excluded from total membership receipts. What that analysis ignores, however, is that the control-group AMA members had to and did pay dues to be members. And those dues (like the dues of all other AMA members) were available to defray periodical costs. In other words, AMA (unlike its commercial counterparts) was able to require those physicians who chose membership to subscribe to AMA’s periodicals, and it thus derived subscription income from such distribution. That being so, it is really irrelevant that a commercial publisher would not have charged the control-group members for periodicals. To the extent AMA in fact derived income from that distribution (see n. 22), that income should be considered in determining the AMA periodicals’ circulation income. This opinion therefore rejects AMA’s contention. Instead this Court construes Reg. (f)(4)(iii) as properly calling for the inclusion of the dues of AMA control-group members in AMA's total membership receipts. AMA Mem. 38 asserts IRS committed one other error in calculating allocable membership receipts under the method 3 formula. This time AMA disputes IRS’ computation of the “cost of other exempt activities” in the denominator of the formula (term “c” in the algebraic version expressed in n. 14). As a preliminary matter, both sides agree that where an organization (1) publishes only one periodical or (2) publishes more than one periodical but treats their income on a consolidated basis (as permitted under Reg. (f)(7)(i)), the total costs of all periodicals are included in both the numerator and denominator of the method 3 formula as “total periodical costs” (term “b” in the formula). That being so, AMA Mem. 38-40 reasons where an organization such as AMA publishes more than one periodical but does not consolidate their income, the total costs for all periodicals should likewise be included in both the numerator and denominator of the formula by reading the regulation this way: 1. “Total periodical costs,” as to any individual periodical for which allocable membership receipts are being computed, are the full costs of that periodical. 2. “Cost of other exempt activities,” as to the same periodical, is the cost of all other exempt activities including the total costs of all other AMA periodicals. IRS took a very different approach to calculating the “cost of other exempt activities” as to any individual periodical: It included only the readership costs of other AMA periodicals. As AMA Mem. 40 points out, that calculation (by decreasing the denominator in each instance) caused more membership receipts to be allocated to circulation income than if exactly the same periodicals had been published by separate organizations or reported on a consolidated basis. That meant a higher circulation income was attributed to the AMA periodicals, in turn leading to a lower readership costs deduction for those periodicals. G. Mem. 23-24 provides no basis whatever for the IRS’s skewed reading of the regulation. Instead it argues AMA was not prejudiced by that reading because AMA could have elected to consolidate its publications, thus avoiding any inequity it perceived from treating each publication separately. Even if true, that assertion provides no support at all for IRS’s distorted construction of the regulation. This Court concludes the “cost of other exempt activities” in the denominator of the method 3 formula includes all costs (not just readership costs) of other AMA periodicals. In addition to rendering uniform the regulation’s application as between organizations that do and do not consolidate, that construction is consistent with the purpose of the regulation: allocation of a portion of total membership receipts to a single periodical in proportion to its costs. After all, membership receipts are available for all exempt activity costs. It makes absolutely no sense to exclude the costs other than readership costs of AMA’s other periodicals in making the allocation, especially when all those costs of the individual periodical are included in the formula. Readership Costs: Allocable Membership Receipts Under Method 1 AMA also complains of two aspects of the IRS treatment of its specialty journals. As already indicated, the membership receipts allocated to the circulation income of those journals are properly calculated by the use of method 1 (Reg. (f)(4)(i)). Under method 1 the subscription price charged to nonmembers for a periodical determines the periodical’s price for purposes of allocating membership receipts to that periodical’s circulation income. When AMA applied method 1 to determine allocable membership receipts for its four specialty journals, it used one-half the two-year subscription rate (or $14) and an equivalent percentage of the dues of members receiving the periodicals but paying reduced dues. When IRS redetermined AMA’s taxes, it used the one-year subscription rate ($18) for all members receiving the periodicals. Neither side offers any reason for choosing one subscription rate over the other. AMA Mem. 63 says simply: The IRS construction ... is wholly unrealistic because it ignores the fact that more than one “subscription price” was charged to nonmembers. And G. Mem. 25 ignores the issues completely. As for the first question, this Court concludes IRS correctly used the one-year subscription rate in determining the price of the specialty periodicals to AMA members who pay full dues each year. After all, the reduced two-year subscription rate is given only to those nonmembers who commit themselves to purchasing an AMA periodical for two consecutive years. To the extent an AMA member makes only a one-year commitment to belong to AMA by paying yearly dues, he or she also makes an identical one-year commitment to purchase AMA periodicals. Each year that same AMA member is thus able to decide whether or not to renew his or her membership, and consequently whether or not to continue to receive the periodicals, by paying his or her dues. There is no guaranty the AMA member will “subscribe” to AMA periodicals for longer than that single year. Accordingly there is no basis for finding the reduced two-year subscription rate should have been considered the rate charged to such AMA members. Only one question remains: the proper treatment of periodicals sent to AMA members whose dues were far below the $250 annual level paid by regular members in 1976 through 1978: interns and residents (who paid only $35 in dues) and medical students (who paid just $15). AMA Mem. 63-64 says it was “illogical” for IRS to have applied method 1 by treating $18 as the price of periodicals sent to those members: Indeed, a construction of [Reg. (f)(4)(i) ] resulting in allocation from individual members of more dues that [sic] they paid would be so irrational as to render that subdivision as so applied invalid. 6. Mem. 25 responds: However, the [IRS] followed the regulations even though it did not necessarily allocate the precise amount of the medical student and other members who paid less dues than regular members. This seems reasonable in that medical students in any event would probably be more interested in reading these publications for their scholarly content than some of the other members. The dues rate of $15 per year is still less than what it would cost a medical student to subscribe to these periodicals. As before, the government has forgotten the entire purpose of the regulation: allocation of a portion of members’ dues to the circulation income of periodicals received by those members, representing payment for the periodicals. Because AMA receives only $15 in total fees from each medical-student member, it would be totally unreasonable to allocate $18 to the circulation income of a periodical sent to that member. And though the numbers for interns and residents are not so graphically absurd, the principle is identical. In each instance the amount received from the low income (or no-income) member of the profession — a kind of “loss leader” — must be viewed as being in payment for all AMA exempt activities, and it must be allocated accordingly. Thus the proper application of method 1 to those members calls for 1. determining what percentage of the annual dues of a full-dues-paying member is allocable to a periodical’s circulation income (based on the one-year subscription rate charged to nonmembers) and then 2. allocating that same percentage of dues from members paying reduced dues to that periodical’s circulation income. ****** In summary, for each of the years 1975 through 1978: 1. AMA was entitled to deduct, as direct advertising costs, all its costs of producing and distributing the readership content of all its periodicals sent free of charge to nonmember physicians in the control group. 2. As for calculations regarding JAMA, AM News and Archives of Surgery under regulation method 3 (Reg. (f)(4)(iii)): (a) AMA’s “total membership receipts” during a year included the dues actually used for AMA’s activity costs, but not the dues placed in an association equity account as a reserve fund to cover any future deficit in meeting activity costs. (b) AMA could not exclude from its “total membership receipts” the dues paid by its members who would have received its periodicals free of charge even had they not been dues-paying AMA members. (c) AMA’s “cost of other exempt activities,” as to any individual periodical, included the total cost of all other AMA periodicals — not merely their readership costs. 3. As for calculations regarding each of AMA’s specialty journals under regulation method 1 (Reg. (f)(4)(i)): (a) AMA’s circulation income from each of its regular-dues-paying members is equal to the one-year subscription price charged to AMA nonmembers (except where the member has paid membership dues for more than one year in advance). (b) AMA’s circulation income from each of its discount-dues-paying members is the amount calculated under the preceding subparagraph, multiplied by the ratio between that individual’s dues and those paid by a regular-dues-paying member. This action is set for a status hearing June 29, 1987 at 9 a.m., when counsel are directed to report on the future conduct of the litigation. SUPPLEMENT TO MEMORANDUM OPINION AND ORDER This Court’s June 22, 1987 memorandum opinion and order (the “Opinion”) set out Findings and Conclusions as to AMA’s claim for a refund of federal income taxes imposed on AMA income earned in the years 1975 through 1978 from the sale of advertising in AMA periodicals. AMA had contended that IRS denied deductions for AMA’s costs of producing and distributing its periodicals that, because those deductions would have been allowed a commercial publisher, are also permitted AMA under Code §§ 511-513. Because IRS construed the Code-implementing Treasury regulations to preclude those deductions, AMA had urged alternatively: 1. IRS construed the regulations erroneously and thus committed six errors in calculating AMA’s deductions. 2. If IRS construed the regulations correctly, they are inconsistent with the Code and therefore invalid. In the Opinion this Court concluded IRS had indeed misconstrued the regulations in most of the ways alleged by AMA, committing four of the six claimed calculation errors. That being so, the Opinion did not reach AMA’s alternative ground for relief: the asserted invalidity of the regulations. Now AMA requests this Court to resolve that issue, asserting: 1. All of Reg. (f) (“Determination of unrelated business taxable income derived from sale of advertising in exempt organization periodicals”) is invalid because it is inconsistent with Code § 512(a)(1). 2. One part of that same regulation, Reg. (f)(4) (“Allocable membership receipts”), is also invalid because it was promulgated without the notice required under the Administrative Procedure Act (“APA”), 5 U.S.C. § 553 (“APA § 553”). According to AMA, such invalidity entitles it, under the remaining regulations, to the full tax refund it has requested. This Court has again reviewed the material previously tendered by the parties, as well as each side's supplemental memorandum addressing the current issues. Because no further factual Findings are needed, this opinion reflects the additional Conclusions required as to challenged regulations. Conclusions of Law I. Reg. (j): Inconsistency with Code 512(a)(1) By way of brief background, Reg. (f) provides rules for the deduction of two categories of periodical costs from taxable advertising income: 1. “Direct advertising costs” are costs “directly connected with the sale and publication of advertising” and are fully deductible. 2. “Readership costs” are costs “directly connected with the production and distribution of the readership content of the periodical” and are deductible only to the extent they exceed “circulation income.” As construed in the Opinion, classification of periodical costs as either “direct advertising costs” or “readership costs” depends on AMA’s motivation in incurring the cost — not on the effect of its having done so. Thus Opinion at 1096 concluded AMA could fully deduct as direct advertising costs even the readership-content costs of periodicals distributed to nonmember control-group physicians, reasoning (id. at 1094 and 1095, footnote omitted): While it may well be those nonmember physicians read and learned from the readership content in the free periodicals, that does not alter AMA’s income-generating motivation in sending them. * * # * * * Just like its commercial competitors, AMA (1) distributes “control group” periodicals solely to generate advertising income, (2) gets no circulation income from that distribution and (3) pays taxes on the advertising revenue generated from that distribution. As for readership-content costs incurred to further AMA’s exempt purposes (rather than to generate advertising revenue), Opinion at 1091-92 n. 11 noted the rationale under the regulations for classifying those as partially-deductible “readership costs”: Because AMA incurs those costs to further its exempt purpose, they are not “directly connected with" the taxable business activity. Therefore Reg. (f) requires that those costs be allocated first to any tax-exempt income derived from the exempt activity, and only second (to the extent readership costs exceed exempt income) to taxable advertising income. AMA Mem. 46 contends such an allocation of costs in Reg. (f) is inconsistent with Code § 512(a)(1) because all readership-content costs (except those incurred solely to generate advertising revenue) are allocated initially to circulation income rather than advertising income. AMA says that inconsistency is demonstrated by two things: 1. Even readership costs incurred to further AMA’s exempt purpose, rather than to generate advertising income, do in fact contribute to advertising income and are therefore “directly connected with” the unrelated business. 2. Commercial publishers are permitted to deduct all readership-content costs from taxable advertising income. But having said all that, AMA Mem. 48 then goes on to acknowledge the reason for Reg. (f)’s limitation on the deductibility of those readership costs: It is true that tax-exempt periodicals have sources of non-taxable income which are not available to nonexempt organizations, and that this income, as well as advertising income, may be attributable to the readership costs of a tax-exempt periodical. The obvious solution to this problem is to allocate the costs of readership content in a reasonable manner between the taxable and the non-taxable income. Thus AMA really admits there is a permissible reason for Reg. (f)'s limitation on the deductibility of readership costs: It concedes that because its circulation income is tax-exempt, it should not be permitted to deduct all readership costs (as a commercial publisher is permitted to do). What AMA actually challenges, then, is simply the Reg. (f) method of limiting the deductibility of those costs. But on that score. AMA fails to identify any persuasive reason why the Reg. (f) allocation is unreasonable or inconsistent with Code § 512(a)(1). All AMA says is that readership costs contribute to taxable advertising income as well as to tax-exempt circulation income. From there it leaps to the impermissible conclusion that the only reasonable way to allocate those costs is in the ratio of advertising income to circulation income. Even if it is assumed AMA’s proposed allocation method is reasonable — indeed even if AMA’s proposed method might arguably be perceived as more reasonable than the one provided under Reg. (f) — that would not be reason to invalidate the regulation. As National Muffler Dealers Association, Inc. v. United States, 440 U.S. 472, 488, 99 S.Ct. 1304, 1312, 59 L.Ed.2d 519 (1979) explains: The choice among reasonable interpretations is for the Commissioner, not the courts. Although readership costs may well contribute to taxable advertising income, AMA unquestionably incurs those costs to further AMA’s exempt purpose. That being so, Reg. (f) cannot be said to be unreasonable in treating such costs as not “directly connected with” an unrelated taxable business. Nor is Reg. (f) unreasonable because it requires that costs incurred to further an exempt purpose be allocated first to any tax-exempt circulation income and only then (to the extent those costs exceed circulation income) to taxable advertising income. After all, every dollar (and not just a portion) of AMA’s circulation income is tax-exempt and fully available to defray periodical costs. In sum, this Court finds Reg. (f) is a reasonable implementation of Code § 512(a)(1). It is therefore a valid regulation, surviving AMA's attack. II. Reg. (f)(4); Compliance with APA Notice Requirements AMA fares better in its challenge of Reg. (f)(4), which sets out three methods for determining the amount of membership receipts to be allocated to the circulation income of a particular periodical (“allocable membership receipts”): 1. the subscription price charged to nonmembers, if 20 percent or more of the total circulation of a periodical consists of sales to nonmembers (“Method 1”); or 2. the subscription price charged to members, if (a) Method 1 does not apply and (b) dues of 20 percent or more of the members are reduced because they do not receive a periodical (“Method 2”); or 3. a pro rata allocation of membership receipts (assigning to any specific periodical a portion of total membership receipts used for all exempt activities in proportion to the periodical’s costs), if neither Method 1 nor Method 2 applies (“Method 3”). As Opinion at 1092 explained, allocable membership receipts represent members’ “payments” through their dues, rather than by a separate subscription charge, for periodicals they receive. AMA urges Reg. (f)(4) is invalid because it was promulgated without the notice required under APA § 553. Absent that “invalid” Reg. (f)(4), AMA asserts the remaining regulations (which in this instance means Reg. (f)(3)(iii)) supports AMA’s refund claim. This opinion treats first with the notice claim, then turns to the effect of AMA’s success on that issue. A. Notice APA § 553 requires that notice of proposed rulemaking be published in the Federal Register 30 days before the adoption of a regulation, so that interested persons have the opportunity to submit their views or arguments about the proposed rule. Such notice must include (APA § 553(b)(3)): either the terms or substance of the proposed rule or a description of the subjects and issues involved. “Proposed” Reg. (f)(4) was announced September 11,1971 in a Notice of Proposed Rule Making (36 Fed.Reg. 18,316). About four years later the final version of Reg. (f)(4) was adopted. AMA invokes American Federation of Labor v. Donovan, 757 F.2d 330, 338 (D.C.Cir.1985) (citations and footnotes omitted) to contend Reg. (f)(4) was adopted without notice because that regulation bears little resemblance to Proposed Reg. (f)(4): It is, of course, elementary that a final rule need not be identical to the original proposed rule. “The whole rationale of notice and comment rests on the expectation that final rules will be somewhat different — and improved — from the rules originally proposed by the agency.” ... However, “[w]here the change between the proposed and final rule is important, the question for the court is whether the final rule is a ‘logical outgrowth’ of the rulemaking proceeding.” ... [I]f the final rule deviates too sharply from the proposal, the affected parties will be deprived of notice and an opportunity to respond to the proposal.” See also Chocolate Manufacturers Association v. Block, 755 F.2d 1098, 1105 (4th Cir.1985) and cases cited there. Comparison of Proposed Reg. (f)(4) with the ultimately adopted version reveals IRS grossly oversimplified matters when it summarized the changes in Reg. (f)(4) this way (40 Fed.Reg. at 58,638): The notice of proposed rulemaking set forth seven factors to be used in making this [membership receipts] allocation. The amendments adopted by this document simplify the allocation rule by reducing the factors to three: (1) subscription price charged to nonmembers; (2) subscription price charged to members and; (3) pro rata allocation of membership receipts based upon cost. For one thing, AMA Supp. Mem. 10 is quite correct in saying Proposed Reg. (f)(4) provided for a “flexible, case-by-case approach” to determining allocable membership receipts, identifying various factors to consider in making that determination. In contrast Reg. (f)(4) sets out three specific methods of actually calculating allocable membership receipts, each applicable to a specific factual matrix. For example, under Proposed Reg. (f)(4) the “subscription price charged to nonmembers” was simply to be “accorded greater weight than any other factor” where sales to nonmembers represent 20 percent or more of total circulation. Reg. (f)(4)’s Method 1 makes that factor the sole determinant under such circumstances, stating the price charged to nonmembers “shall determine the price” for purposes of calculating allocable membership receipts. To be sure, sales of JAMA and. AM News to nonmembers did not represent at least 20 percent of total circulation, thus rendering Method 1 inapplicable. Nor were the dues of 20 percent or more of AMA’s members reduced because they did not receive those periodicals (Method 2). Allocable membership receipts for those periodicals were therefore calculated according to Method 3’s pro rata formula. But had Proposed Reg. (f)(4) been adopted, the amount derived from that pro rata formula would not have been conclusive — instead, it would simply have provided “some indication of the amount of allocable membership receipts____” AMA R. Mem. 5 highlights a second and equally important deviation between the proposed and adopted versions of the regulation. Proposed Reg. (f)(4)’s factor 3 (“Subscription price of comparable periodicals of taxable organizations”), which was not included in the ultimate Reg. (f)(4), had expressly provided: The fact that a taxable organization issues a periodical which is comparable to an exempt organization periodical and makes a practice of distributing substantially all of its circulation at no charge is substantial evidence that none of the membership receipts of the exempt organization are allocable to its periodical. As construed in Opinion at 1097-98, Reg. (f)(4) does not take that fact into consideration at all: As long as AMA charged for a periodical (through dues) and thus may be considered as having derived income from distributing it, Reg. (f)(4) requires the consideration of such income in determining that periodical’s circulation income. Thus Reg. (f)(4) takes an entirely different approach to the determination of allocable membership receipts than did Proposed Reg. (f)(4). Because Reg. (f)(4) deviated so drastically from Proposed Reg. (f)(4), and because it was issued without any separate notice of those highly material proposed deviations, this Court finds Reg. (f)(4) was promulgated without the required notice under APA § 553 and is therefore invalid. B. Effect of Invalid Reg. (f)(4) on AMA’s Refund Claim That is not the end of the story (as AMA would have it), for this opinion must next determine how that invalidity of Reg. (f)(4) impacts on AMA’s refund claim. AMA Supp. Mem. 12 says that in the absence of Reg. (f)(4) the only test for determining allocable membership receipts is that provided in Reg. (f)(3)(iii), which reads (in part): Allocable membership receipts is the amount which would have been charged and paid if— (a) The periodical was that of a taxable organization, (b) The periodical was published for profit, and (c) The member was an unrelated party dealing with the taxable organization at arm’s length. AMA Supp. Mem. 11 n. 4 says that language teaches this: The plain object of the “hypothetical taxable publisher” test for determining “circulation income” in subparagraph (f)(3)(iii) is to disregard sources of income available to exempt publishers attributable to their exempt activities that are not also available to commercial publishers. From that premise AMA Mem. 57-58 reasons that allocable membership receipts are calculated for JAMA and AM News under the Reg. (f)(3)(iii) “test” by determining how much circulation income a commercial publisher would have received for distributing those periodicals. Then, applying that test, AMA Mem. 58 (emphasis added) concludes: As shown by the testimony of Prof. Wecker, a taxable organization publishing JAMA and AM News would have received total subscription income in an amount not much greater than what the AMA in fact received from nonmembers. Thus little or no membership receipts should be allocated to “circulation income.” This is precisely the result that would have been reached under [Proposed Reg.] (f)(4) before it was amended so drastically without notice. Because “circulation income” so determined is lower than that reported by the AMA in the returns upon which its claims for refund are based, the AMA’s claims with respect to JAMA and AM News must be allowed in full. But AMA is obviously wrong in its notion that Reg. (f)(3)(iii), read independently of Reg. (f)(4), provides a meaningful standard (much less an actual test) for determining allocable membership receipts. Reg. (f)(3)(iii) is simply not a self-contained regulation: Immediately after its language on which AMA relies (the single sentence quoted at the beginning of this section B of this opinion), the regulation says: See [Reg. (f)(4) ] for a discussion of the factors to be considered in determining allocable membership receipts of an exempt organization periodical under the standard described in the preceding sentence. In other words, Reg. (f)(3)(iii) simply provides a generalized standard and then goes on to instruct how that standard is to be reduced to practice — hence the regulation’s specific cross-reference to Reg. (f)(4). That being so, even the meaning of the “standard” in Reg. (f)(3)(iii) is not fairly ascertainable except by construing that regulation in conjunction with the implementation provisions in Reg. (f)(4). From the start AMA has failed to recognize that, preferring instead to ignore the specific Reg. (f)(3)(iii) cross-reference to Reg. (f)(4). By thus construing Reg. (f)(3)(iii) in isolation, AMA has managed to read that regulation in a way that renders it inconsistent with its own implementation provisions (and thus internally inconsistent). For example, AMA R. Mem. 5 said Reg. (f)(3) retained the “overall flexible approach” of the proposed regulations while Reg. (f)(4) did not. In the same vein, AMA Supp. Mem. 11 now says: The mechanical allocation “factor” of (f)(4)(iii) is not entirely consistent with the more flexible and case-sensitive standard set forth in (f)(3)(iii)____ That is plainly absurd, because Reg. (f)(3)(iii) itself requires implementation of its “standard” according to what AMA disparagingly labels the “mechanical allocation” methods in Reg. (f)(4). For precisely that reason, Opinion at 1098 rejected AMA’s interpretation of the standard in Reg. (f)(3)(iii) and instead construed that language consistently with Reg. (f)(4). That led the Opinion to reach a conclusion entirely different from that advanced by AMA as to the meaning of the standard in Reg. (f)(3)(iii). But now Reg. (f)(4) has been invalidated, leaving a gap in the regulations: Reg. (f)(3)(iii) provides a standard, but it refers to a now-nonexistent regulation for instructions as to how to implement the standard. This Court cannot perform such major judicial surgery at AMA’s request, by excising the reference to Reg. (f)(4) and reading Reg. (f)(3)(iii) as though it were intended to stand alone (as it clearly is not). And that in turn means this Court cannot fairly determine whether or not IRS was reasonable in calculating allocable membership receipts for JAMA and AM News. On the other hand, this Court also cannot wholly ignore the APA-violative irregularity in promulgation of Reg. (f)(4). Were that to be done, APA § 553 could be ignored at will, and the entire purpose of APA could be flouted. One solution avoids both those unwholesome results. This action is stayed until a regulation can be promulgated that fills the gap created by the invalidation of Reg. (f)(4). Only then will this Court be in a position to decide whether AMA is entitled to the remainder of its refund claim. In summary: 1. Reg. (f) is not inconsistent with Code § 512(a)(1) and is therefore valid (except as next noted). 2. Reg. (f)(4) was promulgated without notice in violation of APA § 553. It is therefore invalid. 3. This action is stayed until a regulation is promulgated that fills the gap created by the invalidation of Reg. (f)(4). At that point this Court will determine whether AMA is entitled to a tax refund because IRS improperly calculated allocable membership receipts for JAMA and AM News. This action is set for a status hearing September 14,1987 at 9 a.m., when counsel are directed to report on the anticipated future conduct of this litigation. APPENDIX PROPOSED RULE MAKING (4) Allocable membership receipts. In determining the allocable membership receipts of an exempt organization periodical all the facts and circumstances involved in each case shall be considered. However, in cases where the factors described in subdivision (i) or (ii) of this subparagraph are significant, or in cases where the factor described in subdivision (iii) of this subparagraph exists, such determination shall be based solely upon the application of all of such factors that do exist. In cases where the preceding sentence does not apply all factors, including (but not limited to) the factors contained in subdivisions (iv) through (vii) of this subparagraph, shall be considered in determining allocable membership receipts. In order to avoid distortions which may arise from consideration of the facts and circumstances for only the taxable year, due regard shall be given to the facts and circumstances, including the application of all relevant factors for all preceding years. The factors used in determining such proper allocation include but are not limited to the following: (i) Subscription price charged to nonmembers. The subscription price charged to nonmembers will be accorded greater weight than any other factor in determining allocable membership receipts where there are significant sales to nonmembers. Sales to nonmembers are significant if the circulation represented by such sales constitutes 20 percent or more of total circulation. Where sales to nonmembers constitute less than 20 percent of total circulation, this factor will not be given greater weight than any other factor and will be given progressively less weight in proportion to the degree that sales to nonmembers are less than 20 percent of total circulation. (ii) Subscription price to members. If membership receipts from a significant percentage of members of an exempt organization are less than those received from the other members because such significant percentage of members do not receive the periodical, such factor shall be given greater weight than the factor described in subdivision (iii) of this subparagraph but less weight than the factor described in subdivision (i) of this subparagraph in determining allocable membership receipts of the organization. Except as provided in the preceding sentence, the subscription price established by the organization for members will otherwise be given relatively little weight. For purposes of this subdivision, the percentage of members not receiving the periodical will be deemed significant only if it represents 20 percent or more of the total number of members. This factor will be given progressively greater weight as such percentage increases. On the other hand, this factor will be given relatively little weight, if any, if substantially all the members