BANKERS LIFE AND CASUALTY COMPANY, Plaintiff-Appellant, v. UNITED STATES of America, Defendant-Appellee.
Upon first glance, this lawsuit seems overly ambitious. Bankers Life and Casualty Company invites us to invalidate a seemingly straightforward and long-standing Treasury regulation. The insurance company's bold argument makes sense in light of one crucial fact: Bankers Life seeks a tax refund of $71 million plus interest from 1983. In effect, Bankers Life is betting on a long shot. It must know its chances of winning are slim, but the huge potential payoff apparently justifies the effort.
In a nutshell, Bankers Life attacks the validity of Treasury Regulation § 1.815-2(b)(3). The regulation had been on the books for 22 years-since the end of the Eisenhower Administration-at the time it was applied to Bankers Life. The regulation measured the value of a life insurance company's distribution of real property to its shareholders according to market value-rather than adjusted basis-for the purpose of determining whether the distribution triggered a tax on previously untaxed income. The district court, after upholding the validity of the regulation, granted summary judgment in favor of the government, and Bankers Life, on this appeal, asks us to reverse the judgment.
As one might expect in a case involving the interpretation of the tax code and its accompanying regulations, a certain amount of dry exposition is necessary to lay out the controversy. This story begins with the death of John D. MacArthur, the owner of Bankers Life and Casualty, in 1978. MacArthur's death invoked the terms of a trust agreement that transferred ownership of the insurance company to the John D. and Catherine T. MacArthur Foundation (a name well-known to public television and radio audiences). In 1983, because of legal limits on the foundation's ability to hold corporate stock, see I.R.C. § 4943(c)(6), the insurance company transferred some of its real estate holdings to the foundation as a distribution to the sole shareholder. Eventually, in 1984, the foundation divested itself of Bankers Life by selling the company to the ICH Corporation. This case focuses on the amount of tax owed by Bankers Life because of the 1983 distribution of real estate assets. The fair market value of these real estate holdings was a whopping $875 million. Their aggregate adjusted tax basis, however, was just under $210 million. The difference between the two valuations results in a $71 million tax differential.
The distribution invoked the arcane rules for taxation of life insurance company income. See United States v. Atlas Life Ins. Co., 381 U.S. 233, 235 n. 2, 85 S.Ct. 1379, 1381 n. 2, 14 L.Ed.2d 358 (1965) (describing the life insurance taxation system); Continental Bankers Life Ins. Co. v. Commissioner, 93 T.C. 52, 60, 1989 WL 75161 (1989) (same). For our purposes, matters are even more obscure because Congress changed these rules in 1984.1 See Tax Reform Act of 1984, Pub.L. No. 98-369, §§ 211-24, 98 Stat. 494. In any event, under these complex rules, life insurance companies paid taxes in three distinct “phases.” Under “Phase I” taxation, an insurance company paid a tax on the lesser of either its gain from investment or its gain from investment and underwriting. See I.R.C. § 802(b)(1). Under “Phase II” taxation, an insurer paid taxes on one-half of the amount by which the gain from underwriting exceeded the gain from investment. See I.R.C. § 802(b)(2). In effect, Phases I and II taxed the insurance company's income-except for the remaining one-half of the gains from underwriting. This remaining income received special deferred taxation-“Phase III” taxation-the general subject of this case. The law required the insurance company to keep a running tally of this Phase III income and to pay a tax on the money when the company issued certain distributions to its shareholders. See I.R.C. § 802(b)(3).
In order to keep track of the Phase III income and determine when it owed taxes on the money, the revenue code required insurance companies to maintain two sets of bookkeeping entries, or, in IRS parlance, “tax return memorandum accounts.” The first of these accounts-the “shareholders surplus account” (“SSA”)-recorded the amount of accumulated after-tax earnings from Phase I and II income. See I.R.C. § 815(b). The second account-the “policyholders surplus account” (“PSA”)-recorded the amount of accumulated Phase III income. See I.R.C. §§ 802(b)(3), 815(c).
The laws allowed an insurance company to distribute its Phase I and II after-tax earnings to its shareholders without consequence. Accordingly, whenever a life insurer issued a distribution, it deducted the amount of the distribution from the SSA without tax consequences. This made complete sense because the company already paid Phase I and II taxes on this money. If, however, the distribution exceeded the amount of money recorded in the SSA, the laws treated the excess amount of the distribution as a transfer of Phase III income and required the insurance company to pay Phase III taxes. Thus, the insurer deducted the excess from the PSA and paid a Phase III tax on this amount. See I.R.C. § 815.2
To sum up, a life insurance company could have distributed its after-tax earnings (the amount of money in the SSA) without paying any additional tax. But when the company dipped into the previously untaxed earnings (the amount of money in the PSA) to make a distribution, it paid the Phase III income tax. In other words, the laws allowed the life insurance company simply to avoid paying taxes on a big chunk of its income until it distributed that income to the shareholders.
Why did life insurance companies receive this unusual treatment? 3 The Senate Finance Committee discussed the matter when it enacted the Life Insurance Company Income Tax Act of 1959 which established the three-phase tax structure. See Pub.L. No. 86-69, 73 Stat. 112. According to the committee report, Phase III taxation addressed the “long-term nature of life insurance contracts” by allowing the company to keep one-half of the annual underwriting gain as “a desirable ‘cushion’ for special contingencies which may arise.” S.Rep. No. 86-291, at 26 (1959). The report explained that underwriting gains may appear as income under an annual tax system when, in fact, these gains merely serve to fulfill life insurance contracts down the road. See id.; see also H.R.Rep. No. 86-34, at 1, 15 (1959). The special deferred tax treatment adjusted for the conflict between a tax code based on annual income and the timetable of the insurance business in which income remains uncertain until actuarial predictions come to life (or death, as the case may be).
Congress reasoned that when a life insurance company made a distribution so large that it dipped into the PSA, then the company no longer needed a big reserve for a rainy day. As the Senate committee explained, “If the insurance company itself decides to distribute these amounts to stockholders, it has demonstrated that this cushion is no longer needed.” S. Rep., supra, at 26. Thus, a distribution larger than the SSA triggered the Phase III tax.
So far, we have explained one part of the mechanism the IRS used to calculate Phase III taxes. We still need more information to determine the amount of tax. The missing link is the method for valuing the distribution. To determine the value of a distribution for purposes of Phase III taxation, the IRS relied on a regulation-Treasury Regulation § 1.815-2(b)(3)-developed under its general authority to promulgate necessary rules. See I.R.C. § 7805(a). Before issuing this regulation, the IRS followed full notice and comment procedures. Treasury Regulation § 1.8152(b)(3) provided that certain noncash distributions would be measured by fair market value:
Except in the case of a distribution in cash and as otherwise provided herein, the amount to be charged to the special surplus accounts referred to in subparagraph (1) of this paragraph with respect to any distributions to shareholders (as defined in section 815(a) and paragraph (c) of this section) shall be the fair market value of the property distributed, determined as of the date of distribution. However, for the amount of the adjustment to earnings and profits reflecting such distributions, see section 312 and the regulations thereunder. For a special rule relating to the determination of the amount to be charged to such special surplus accounts in the case of a distribution by a foreign life insurance company carrying on a life insurance business within the United States, see section 819(c)(1) and the regulations thereunder.
If we put some real world numbers to all of this abstract tax talk we can see why Bankers Life has a bone to pick with the IRS. In 1983 Bankers Life had an SSA of $241,014,695 and a PSA of $154,434,319. The real estate assets it distributed to the foundation, as we noted earlier, had a fair market value of over $875 million but an aggregate adjusted basis of only $209,650,747. Pursuant to Treasury Regulation § 1.815-2(b)(3), Bankers Life valued the distribution at fair market value. As a result, the life insurance company depleted its SSA and emptied its PSA-paying Phase III taxes on the entire amount of the PSA. After computation of various adjustments, Bankers Life ended up paying a Phase III tax of $70,750,058.
Incidently, another abstruse twist in the tax laws prevented the IRS from just requiring Bankers Life to pay Phase I and II taxes on the distribution under the normal rule that when a corporation distributes appreciated property to shareholders, the corporation must recognize the gain. That obscure convolution in the tax code was the General Utilities doctrine (from General Utilities & Operating Co. v. Helvering, 296 U.S. 200, 56 S.Ct. 185, 80 L.Ed. 154 (1935))-“was” because Congress extinguished the doctrine with the 1986 tax reforms. See Tax Reform Act of 1986, Pub.L. No. 99-514, 100 Stat. 2085. The General Utilities doctrine provided that a corporation did not pay normal tax on a noncash distribution. Bankers Life made sure that this exception to the norm applied-it sought and received a private letter ruling that the 1983 distribution did not give rise to a taxable event for purposes of Phase I or Phase II taxation. See Private Letter 8540070 (July 10, 1985). (Bankers Life also received permission from Illinois regulators to make the distribution as an “extraordinary dividend” under state insurance law.) If Bankers Life had followed the current tax law regarding noncash distributions, it would have paid Phase I and II taxes on the appreciation, then included the after-tax gain in the SSA as a credit, and simultaneously entered a debit reflecting the full value of the distribution to the foundation.
In any event, Bankers Life predictably examined the enormous Phase III tax payment very closely. Not surprisingly, the company concluded that Treasury Regulation § 1.815-2(b)(3) conflicted with the tax code. Bankers Life argues that the tax code required the IRS to value distributions at aggregate adjusted basis-rather than fair market value-for purposes of calculating Phase III tax. According to Bankers Life, the IRS should have valued the distribution at just under $210 million, an amount lower than the balance of the SSA. The SSA would have absorbed the entire distribution and Bankers Life would not have deducted any amount from the PSA. Thus, under this approach, Bankers Life would have owed no Phase III tax at all.
Consequently, Bankers Life filed for a refund seeking a return of the Phase III tax plus interest since 1983. The IRS stood by its regulation and disallowed the claim. Bankers Life then filed this lawsuit. Because the parties did not dispute the facts, the district court (Judge Paul E. Plunkett) needed to determine only the validity of Treasury Regulation § 1.815-2(b)(3). On March 22, 1996, Judge Plunkett upheld the regulation and granted summary judgment in favor of the government. Bankers Life now appeals. We review de novo. See Bell Fed. Sav. & Loan Ass'n v. Commissioner, 40 F.3d 224, 226 (7th Cir.1994).
Before explaining Bankers Life's substantive arguments or the government's responses, we need to cover the basic structure of a challenge to a tax regulation. Essentially, we must determine the degree of deference owed to a Treasury Regulation issued under I.R.C. § 7805(a) with notice and comment procedures. This seemingly simple inquiry leads us into a free-fire zone of judicial debate over the proper level of judicial deference to various IRS interpretations of the revenue laws. The basic question is whether Chevron deference (from Chevron U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), and its progeny) applies to tax regulations. See, e.g., Central Pa. Sav. Ass'n v. Commissioner, 104 T.C. 384, 390-91, 1995 WL 138584 (1995) (discussing the “checkered career” of Chevron deference in the tax arena), supplemented by, 71 T.C.M. (CCH) 2724, 1996 WL 165489 (Apr. 10, 1996). In addition to the debate in the case law, the topic has generated heated scholarly commentary. See Ellen P. Aprill, Muffled Chevron: Judicial Review of Tax Regulations, 3 Fl. Tax Rev. 51 (1996) (describing and advocating a fusion of Chevron and National Muffler, 440 U.S. 472, 99 S.Ct. 1304, 59 L.Ed.2d 519 (1979)); Paul L. Caron, Tax Myopia Meets Tax Hyperopia: The Unproven Case of Increased Judicial Deference to Revenue Rulings, 57 Ohio St. L.J. 637 (1996) (critiquing attempts to apply discrete levels of deference to tax interpretations); John F. Coverdale, Court Review of Tax Regulations and Revenue Rulings in the Chevron Era, 34 Geo. Wash. L. Rev. 35 (1995) (arguing for distinct levels of non-Chevron deference for each type of tax interpretation); Linda Galler, Judicial Deference to Revenue Rulings: Reconciling Divergent Standards, 56 Ohio St. L.J. 1037 (1995) (attempting to provide a hierarchy of uniformly applied levels of deference for tax interpretations).
Unfortunately, we walk into this melee without very much help from the parties. Bankers Life inexplicably urges us to evaluate its challenge to Treasury Regulation § 1.815-2(b)(3) under what many commentators would consider the highly deferential Chevron standard. Frustratingly, the government neglects to address the applicability of Chevron deference. Therefore, before resolving this issue, we need some background information on the different forms of IRS interpretation and the corresponding degrees of deference.
The IRS interprets the tax code in four significantly different ways: (1) regulations issued pursuant to a specific directive from Congress, (2) regulations issued under the IRS's general authority to interpret the tax laws, (3) revenue rulings, and (4) private letter rulings. Each of these categories often receives different deference from the courts. While the first three categories constitute interpretations of general applicability, letter rulings apply only to the parties who specifically request them. Neither the courts nor the IRS may rely on letter rulings as precedent. See I.R.C. § 6110(j)(3) (1988); Treas. Reg. § 301.6110-7(b) (West 1998); cf. Fox Valley & Vicinity Const. Workers Pension Fund v. Brown, 897 F.2d 275, 280 n. 2 (7th Cir.1990) (en banc ). At the other end of the spectrum from letter rulings are regulations issued under specific grants from Congress. The tax code contains a myriad of specific congressional instructions regarding rulemaking. See Coverdale, supra, at 52 (estimating the number of specific grants at over 1000 and providing examples). Pursuant to the Administrative Procedures Act (“APA”), the IRS issues these rules with full notice and comment. See 5 U.S.C. § 553 (1994). In the middle, between letter rulings and specific authority regulations, are general authority regulations and revenue rulings. The IRS issues general authority regulations under its power to “prescribe all needful rules and regulations.” See I.R.C. § 7805(a) (1994). While the IRS takes the position that regulations issued solely under this general authority do not require notice and comment, the agency nevertheless usually follows full notice and comment procedures. See First Chicago NBD Corp. v. Commissioner, 135 F.3d 457 (7th Cir.1998). In contrast, the IRS does not follow notice and comment procedures for revenue rulings. Revenue rulings typically contain the IRS's interpretation of how the law applies to a set of hypothetical facts. Revenue rulings do not have broad application like regulations, but the IRS does consider them authoritative and binding. See Rev. Proc. 95-1, § 2.05 (codified at Treas. Reg. § 601.201 (West 1998)); Rev. Proc. 89-14, § 7.01 (codified at Treas. Reg. § 601.601 (West 1998)).
Revenue rulings receive the lowest degree of deference-at least in this circuit. In First Chicago, we held that revenue rulings deserve “some weight,” 135 F.3d at 459, and are “entitled to respectful consideration, but not to the deference that the Chevron doctrine requires in its domain,” id. at 458 (citations omitted). In other circuits this question has generated inconsistent rulings ranging from Chevron deference to no deference. See Coverdale, supra, at 81-84 (citing cases); Galler, supra, at 1059-74 (citing cases).
Determining the level of deference accorded to regulations is more difficult. Initially it may appear that we can resolve the problem by resorting to the APA's distinction between legislative and interpretive regulations. See 5 U.S.C. § 553(b)(3)(A) (1994). Administrative law scholars usually treat legislative regulations as rules of full legal effect-they create new legal duties binding on the parties and the courts and, therefore, require full notice and comment procedures. Interpretive rules, on the other hand, only clarify existing duties and do not bind; thus, they do not require notice and comment. In the tax world, however, these terms and classifications seem to provide more confusion than clarity. Tax experts refer to specific authority regulations as “legislative” and to general authority regulations as “interpretive.” The confusion arises because the “interpretive” designation does not mesh with the characteristics of the IRS's general authority regulations. While the IRS calls its general authority regulations interpretive, the agency promulgates them according to the same formal procedures it employs for its specific regulations. Moreover, both the specific authority and general authority regulations, create duties and have binding effect. See Aprill, supra, at 55-56; Coverdale, supra, 48-50.
If this legislative versus interpretive distinction held up for tax regulations, we could apply the simple rule described in Hanson v. Espy, 8 F.3d 469 (7th Cir.1993). In that case, we explained the distinction between nontax legislative and interpretive regulations and said that Chevron applied to legislative regulations only. See id. 472 n. 3. General tax regulations seem to carry the force of law, they are developed according to notice and comment, and they have the imprimatur of a congressional delegation of authority. In substance, general tax regulations fall short of being full legislative regulations only because the congressional delegation is general rather than specific. This distinction, however, may not have any effect at all on the standard of deference because Chevron itself dealt with a regulation promulgated under an arguably general grant of authority to the EPA under the Clean Air Act. See Coverdale, supra, at 50. Furthermore, Chevron stated that its framework applied to implicit congressional delegations as well as to specific and explicit directives. See Chevron, 467 U.S. at 844, 104 S.Ct. at 2782-83.4
In any event, courts uniformly give tremendous deference to regulations issued under a specific directive from Congress. In fact, some courts give these regulations the same force as statutes, focusing only on whether the agency promulgated the regulation properly within its scope of authority. See Coverdale, supra, at 53 (citing cases). We basically follow the same approach. In Gehl Co. v. Commissioner, 795 F.2d 1324 (7th Cir.1986), we explained that “a court's main focus in examining a legislative regulation issued under a specific statutory authority ‘is whether the interpretation or method is within the delegation of authority.’ ” Id. at 1329 (quoting Rowan Companies v. United States, 452 U.S. 247, 253, 101 S.Ct. 2288, 2292, 68 L.Ed.2d 814 (1981)). One commentator suggests that this level of deference accords even greater respect to the IRS's views than Chevron. See Coverdale, supra, at 55. We leave that hypothesis for another day.
Where then does all of this leave tax regulations promulgated under general authority? The question would be easier if the IRS did not employ notice and comment rulemaking procedures-in other words, if general authority tax regulations more truly resembled interpretive regulations. In Atchison, Topeka and Santa Fe Railway v. Pena, 44 F.3d 437 (7th Cir.1994) (en banc ), aff'd. without discussion, Brotherhood of Locomotive Engineers v. Atchison, Topeka and Santa Fe Railway, 516 U.S. 152, 116 S.Ct. 595, 133 L.Ed.2d 535 (1996), we held that courts should accord only moderate deference to interpretive regulations issued without notice and comment by an agency without general rulemaking powers.5 See id. at 442; see also Central Midwest Interstate Low-Level Radioactive Waste Comm'n v. Pena, 113 F.3d 1468, 1473 (7th Cir.1997) (“we do not apply Chevron's ‘rubber stamp’ to interpretive rules”). As the Atchison court explained:
Whatever degree of deference due these interpretive rules is dictated by the circumstances surrounding the agency's adoption of its statutory interpretation. “The weight given to an agency interpretation depends on many factors, including the validity of its reasoning, its consistency with earlier and later agency pronouncements and whether the administrative document was issued contemporaneously with the passage of the statute being interpreted.” Doe v. Reivitz, [830 F.2d 1441, 1447 (7th Cir.1987), amended, 842 F.2d 194 (7th Cir.1988) ]. In short, we look to “the thoroughness, validity, and consistency of the agency's reasoning.” Orrego v. 833 West Buena Joint Venture, 943 F.2d 730, 736 (7th Cir.1991) (citing United States v. Markgraf, 736 F.2d 1179, 1184 (7th Cir.1984)).
Atchison, 44 F.3d at 442.
In accord with its emphasis on the “circumstances surrounding the agency's adoption” of the regulation, the Atchison court understood Chevron either to apply or not to apply based on the extent to which an agency follows full notice and comment rulemaking procedures. The court explained that the notice and comment process provides a critical component in our decision to grant Chevron deference:
The principal rationale underlying [Chevron] deference is that in this context the agency acts as a congressional proxy; Congress develops the statutory framework and directs the agency to flesh out the operational details. But Congress typically does not permit the agency to run free in this endeavor; the Administrative Procedures Act establishes certain procedures that the agency must follow. Chief among them is the notice-and-comment provision of the APA. 5 U.S.C. § 553. This rulemaking process bears some resemblance to the legislative process and serves to temper the resultant rules such that they are likely to withstand vigorous scrutiny. It is this process that entitles the administrative rules to judicial deference.
Id. at 441-42.
In this regard Atchison is not unique. For instance, in Pennington v. Didrickson, 22 F.3d 1376 (7th Cir.1994), we explained “[w]e are mindful of our obligation to defer to the interpretation of the agency whenever that interpretation can be said to embody a deliberate and considered interpretation of the legislative intent.” Id. at 1383. When an agency undertakes notice and comment procedures it elevates the status of a regulation above mere interpretation. Similarly, in Doe v. Reivitz, 830 F.2d 1441, 1446-47 (7th Cir.1987), amended, 842 F.2d 194 (7th Cir.1988), we characterized regulations promulgated with the full panoply of procedures as “high powered” rules meriting Chevron deference. Thus, in this circuit it appears that rules developed pursuant to notice and comment procedures constitute full legislative regulations.
So, is there an answer to the question of whether we give Chevron deference to general authority tax regulations promulgated with notice and comment? At first it might appear that we shut the door on that question in Bell Federal Savings & Loan Ass'n v. Commissioner, 40 F.3d 224, 226 (7th Cir.1994). Unfortunately, that case does not provide definitive closure.
In Bell Federal we considered an IRS regulation promulgated under the general authorization statute, I.R.C. § 7805(a), just like the regulation in this case. We weighed whether to apply Chevron deference (as the Sixth Circuit did in Peoples Federal Savings & Loan Ass'n v. Commissioner, 948 F.2d 289, 304-05 (6th Cir.1991)) or whether to apply a more narrow traditional rule of deference to tax regulations (as the Ninth Circuit did in Pacific First Federal Savings Bank v. Commissioner, 961 F.2d 800, 805-08 (9th Cir.1992)). The Bell Federal court opted for the traditional rule apparently because it was more narrowly tailored to tax regulations:
The proper test for the validation of a challenged regulation thus requires only that the regulation issued by the Commissioner constitute a reasonable implementation of Congress's mandate. The Commissioner's interpretation need not be the only, or even the most, reasonable interpretation possible as “[t]he choice among reasonable interpretations is for the Commissioner, not the courts.”
40 F.3d at 227 (quoting National Muffler Dealers Ass'n v. United States, 440 U.S. 472, 488, 99 S.Ct. 1304, 1312, 59 L.Ed.2d 519 (1979)). Bell Federal continued by explaining that courts should inquire into whether the regulation “ ‘harmonizes with the plain language of the statute, its origin, and its purpose.’ ” Id. (quoting National Muffler, 440 U.S. at 477, 99 S.Ct. at 1307).
Bell Federal's choice of the traditional rule appeared eminently reasonable at the time. In Cottage Savings Ass'n v. Commissioner, 499 U.S. 554, 560-61, 111 S.Ct. 1503, 1507-08, 113 L.Ed.2d 589 (1991), the Supreme Court conspicuously did not cite Chevron and appeared to apply the traditional rule of tax deference: “Because Congress had delegated to the Commissioner the power to promulgate ‘all needful rules and regulations for the enforcement of [the Internal Revenue Code],’ 26 U.S.C. § 7805(a), we must defer to his regulatory interpretations of the Code so long as they are reasonable.” Id. at 560-61, 111 S.Ct. at 1508 (citing National Muffler).
The Bell Federal court equivocated, however, on whether there was any real difference between Chevron and the traditional rule. Chevron upholds a regulation if the agency based its interpretation on a permissible construction of a statute, see 467 U.S. at 843, 104 S.Ct. at 2781-82, while the traditional rule validates the regulation if the agency reasonably implemented Congress's mandate. Noting that “both approaches apply essentially the same test,” the Bell Federal court stated that “the difference between these two approaches is negligible at best-any regulation which is ‘based upon a permissible construction’ of an ambiguous statute will almost always ‘implement the congressional mandate in some reasonable manner’ and vice versa․” 40 F.3d at 227.
While Bell Federal's hedging about the difference between Chevron and the traditional rule gives us pause, we hesitate to apply the traditional rule for another reason. We think that the en banc decision in Atchison reopened the door that Bell Federal seemed to have shut. As Chief Judge Posner stated in First Chicago, “we consider [Chevron's] application to interpretive rules issued by the Treasury to be open in this circuit, since the Treasury has decided to use the notice and comment procedure for interpretive rules, though not required to do so by statute.” 135 F.3d at 459 (citation omitted).
Atchison explained that the notice and comment procedure was the sine qua non for Chevron deference.6 Because the IRS promulgates general authority regulations through a notice and comment process, Atchison militates forcefully in favor of according Chevron deference to those regulations. Thus, after Bell Federal, Atchison altered the balance and tipped the scales in favor of Chevron.7
Furthermore, we think Bell Federal's description of a “negligible” difference between the approaches might overstate the true extent of the divergence. Chevron held that “a court may not substitute its own construction of a statutory provision for a reasonable interpretation made by the administrator of an agency.” 467 U.S. at 844, 104 S.Ct. at 2782. While the two approaches articulate the level of deference differently, they both come down to one operative concept-reasonableness. Thus, Chevron and the traditional rule constitute two different formulations of a reasonableness test. There may be some subtle difference in the phrasing of each framework, but we should be wary of attempts to discern too many gradations of reasonableness. As we explained in Boyce v. Fernandes, 77 F.3d 946 (7th Cir.1996), “human ability to make fine distinctions is limited.” Id. at 948 (describing the difficulty in distinguishing between the inquiries into a police officer's probable cause for an arrest and whether that officer receives qualified immunity). Viewed from this perspective at least, the supposed gap between Chevron and the traditional rule is a distinction without a difference.
The Tax Court shares our suspicion that Chevron merely repackages the traditional rule: “[W]e are inclined to the view that the impact of the traditional, i.e., National Muffler standard, has not been changed by Chevron but has merely been restated in a practical two-part test with subtle distinctions as to the role of legislative history and the degree of deference to be accorded to a regulation.” Central Pa. Sav. Ass'n v. Commissioner, 104 T.C. 384, 1995 WL 138584 (1995), supplemented by 71 T.C.M. (CCH) 2724, 1996 WL 165489 (Apr. 10, 1996). Thus, as one might expect, other circuits and the Tax Court apply Chevron to general authority tax regulations. See, e.g., Western Nat'l Mut. Ins. Co. v. Commissioner, 65 F.3d 90, 93 (8th Cir.1995); Redlark v. Commissioner, 106 T.C. 31, 38-39, 1996 WL 10243 (1996).
This all leads us to the $64,000 question: If the deference test walks like Chevron and talks like Chevron, why shouldn't we just call it Chevron? Or, phrased differently, if the two standards are the same, does our denomination make any difference? Our answer lies with consistency. While we do not doubt that “foolish consistency is the hobgoblin of little minds, adored by little statesmen and philosophers and divines,” Ralph Waldo Emerson, Self Reliance, Essays: First Series (1841), we hope that our emphasis on consistency does not qualify as foolishness. On the contrary, consistency in the law forms the backbone of effective jurisprudence. See, e.g., Piper Aircraft Corp. v. Wag-Aero, Inc., 741 F.2d 925, 937 (7th Cir.1984) (Posner, J., concurring) (“The most important function of appellate review is to maintain the consistency of the law; if consistency is not a desideratum, the argument for appellate review is weakened.”). Because the APA imposes uniform procedures on agencies that formulate rules, there is some incentive for courts to apply a uniform framework for challenges to regulations developed with notice and comment. A consistent approach might help alleviate the tensions evident in the now-resolved conflict between the Sixth Circuit and the Tax Court over Chevron. See Peoples Fed. Sav. & Loan Ass'n v. Commissioner, 948 F.2d 289, 304-05 (6th Cir.1991) (“[W]e conclude that the Tax Court used the wrong standard to decide this case. Entirely ignoring Chevron ․ the Tax Court employed the standard in National Muffler Dealers which seemed to allow a plenary review of the legislative history without the deference requirements found in Chevron․”). More practically, our decision signals an effort to move toward a resolution of the current circuit split on the issue. Currently the Sixth Circuit accords Chevron deference to all tax regulations. See Peoples Federal. Meanwhile, the Third and Fifth Circuits have rejected Chevron for general authority tax regulations. See E.I. du Pont de Nemours & Co. v. Commissioner, 41 F.3d 130, 135-36 & n. 23 (3d Cir.1994); Nalle v. Commissioner, 997 F.2d 1134, 1138-39 (5th Cir.1993). Furthermore, we wish to avoid the situation in the Third Circuit-divergent treatment for notice and comment regulations depending on whether they are tax or nontax rules. See Elizabeth Blackwell Health Center for Women v. Knoll, 61 F.3d 170, 182 (3d Cir.1995) (applying Chevron to a nontax interpretive regulation), cert. denied, 516 U.S. 1093, 116 S.Ct. 816, 133 L.Ed.2d 760 (1996). We see no reason that nontax regulations issued with notice and comment should receive Chevron deference while tax regulations promulgated with the same procedures receive something different.
More importantly, in contrast to our foregoing analysis of the similarities between the traditional rule and Chevron, there are some important differences. These considerations also argue in favor of Chevron. Chevron is the focus of ongoing development by the Supreme Court, especially regarding the scope of the initial inquiry into statutory meaning. See, e.g., National R.R. Passenger Corp. v. Boston & Maine Corp., 503 U.S. 407, 417-18, 112 S.Ct. 1394, 1401-02, 118 L.Ed.2d 52 (1992) (reformulating the first step of Chevron analysis by focusing on the plain language of the statute rather than the intent of Congress). We do not labor under the illusion that Chevron will provide monolithic certainty for all courts reviewing administrative rules-we recognize the Court's own inconsistency regarding Chevron. See Thomas W. Merrill, Judicial Deference to Executive Precedent, 101 Yale L.J. 969 (1992) (explaining that “the Chevron framework is used in only about half the cases that the Court perceives as presenting a deference question”); see also Thomas W. Merrill, Textualism and the Future of the Chevron Doctrine, 72 Wash. U. L.Q. 351 (1994). We do hope, however, that by employing a uniform framework for challenges to administrative regulations we can better stay abreast of the Court's dictates.
We also favor Chevron because many courts contend that the traditional rule accords less than Chevron deference to tax regulations. The Sixth Circuit, for instance, characterizes National Muffler as allowing “plenary review” contrasted with Chevron deference. See Peoples Federal, 948 F.2d at 305. The Third and Fifth Circuits believe that the traditional rule entitles a regulation to “less deference” than under Chevron. See E.I. du Pont, 41 F.3d at 135 n. 23; Nalle, 997 F.2d at 1139. As Atchison explains, we owe full Chevron deference to a regulation issued with the full deliberative procedures. Although we acknowledge the gulf between the idealized Chevron and the realized one, we do believe that the structure of Chevron encourages a court to defer rather than to interpret. We, therefore, prefer it.
Under Chevron deference we apply a two-step analysis: (1) We examine the text of the statute-in this case, the relevant section of the tax code. If the plain meaning of the text either supports or opposes the regulation, then we stop our analysis and either strike or validate the regulation. But if we conclude the statute is either ambiguous or silent on the issue, we continue to the second step: (2) We examine the reasonableness of the regulation. If the regulation is a reasonable reading of the statute, we give deference to the agency's interpretation. See National R.R. Passenger Corp., 503 U.S. at 417-18, 112 S.Ct. at 1401-02; Chevron, 467 U.S. at 842-43, 104 S.Ct. at 2781-82.
While this circuit has examined legislative history during the first step of Chevron, see Illinois EPA v. United States EPA, 947 F.2d 283, 289 (7th Cir.1991) (“To ascertain the intent of the law, we must examine the plain meaning of the statute at issue, the language and design of the statute as a whole, and, where necessary, less satisfactory indicia of congressional intent such as legislative history.” (citations omitted)), we now seem to lean toward reserving consideration of legislative history and other appropriate factors until the second Chevron step, cf. Alex v. City of Chicago, 29 F.3d 1235, 1239 (7th Cir.1994) (“when statutory meaning is clear with respect to the issue at bar judicial inquiry normally should end without heed to the embellishments of secondary materials like legislative history”). In the second step, the court determines whether the regulation harmonizes with the language, origins, and purpose of the statute. While not dispositive, a court may find various considerations informative-these considerations might include the consistency of the agency's interpretation, the contemporaneousness of the interpretation, and the robustness of the regulation following congressional re-enactment of the underlying statute. Although we sometimes describe Chevron as a “rubber stamp,” see Atchison, 44 F.3d at 442, we know that agencies occasionally act unreasonably. Given the scope of the permissible inquiry under Chevron's second step, we believe that courts can rein in the excesses of unreasonable administrative rulemaking. With that said, however, we reiterate Chevron's fundamental dictate that a court must not “substitute its own construction of a statutory provision for a reasonable interpretation” by the agency in question. 467 U.S. at 844, 104 S.Ct. at 2782.
At long last, then, we begin our examination of Bankers Life's challenge to Treasury Regulation § 1.815-2(b)(3) under the Chevron framework starting with the tax code. Bankers Life seeks to show that the plain meaning of the tax code required the IRS to value distributions according to aggregate adjusted basis. Next, it argues that the regulation unreasonably departed from the harmony and order of the statutory scheme.
Bankers Life, of course, starts by parsing the language of the code section that created the Phase III tax system-I.R.C. § 815:
For purposes of this section and section 802(b)(3), any distribution to shareholders after December 31, 1958, shall be treated as made-
(1) first out of the shareholders surplus account, to the extent thereof,
(2) then out of the policyholders surplus account, to the extent thereof, and
(3) finally out of other accounts.
I.R.C. § 815(a) (emphasis added). The insurance company focuses on the meaning of the terms “distribution” and “out of.” Their argument is that “an element cannot be distributed out of an account unless the element had been included in the account.” Specifically, that SSA and PSA never included the portion of the fair market value of the 1983 distribution attributable to unrealized appreciation. Thus, the IRS should not have deducted the appreciation “out of” the SSA or the PSA because the rules never placed the appreciation into those accounts.
Or, phrased another way, the IRS could not have distributed the fair market value out of the SSA and PSA because “distribution involves the concept of being dispensed from a ‘whole’ that only includes ‘parts' within that ‘whole,’ and a distribution, therefore, cannot include elements that are not a part of the source of the distribution.” Even more basically, if we never put the appreciation in the boxes marked SSA and PSA, how could we have taken the appreciation out of those boxes?
Bankers Life also focuses on the statute's use of the word “accounts.” Since § 815 referred explicitly to “accounts,” logic dictates that it must have mandated some standard form of accounting. The tax code generally employs uniform methods of debit/credit accounting. See, e.g., I.R.C. § 312(a)(3) and Treas. Reg. § 13121(a)-(c) (employing credits/debits system and adjusted basis to measure distributions out of a corporate earnings and profits account during partial liquidation). Since the SSA included only Phase I and II income-after-tax gains-it could not have included unrealized appreciation of assets. Thus, the statute could not have directed the withdrawal of unrealized appreciation from the SSA unless it also allowed for the simultaneous credit of the after-tax gain from that appreciation to the account.
These several arguments lead Bankers Life to the conclusion that Treasury Regulation § 1.8152(b)(3) defied the statute. It defied the plain meaning of “distribution” and “out of.” It defied fundamental principles of accounting evoked by the word “account.” Therefore, according to Bankers Life, we should apply the first step of Chevron and strike the regulation.
Moving beyond the language of the statute, Bankers Life argues that the regulation constituted an unreasonable interpretation of the code. Bankers Life backs up statutory language argument about “accounts” with reference to the “regular”-as opposed to “phased”-taxation rules. The company explains that, normally, unrealized appreciation inherent in an asset does not become part of taxable income until the owner causes a taxable event by selling or exchanging the asset. As Bankers Life explains, the tax laws mark value with reference to basis, they do not “mark to market.” See Cottage Sav. Ass'n v. Commissioner, 499 U.S. 554, 559, 111 S.Ct. 1503, 1507, 113 L.Ed.2d 589 (1991); see also Young v. Commissioner, 923 F.2d 719, 721 (9th Cir.1991) (stating that principles of taxation do not recognize unrealized gain or loss). Since, in this case, the tax laws did not recognize a taxable event under Phase I or II taxation, Bankers Life contends that § 815 could not have suddenly marked to market and created a taxable event for purposes of Phase III taxation alone.
The core of this argument is that by creating a taxable event under Phase III rules the IRS impermissibly imposed a debit without the corresponding credit. Under the regular tax rules a taxable event imposes a tax burden but also confers an accounting credit by adding after-tax income to the SSA and/or the PSA. A distribution would then result in a debit-but only after Bankers Life received the benefit of the accounting credit. By computing the value of the 1983 distribution according to market value, the IRS subjected Bankers Life to a debit without ever having given it the corresponding credit. By creating a taxable event solely under Phase III the IRS imposed the detriment of a taxable event without the benefit. Returning to credit/debit accounting principles, since the SSA received no credit for the unrealized appreciation, the IRS could not attempt to extract the unrealized appreciation from the SSA.
To support its interpretation of § 815, Bankers Life points to language in the Senate Finance Committee report. The report described the SSA as containing “amounts on which the tax has already been paid.” S. Rep., supra, at 26. Since Bankers Life did not yet pay tax on the appreciation element of the real property assets, the SSA could not contain that element and the IRS could not deduct the appreciation from the SSA.
Similarly, the report referred to the PSA as a “cushion.” Since the statute did not allow Bankers Life to add the appreciation element to the PSA, it never became part of the “cushion.” Thus, § 815 could not have conceived of Bankers Life relying upon the appreciation as part of its special reserve. If the appreciation constituted a portion of the Phase III income protected by special treatment, then the law should have attributed that element to the PSA.
Next, the company focuses on the issue of consistency throughout the tax code. The insurer points to a number of areas in which the tax laws treated the value of a distribution as the value of the adjusted basis rather than the fair market value. See, e.g., Fulman v. United States, 434 U.S. 528, 98 S.Ct. 841, 55 L.Ed.2d 1 (1978) (explaining that personal holding company dividends are charged to distributing company at fair market value); H.H. Robertson v. Commissioner, 59 T.C. 53, 1972 WL 2523 (1972) (explaining that debits to foreign subsidiary's earnings and profits account were valued by adjusted basis), aff'd, 500 F.2d 1399 (3d Cir.1974). The regulation destroyed an otherwise harmonious measurement of distributions by adjusted basis.
The company also explains the anomaly produced by the difference between the treatment of cash, compared to noncash, distributions. Because of the General Utilities doctrine, noncash distributions resulted in no taxable event at the corporate level. The regulation constituted a brazenly impermissible end run around General Utilities-an attempt by the IRS to create a Phase III taxable event when no other rule would have imposed a tax.
These arguments lead Bankers Life to the conclusion that, even if Treasury Regulation § 1.815-2(b)(3) did not defy the plain meaning of the statute, it was fundamentally unreasonable given the overall structure of the code. It effectively created a taxable event without any support from the tax law. And it defied the meaning of the statute as revealed in legislative history.
The government responds with several arguments. First, the government points, of course, to the plain language of § 815. It notes that the statute stated that “any distribution shall be treated as made” from out of the SSA and PSA. I.R.C. § 815(a) (emphasis added). The use of “treated as” overcomes any attempt to ascribe precise meaning to “distribute” and “out of” because “treated as” presumed that someone must determine the proper treatment of the distribution. The “treated as” language suggested some level of abstraction, rather than the withdrawal of a concrete asset deposited in a real account. The government expands on this concept by noting that the SSA and PSA were but memorandum accounts, not real asset holding accounts. They constituted bookkeeping entries for the computation of tax, not bank accounts.
This line of argument leads the government to the conclusion that nothing in the controlling statute explained how the IRS should value the distribution. Under Chevron analysis, the government explains that the statute was silent on the issue of valuation-the statute contained a gap on this topic and the IRS filled that gap by issuing Treasury Regulation § 1.815-2(b)(3).
The government also argues that “out of” did not conclusively address the question of how to value distribution. See Fulman, 434 U.S. at 534, 98 S.Ct. at 845 (“out of ․ earnings and profits” language in I.R.C. § 316 did not answer question of how “distribution” is valued). The government claims that “out of” created ambiguity. Thus, again, the government argues that the statute left this question for development by the IRS through regulation.
According to the government, therefore, Treasury Regulation § 1.815-2(b)(3) did not defy the statute. The government argues that the plain language of the statute was either silent or ambiguous on the valuation of noncash distributions. Consequently, we must reach the second step of Chevron analysis and decide whether the regulation constituted a reasonable interpretation of the statute.
To debunk Bankers Life's arguments about inconsistencies, disparities, and anomalies, the government explains that we cannot compare phased taxation to regular taxation. The government finds numerous references in the Senate and House reports to the special tax treatment afforded to life insurance companies under the three-phase structure. Thus, because of the uniqueness of the three-phase structure, the government argues that there is simply no reason for conventional tax accounting principles to apply to the calculation of Phase III taxes. Congress created a unique tax structure and arguments by analogy to conventional taxation shed no light on the reasonableness of Phase III taxation.
The government also advances an argument based on the regulation's longevity and its survival in the face of congressional re-enactment of the underlying statute. This congressional acquiescence argument contends that when a long-standing regulation survives substantial changes to the underlying tax code, we should deem the regulation to have tacit congressional approval and, therefore, the effect of law. See Cottage Sav. Ass'n, 499 U.S. at 561, 111 S.Ct. at 1508. At oral argument, the government attempted to demonstrate congressional awareness of the regulation in legislative history of the Tax Reform Act of 1969. The government bolstered the re-enactment argument by explaining that the IRS promulgated the regulation contemporaneously with the enactment of the three-phase tax structure. See Atchison, 44 F.3d at 442 (listing contemporaneousness among the factors supporting a regulation).
Next, the government examines the legislative history of the statute to elucidate congressional intent and shed light on the law's meaning. Both the House and Senate reports on the statute explained that Phase III taxes applied when the life insurance company made distributions to shareholders: “underwriting gains made available to shareholders will be subject to the full payment of tax.” S. Rep., supra, at 13; H.R. Rep., supra, at 9. Because Bankers Life could have sold the real estate assets and realized gains, the IRS argues it properly deducted this value from the SSA and PSA. In effect, the potential income attributable to the appreciation element constituted a reserve held by the company. The government finds support for this line of reasoning in the reports' description of how Phase III taxation should treat distributions of dividends: “where a life insurance company has distributed dividends to stockholders which are in excess of previously taxed income, it becomes clear that the company itself has made a determination that additional amounts constitute income which was not required to be retained to fulfill the policyholders' contracts.” S. Rep., supra, at 25; H.R. Rep., supra, at 15. The government explains that when Bankers Life transferred assets worth over $875 million to the foundation it clearly signaled that it no longer needed an untaxed cushion of $154 million (the amount in the PSA).
In other words, the concept of a cushion, or reserve, embodied the idea of true value. A life insurance company would not let money in a reserve sit idly in a safety deposit box. On the contrary, the company would invest the money in hope of realizing a gain. Taking this argument to its logical conclusion, the government contends that, by conceiving of Phase III as a deferment mechanism that created a special reserve fund, Congress evinced an intent that Phase III taxation value distributions according to fair market value. Thus, congressional intent embodied in the statute supported the regulation.
Finally, in response to Bankers Life's arguments, the government explains that the regulation did not create taxable income. The entire logic of Phase III taxation involved not the creation of taxable income, but instead the release of taxable income from its special deferred status. The taxable income came into being when the life insurance company made gains from underwriting-the tax code allowed the company to postpone the payment of taxes on half of this income. The deduction of distributions from the memorandum accounts and the valuation of those distributions constituted the methods used to determine when to levy the tax on the Phase III income. In other words, the regulation amounted to no more than a mechanism to determine when and to what extent the IRS should subject Phase III income to its normal tax.
When these arguments were tested in the district court, Judge Plunkett concluded that Bankers Life failed to meet its burden of showing the regulation to be invalid. He reasoned that memo accounts did not necessarily follow the same rules as real accounts. The judge also concluded that, given the unique purpose of special taxation, the regulation was not irrational.
We agree with the district court. Beginning with the statutory language, we cannot glean enough from the plain meaning to determine whether the statute supports or condemns the regulation. The “treated as” language injects so much ambiguity into the meaning of the statute that we cannot accept Bankers Life's interpretation of “distribution” and “out of” as controlling. The “treated as” language makes it impossible to construe the “accounts” as containing specific funds derived from specific sources.
Even more fundamentally, the plain language of the statute simply fails to address the question of valuation. In other words, the statute is silent on this question. In Fulman, the Supreme Court faced a similar situation in which “while obviously some rule of valuation must be applied, Congress ․ failed expressly to provide one.” 434 U.S. at 533, 98 S.Ct. at 845. Here, too, Congress failed to expressly provide a method of valuation. Because the statute is silent, we should not attempt to infer a method of valuation under the guise of determining the plain meaning of the provision.
Moving to the second step of Chevron, Bankers Life fares only a little better. Again, Fulman provides us with guidance. The Fulman Court recognized the “logical force” of the argument against the regulation in question. See id. at 534, 98 S.Ct. at 845. Logical force, however, did not carry the day. The Court explained that “but for the Regulation, the argument might well prevail.” Id. at 536, 98 S.Ct. at 846. The Court continued: “[A]s we have indicated, the issue before us is not how we might resolve the statutory ambiguity in the first instance, but whether there is any reasonable basis for the resolution embodied in the Commissioner's Regulation.” Id. Similarly, Bankers Life's interpretation does not lack logical force, but neither does it rob the regulation of its reasonableness.
The arguments over congressional intent as revealed in the legislative history are in equipoise. The two sides offer conflicting plausible interpretations of Congress's purpose. And Bankers Life certainly demonstrates that the regulation conflicts with the methods of conventional taxation. But conflicts do not equal irrationality. The IRS may adopt regulations that produce unusual results, so long as those results have some colorable support. As Chevron itself explained, an agency may choose among conflicting policies. See 467 U.S. at 845, 104 S.Ct. at 2783. Furthermore, as the government argues, because of the singularity of the Phase III system, any conflicts with regular taxation do not mean very much.
The government also carries the day on the question of whether the regulation creates a taxable event. The regulation merely provides the timing mechanism for the release of Phase III income to taxation-the taxable event occurred when the life insurer earned the Phase III income. Thus, Bankers Life contentions about impermissible marking to market also fail. The IRS did not mark to market to measure the amount of Phase III income earned, it used market value to measure the amount of Phase III income released from its special deferred status.
While we find the government's longevity and reenactment arguments informative, we choose not to place too much weight on them or on the government's argument for the regulation's validity based on its contemporaneity with the passage of the statute. See Smiley v. Citibank (South Dakota), N.A., 517 U.S. 735, 740, 116 S.Ct. 1730, 1733, 135 L.Ed.2d 25 (1996) (“neither antiquity nor contemporaneity with the statute is a condition of validity”). Arguments based on congressional silence presume much. See Brown v. Gardner, 513 U.S. 115, 120, 115 S.Ct. 552, 556, 130 L.Ed.2d 462 (1994) (explaining that arguments based on congressional silence and re-enactment lack persuasive significance). Here, the government's stronger arguments lie elsewhere.
In the end, the government offers a reasonable explanation for the regulation given the origin and purpose of the three-phase taxation scheme. In effect, the IRS adopted a policy that attempted to limit the length of time the insurance company could defer Phase III taxes. Nothing in the three-phase scheme suggested that the IRS could not constrain the benefit of deferment as narrowly as possible. On the contrary, the scheme evinces an intent to tax Phase III income upon large distributions to shareholders. Treasury Regulation § 1.815-2(b)(3) did no more than measure distributions of appreciated real property in the largest manner possible. Because this is not unreasonable given the purpose of the statute, we must affirm.
1. Thus, unless otherwise noted, all citations to the Internal Revenue Code and Treasury regulations are to the statutes and regulations in effect in 1983-namely, the 1954 I.R.C.
2. Strictly speaking, the 1984 tax reforms did not abolish all remnants of the three-phase structure. The new laws grandfathered existing PSA accounts. The 1984 statutes mandated that life insurers maintain existing PSA accounts, see I.R.C § 815(d) (1984), and pay a tax on “the amount of direct and indirect distributions ․ to shareholders from such account,” see I.R.C. § 815(a)(2) (1984).
3. The three-phase tax structure was not really a special deal won by effective lobbying. On the contrary, the scheme enacted a total income approach-taxing revenues from both investment and underwriting. Previously, life insurers paid taxes only on investment revenue. See S.Rep. No. 86-291, at 20-21 (1959).
4. Chevron arguably, therefore, altered the understanding of a line of Supreme Court cases often applied to “interpretive” tax regulations. See National Muffler Dealers Ass'n v. United States, 440 U.S. 472, 99 S.Ct. 1304, 59 L.Ed.2d 519 (1979); Rowan Companies, Inc. v. United States, 452 U.S. 247, 101 S.Ct. 2288, 68 L.Ed.2d 814 (1981); United States v. Vogel, 455 U.S. 16, 102 S.Ct. 821, 70 L.Ed.2d 792 (1982). These cases all distinguished between regulations issued under specific grants of authority (full deference) and regulations issued under general grants of authority (less deference).
5. Because Atchison dealt with a regulation issued by an agency that lacked an express authorization-either general or specific-to promulgate rules, it seems unlikely that general authority tax regulations would fall under Atchison's rule of moderate deference.
6. Having described the fundamental importance of notice and comment procedures, we note that courts could employ other standards. In American Medical Ass'n v. United States, 688 F.Supp. 358 (N.D.Ill.1988), aff'd. in part and rev'd in part, 887 F.2d 760 (7th Cir.1989), the district court eschewed a brightline rule and proposed distinguishing between legislative and interpretive regulations based on whether the regulation defined or gave “substantive content to language undefined in the Code.” Id. at 364. Under this test, “the existence or nonexistence of a precise statutory definition of the term later elaborated on by the Treasury Regulation would mark the watershed.” Id. Thus, this district court advocated an examination of “the nature of the regulation, and not merely the source of authority for its promulgation.” Id. at 365 (citing Chamber of Commerce v. OSHA, 636 F.2d 464 (D.C.Cir.1980)). Other courts have observed that an agency's interpretation deserves diminished deference when it carries the indicia of a litigating position. See Pennington, 22 F.3d at 1383. But cf. Amoco Corp. v. Commissioner, 138 F.3d 1139, 1146 (7th Cir.1998) (explaining that the court should consider a regulation “promulgated through notice-and-comment rulemaking” even if the rule “bears the marks of a litigating position.”).
7. We note the use of the traditional rule in Indianapolis Life Insurance Co. v. United States, 115 F.3d 430, 436 (7th Cir.1997). The Indianapolis Life court, however, did not weigh the applicability of Chevron or the effect of Atchison.
TERENCE T. EVANS, Circuit Judge.