Like fences, the statutes in subtitle F of the code define the operational area in which the IRS administers the tax laws. Statutory words — like pickets — place important boundaries on IRS action. Statutory silence — the spaces surrounding the words — can be just as important in understanding the boundaries. As with any other aspect of language, meaning comes from both what is said and what is not said.
This report examines a dispute over the meaning of the spousal relief provisions in section 6015 where the silence is particularly disquieting. The problem is this: While there is a statutory two-year limitation period for taxpayers to seek spousal relief under section 6015(b) or section 6015(c), the statute contains no such limitation period for those seeking section 6015(f) relief. Only silence surrounds section 6015(f). The IRS has taken that silence and, in reg. section 1.6015-5(b)(1), used it to construct an immovable two-year limitation period that the IRS applies against any person seeking section 6015(f) relief. In Lantz v. Commissioner, the Tax Court said the regulation was invalid. As this report was being submit- ted, the Seventh Circuit Court of Appeals reversed the Tax Court, but its opinion is unlikely to settle the issue.
As usual, I divide my report up into easily digestible chunks so you can read part now, part later, and part never. Part A reviews the statutory context of section 6015(f). Part B presents the facts of Lantz and contrasts the majority and dissenting Tax Court opinions. Part C offers my humble (oh please, don’t laugh) thoughts on the matter. I submit that Chevron does not provide a helpful analytical structure for questions about who bears what responsibility for interpreting tax code administrative provisions. Finally, Part D looks at Judge Richard Pos- ner’s opinion and explains why it is wrong. My bottom line is that the IRS’s attempt to appropriate this statutory silence to suit its (admittedly legitimate) administrative needs should not be allowed.
A. Statutory Background
1. Overall context. Tax administration consists of two basic tasks: determining tax and collecting tax. The determination process generally ends when the IRS records a taxpayer’s liability on its computerized books of account. That act is called assessing. Because it processes so many returns, the IRS uses an operational presumption that taxpayers properly report their financial affairs, and so it determines the tax by assessing the tax reported on properly filed tax returns. Sometimes, however, the IRS examines a return and decides that the taxpayer understated his taxes. For income, estate, and gift returns, that resulting discrepancy is called a deficiency, and the IRS can only assess it once the taxpayer has a chance to contest it in Tax Court.
The enforced collection process begins only after as- sessment. It too uses an operational presumption: that those taxpayers whose assessed taxes remain unpaid are able but unwilling to pay. They are ‘‘won’t pays.’’ The highly automated collection system proceeds on that assumption — searching for and seizing taxpayer assets until taxpayers persuade the IRS that they are ‘‘can’t pays’’; they simply cannot fully pay their assessed liabilities. Taxpayers generally do this by applying for collection alternatives such as installment agreements, offers in compromise, or currently not collectible (CNC) status. Taxpayers also have some opportunities during the collection process to persuade the IRS that the liability determination was incorrect. However, taxpayers generally must fully pay the assessed liability before they can seek judicial review of the administrative liability determination.
The liability determination process is about how much tax a taxpayer owes. The collection determination proc- ess is about how much of an admittedly owed tax a taxpayer should be made to pay. Taxpayers who get the same relief under either a liability or collection determi- nation don’t much care about that difference. Either way, they’re off the hook. But the differences are important because of the vastly different substantive and pro- cedural rules that govern each process. Each process presents a different maze of sharp corners for taxpayers to navigate. Also, decisions under each process can have different scope. Section 6015 straddles both processes. It can apply to either change a previous liability decision or change a current collection decision.
2. Description of section 6015(b), (c), and (f). Congress enacted section 6015 as part of the Internal Revenue Service Restructuring and Reform Act of 1998 (RRA ’98) to mitigate some of the harshness of the joint and several liability for income taxes that attaches to jointly filed returns. As I recount in much greater detail elsewhere, the final product was a quick and dirty conference committee compromise. The House proposed a straight- forward reform of the innocent spouse provisions that would apply only to deficiencies. The Senate proposed an opt-out model whereby any taxpayer, whether married or divorced, could choose to basically undo a previously filed joint return and elect to be held liable only for his proportionate share of the liability. The Senate version would apply to any liability, whether self-reported or part of a deficiency, and whether previously paid or not. The compromise was to mix the two together, resulting in what is commonly viewed as three paths to spousal relief. Regardless of the specific compromise, however, the statute’s overall raison d’être is to give taxpayers a remedy when the sharp corners of the tax code create hardship or unfairness.
Specifically, section 6015 marks out three different substantive paths to relief from joint liability for a spouse who so requests (requesting spouse). First, section 6015(b) marks the well-trod traditional innocent spouse route. The requesting spouse must prove that he was (1) innocent of knowledge about the items giving rise to the understatement of tax; and (2) innocent of taking benefit from the understatement.
Importantly, relief under section 6015(b) turns on a determination of prior, unchanging facts. For example, assume Alex and Blair are married and Blair has a secret drug-dealing business. They file jointly, and the revenue agent sniffs out the drug business on audit. To get relief under section 6015(b), Alex will have to prove: (1) the understatement of tax was because of Blair’s unreported income and Alex neither knew about the income nor had any reason to know about it; and (2) Alex derived no benefit from that understatement. Both determinations turn on what happened in the past.11
The second path to relief is through section 6015(c) apportionment. That section gives a type of ‘‘no fault’’ relief for taxpayers who have divorced or are sufficiently separated so that it no longer makes sense to treat them as an economic unit for the liability in question. Section 6015(c) requires the requesting spouse to show the proper allocation of income and deduction items as between the spouses. The requesting spouse’s liability is then limited to just ‘‘the portion of such deficiency properly allocable to the individual under subsection (d).’’ Once the allocation is made, the relief is granted unless the IRS can show either (1) the requesting spouse had ‘‘actual knowledge, at the time such individual signed the return, of any item giving rise to a deficiency… which is not allocable to such individual’’; or (2) the spouses previously transferred assets ‘‘as part of a fraudulent scheme’’ (which the regulations clarify means a scheme to avoid tax and not, for example, just a scheme to defraud other people).
Note again that relief under section 6015(c) turns on an evaluation of past events: what did the requesting spouse know, when and what asset transfers took place, and why.
Using the same Alex and Blair example from above, if Alex and Blair separated or divorced, and the IRS came after Alex to pay the deficiency, Alex would be able to elect the section 6015(c) apportionment regime, under which the unpaid deficiency would most likely be attrib- uted to Blair. Again, this determination would be made on the basis of past facts, and it would be a determination that Alex should not be liable for the deficiency.
The third path to relief is through section 6015(f), which is very short and general. The statute requires the IRS to create procedures to relieve a taxpayer from joint and several liability when:
1. taking into account all the facts and circum- stances, it is inequitable to hold the individual liable for any unpaid tax or any deficiency (or any portion of either); and
2. relief is not available to such individual under subsection (b) or (c).
The section 6015(f) regulations are also very short, and they basically tell taxpayers to ‘‘see… guidance published by the Treasury and the IRS.’’ That guidance is currently found in Rev. Proc. 2003-61, 2003-2 C.B. 296, Doc 2003-17345, 2003 TNT 143-10. It is in the revenue proce- dure that the IRS details what it considers to be the relevant facts and circumstances for granting relief under section 6015(f). I will explain those details in the next subsection.
3. Comparing and contrasting (b) and (c) with (f). Section 6015(f) differs from section 6015(b) and (c) in three important respects. First, subsections (b) and (c) are true liability determinations, but I submit that subsection (f) is best read as a collection determination, notwith- standing its linguistic similarity to (b) and (c). Unlike (b) and (c), subsection (f) is not limited to relief from the assessment of a deficiency. It allows a taxpayer to request relief from underpayments as well as from deficiencies. Whereas relief under (b) or (c) involves a decision about a requesting spouse’s true tax liability, relief under (f) involves a decision that collecting a particular joint liability from one of the spouses would be unfair, even when there is no doubt that, legally, the spouse owes the tax.
One can see this aspect of (f) in the legislative history, the IRS implementation, and court cases. First, the RRA’98 conference committee report tells us that:
The conferees intend that the Secretary will con- sider using the grant of authority to provide equi- table relief in appropriate situations to avoid the inequitable treatment of spouses in such situations [underpayment of self-reported liabilities]. For ex- ample, the conferees intend that equitable relief be available to a spouse that does not know, and had no reason to know, that funds intended for the be considered in determining whether it would be inequitable to hold that spouse jointly liable. Although ‘normal support’ is not considered a significant benefit for this purpose, unusual or lavish support or gifts to the spouse seeking relief are consid- ered even when the benefit is received ‘several years after the years in which the omitted item should have been included in gross income.’’’) (citations and quotes omitted). Courts some- times confuse the section 6015(b) equity test with the section 6015(f) equity test. See, e.g., Doyle v. Commissioner, 94 Fed. Appx. 949 (3d Cir. 2004), Doc 2004-8522, 2004 TNT 77-11. As I explain in the text below, they are different tests. payment of tax were instead taken by the other spouse for such other spouse’s benefit.
One also sees this difference in the revenue procedure. Section 4.01(7) of Rev. Proc. 2003-61 provides that a requesting spouse must, as a ‘‘threshold condition,’’ not be liable for the tax the requesting spouse does not want to pay. That is, the requesting spouse must show that ‘‘the income tax liability from which the requesting spouse seeks relief is attributable to an item of the individual with whom the requesting spouse filed the joint return (the ‘nonrequesting spouse’).’’
The revenue procedure then sets out several excep- tions to this no-liability threshold. One is the ‘‘misappro- priated funds’’ exception. That is, the threshold condition does not apply when ‘‘the requesting spouse did not know, and had no reason to know, that funds intended for the payment of tax were misappropriated by the nonrequesting spouse for the nonrequesting spouse’s benefit.’’ Tracking the conference committee example, the IRS might make a decision in such situations to not collect a tax for which the requesting spouse is legally liable.
Another exception in Rev. Proc. 2003-61 to the no- liability threshold is the ‘‘abuse not amounting to coer- cion’’ exception. That is, the threshold condition does not apply when ‘‘the requesting spouse establishes that he was the victim of abuse before the time the return was signed, and that, as a result of the prior abuse, the requesting spouse did not challenge the treatment of any items on the return for fear of the no requesting spouse’s retaliation.’’
Case law also treats (f) determinations as akin to collection decisions and not liability decisions. For ex- ample, in Maluda v. Commissioner, T.C. Memo. 2009-281, Doc 2009-26778, 2009 TNT 233-15, Mr. Maluda asked for (f) relief because, he claimed, his ex-wife took the money they had set aside to pay taxes and used it for her own benefit (thus invoking the misappropriated funds excep- tion). The IRS denied Mr. Maluda subsection (f) relief from an underpayment of self-reported liabilities on the grounds that the liabilities were completely attributable to his activities and not his ex-wife’s activities. The Tax Court held that the IRS ‘‘inappropriately denied the requested relief solely because the liability was attribut- able to petitioner’s income.’’ However, the IRS still won because the taxpayer’s stipulations failed to prove his claim of misappropriation.
Understanding (f) as a collection decision is important because relatively little IRS collection activity involves the collection of deficiencies. It mostly involves collecting from taxpayers who self-report the correct tax but simply underpay. Consequently, section 6015(b) and (c) do not apply in most collection situations because they allow relief only from deficiencies, not from underpayments. If what the IRS is trying to collect is simply a self-reported but unpaid liability, the requesting spouse cannot invoke subsections (b) and (c).
The second difference flows from the first. Relief under subsections (b) and (c) requires a determination of past facts — which is typical for liability decisions — but relief under section 6015(f) requires a determination of a present status — which is typical for collection decisions. To see this, one must look at how the IRS has imple- mented the broad command of subsection (f) to take into account all the facts and circumstances. Once the thresh- old conditions are met, Rev. Proc. 2003-61 lists various factors it will use in making an equitable determination on whether to pursue collection of a tax that the request- ing spouse admittedly owes. Six of these factors require a determination of present status — that is, they are factors that can change at any time over the 10-year collection period:
a. Economic hardship. Perhaps the most important factor for section 6015(f) relief is ‘‘whether the requesting spouse would suffer economic hardship. The Service will base its determination of whether the requesting spouse will suffer economic hardship on rules similar to those provided in Treas. Reg. section 301.6343-1(b)(4).’’
b. Compliance with income tax laws. A second relevant factor for the IRS is ‘‘whether the request- ing spouse has made a good faith effort to comply with income tax laws in the taxable years following the taxable year or years to which the request for relief relates.’’
c. Mental or physical health. A third relevant factor that ‘‘if present in a case, will weigh in favor of equitable relief, but will not weigh against equi- table relief if not present’’ is the health of the requesting spouse at the time the IRS evaluates the request. The revenue procedure says that the IRS will consider the nature, extent, and duration of illness when weighing this factor.
d. Nonrequesting spouse’s legal obligation. A fourth relevant factor is ‘‘whether the nonrequest- ing spouse has a legal obligation to pay the out- standing income tax liability pursuant to a divorce decree or agreement.’’ Note that this factor might change at any point during the 10-year collection period.
e. Abuse. Fifth, the IRS will factor in ‘‘whether the non requesting spouse abused the requesting spouse.’’ Like the others, this factor is one that can change at any time.
f. Marital status. A final status factor that can change over time is whether the requesting spouse is still part of the economic unit that generated the liability in the first place. Generally, when the marital unit files a return and reports a liability as a marital unit, the marital unit is liable for the re- ported liability. But if one is no longer part of that marital unit, then the IRS will consider that status change in deciding whether to collect the tax. Changes in marital status, of course, can occur at any point during the 10-year collection period.
Of these six present-status determinations, the first four are typically decisions one finds IRS employees making during the collection process, not the liability determination process. The economic hardship factor explicitly relies on reg. section 301.6343, which concerns whether to release a levy because of economic hardship. That reg. section is also used to decide whether to accept a taxpayer’s offers in compromise based on Effective Tax Administration (an ETA OIC). The investigation into the current compliance status is redolent of decisions on whether to accept installment agreements and OICs. The physical and mental health of the taxpayer is also fac- tored into decisions on whether to grant ETA OICs.
Finally, and consistent with the Conference Committee Report, the factor about who was expected to pay for the joint liability has nothing to do with who is liable for the tax but has everything to do with whether it is fair to collect the tax from the requesting spouse.
I cannot overemphasize how each of these status determinations is dependent on current and ever- changing facts. A taxpayer might be flush one year and destitute the next. A taxpayer might be noncompliant one year but then come into full compliance in a later year. And cancer has no respect for time but, indeed, creates its own limitations period. It is true that some of the (f) factors are also important for (b) or (c) relief — notably the inquiry into the requesting spouse’s knowledge about the items giving rise to the unpaid liability at the time the return was filed — but the changing nature of the six factors listed above is unique to the section 6015(f) equity determination.
It is true that the section 6015(b) also contains an equity determination, but not one dependent on present status. The implementing regulation’s explanation of what goes into the equity determination carries forward prior law regarding whether the requesting spouse re- ceived a significant benefit from the understatement. None of the other equity factors listed in that regulation concerns present status.
It is also true that the section 6015(b) regulation assumes that what is now contained in Rev. Proc. 2003-61 provides ‘‘guidance concerning the criteria to be used in determining whether it is inequitable to hold a request- ing spouse jointly and severally liable under this sec- tion,’’ but no one appears to have told the revenue procedure authors about that. The revenue procedure, by its terms, applies only to subsection (f) equity determi- nations. Look at section 4.03 of Rev. Proc. 2003-61 (‘‘Fac- tors for determining whether to grant equitable relief’’). It says that ‘‘this section 4.03 applies to requesting spouses who… satisfy the threshold conditions of section 4.01.’’ When one looks at section 4.01 (‘‘Eligibility for equitable relief’’), one finds that one of the ‘‘threshold conditions’’ is that ‘‘relief is not available to the requesting spouse under section 6015(b) or (c).’’
Further, even if the revenue procedure were not pel- lucid that it applied only to (f) situations, the factors listed in it concerning present status have no logical role to play in the section 6015(b) analysis. The equity analysis in (b) goes to the innocence of the requesting spouse. Section (f) relief, by its terms, goes beyond that, at times giving relief to requesting spouses for whom the equity determination in (b) would provide no relief. You can see how this works in Rev. Proc. 2003-61 section 4.03 that lists ‘‘actual knowledge of the item giving rise to the defi- ciency’’ as a ‘‘strong factor ’’ in determining equity but acknowledges that ‘‘this strong factor may be overcome if the factors in favor of equitable relief are particularly compelling.’’
In sum, the (b) analysis is only whether it is fair to hold a spouse liable for a tax when he had no reason to know about the income item (for example, drug money) giving rise to the understatement. The (f) analysis is about fairness of collecting from an otherwise liable spouse. While present economic status has much to do with a collection decision, it has little connection with a decision about the taxpayer’s true and correct past tax liability. So the regulation is simply wrong. The nonregu- latory guidance on what equitable factors should be used to evaluate the statutory command ‘‘whether it is in- equitable’’ applies only to the (f) situation.
The third difference between (f) and the other two subsections is the breadth of the congressional command. Congress gave the IRS well-defined directions in how to apply subsection (b), which it wrote in light of decades of case law. Subsection (c) was a new path that Congress carved out in 1998, and so the directions are even more detailed. Subsection (f), however, basically just tells the IRS to ‘‘do the right thing’’ and leaves it up to the agency to fill in that considerable gap with regulations. Congress was silent on the contours of what equity means and was silent as to just how taxpayers were supposed to claim (f) relief. The question in Lantz and similar cases is whether one such gap-filling regulation — the one imposing a two-year limitation period on a requesting spouse — is permissible exercise of agency authority.
B. Tilting at Windmills in Lantz
Lantz’s is the quintessential case for subsection (f) relief. She was married to Dr. Chentnik, a dentist who embezzled money from Medicaid and failed to report the embezzled funds as income on the couple’s joint 1999 return.24 He was caught in 2000, prosecuted, and sent to prison for 50 months in 2001. In 2002 the IRS issued a notice of deficiency against the couple to reflect the unreported income from the embezzlement. When nei- ther taxpayer contested the deficiency, the IRS assessed almost $1 million in taxes, penalties, and interest on August 12, 2002, for the couple’s 1999 tax year.
The IRS sent both spouses a CDP notice in May 2003 while Dr. Chentnik was in prison. Dr. Chentnik promised Lantz he would take care of the matter, and he timely requested a CDP hearing. In his correspondence with the IRS, he wrote that Lantz was the innocent spouse and asked for the form he could use to get relief for her. As a result of his efforts from prison, the Office of Appeals put his account in currently not collectible (CNC) status in February 2004. The Appeals officer wrote that ‘‘the tax- payer’s financial condition reflects that the account is noncollectible at this time. Therefore, serving a levy would cause undue hardship for the taxpayer at this time.’’
Lantz had every reason to believe that Dr. Chentnik had taken care of the problem. He repeatedly assured her, from prison, that he was working on the matter. After he was released into a halfway house in early 2004, he sent her a copy of the Appeals letter placing the account in CNC status as proof of his good work. He died shortly thereafter. And when Lantz filed her 2004 return, the IRS sent her the refund she claimed.
There was only one small problem with Dr. Chentnik’s efforts: He had done all of this in his name only. He sought the CDP hearing in his name only, and the letter from Appeals was addressed just to him. Although he had asked for a form to relieve Lantz of the liability, nothing in the record shows that he either received or sent in a form. So when he died in August 2004, even if he thought he had taken care of her problem, he had not.
The first indication Lantz received that her late hus- band had not, after all, kept his promise came in 2006 when, unlike the year before, she did not receive her tax refund. Instead, the IRS sent her a letter in June 2006 explaining that it had snagged her refund and applied it to the 1999 joint liability. She then filed a request for innocent spouse relief.
Lantz’s June 2006 request for spousal relief was too late to claim section 6015(b) or (c) relief. The statute requires a requesting spouse to claim (b) or (c) relief ‘‘not later than two years after the date on which the Secretary has begun collection activities with respect to the’’ re- questing spouse. Reg. section 1.6015-5(b)(2) provides that a CDP notice is a ‘‘collection activity’’ that starts the running of the statutory two-year period for seeking (b) or (c) relief. Because Lantz’s request came more than two years after the May 2003 CDP notice, we will never know if Lantz would have qualified on the merits for spousal relief under section 6015(b) or (c).
Lantz’s only hope was for equitable relief under subsection (f). However, this road to relief was blocked by reg. section 1.6015-5(b), which subjects requests for (f) relief to the same two-year limitation period that Con- gress put in the statute for (b) and (c) relief. Lantz would first have to convince the Tax Court to push aside this properly promulgated regulation. In light of the consid- erable deference courts generally give to agency regula- tions, this was a long shot, at best.
Lantz, however, won in Tax Court. A 12-judge Tax Court majority found the regulation ‘‘an invalid interpre- tation of section 6015.’’ It remanded Lantz’s petition to the IRS ‘‘to consider all facts and circumstances in petitioner’s case.’’ The IRS appealed the case to the Seventh Circuit, and a panel heard oral arguments on April 9. On appeal, the IRS conceded that Lantz would qualify for relief under subsection (f), thus narrowing the issue in the case to whether Treasury may create, by regulation, a two-year SOL for requesting (f) relief.
The basis for the Tax Court’s decision was that it thought the statute prohibited the creation of an absolute limitation period for (f) relief. ‘‘We find that by explicitly creating a 2-year limitation in subsections (b) and (c) but not [subsection] (f), Congress has ‘spoken’ by its audible silence.’’ And, later, ‘‘the Secretary’s adoption of the very timing rule that Congress had imposed on subsec- tions (b) and (c), but had specifically omitted from [subsection] (f) runs directly contrary to the nature of the relief provided by Congress.’’
As an alternative holding, the Tax Court decided that even if the statute were considered ambiguous, the regulation was an unreasonable interpretation of section 6015(f) because its unyielding nature defied the congres- sional mandate for a flexible approach. ‘‘While a tax- payer’s delay in applying for relief under section 6015(f) is a factor to be considered in applying ‘all the facts and circumstances’ test of section 6015(f), the Secretary must be reasonable when creating restrictions that categorically exclude taxpayers from relief… it is clear from the omission of a 2-year limitations period in section 6015(f) that such a 2-year limitations period is impermissible.’’
Three of the five Tax Court judges who disagreed with these holdings wrote dissents. Judges Thornton and Holmes thought the statute was ambiguous for two reasons. First, they bought the government’s double- negative argument that ‘‘Congress didn’t say we couldn’t create a 2-year SOL.’’ Relying on dicta from a Supreme Court decision involving a Bureau of Prisons regulation, they said that the only way a statute such as subsection (f) — one that grants enormous discretion to an agency to make administrative adjudications — could be unam- biguous was if Congress clearly expressed an intent to withhold the authority being challenged. Second, the duo thought that the language ‘‘under procedures pre- scribed by the Secretary’’ created an ambiguity because of the long-standing debate on whether statutes of limita- tion were procedural or substantive in nature.
Once they concluded the statute was ambiguous, the dissent found it easy to conclude that the IRS acted reasonably in prescribing a deadline identical to the deadline for equitable relief under section 66(c).
Judge Halpern added another reason for why he considered the regulation reasonable: He did not believe the regulation was inflexible. He interpreted another set of regulations as giving the IRS discretion to waive the regulation’s two-year SOL. Accordingly, what looked nondiscretionary was, in his view, discretionary.
C. The Tax Court Got It Right
1. I hate Chevron. Everybody starts with Chevron. What a crock. My colleague Ann Graham has persuasively demonstrated in her study of the Roberts Court that “classic Chevron analysis dead or at least critically wounded.’’ The classic Chevron two-step is, of course, as follows: (1) If Congress has directly spoken to the issue before the court, then the court enforces the statute over the agency interpretation, but if the statute is not suffi- ciently clear, then (2) the court must defer to a reasonable agency interpretation of the ambiguous statute.
These two steps have no analytical bite. The first problem is that Step 1 is indeterminate. Courts com- monly describe the degree of clarity they look for in terms of ‘‘ambiguity,’’ as did the Tax Court in Lantz:
If Congress has directly spoken to the precise question at issue, we give effect to the unambigu- ously expressed intent of Congress…. If the statute is ambiguous with respect to the specific issue, we determine whether the regulation is a permissible construction of the statute.
But ‘‘ambiguity’’ is itself ambiguous. The Chevron opinion instructs courts to use ‘‘traditional tools of statu- tory construction’’ in Step 1 to ‘‘reject administrative constructions which are contrary to clear congressional intent.’’ Those tools — embodied in various canons of statutory construction — are far from precise. If there is one universal canon of statutory interpretation, it is that there is no universal canon of statutory interpretation. There is no unified field theory. In a famous law review article, Karl Llewellyn demonstrated that each of the canons of statutory construction has an equal and oppo- site canon. Worse, as Judge Posner put it, there is no ‘‘choice of canon’’ canon. The canons can often lead different judges to different conclusions on the meaning of a statute, even when each judge strives in good faith to find the law. Equally smart judges can disagree on whether Congress has ‘‘directly spoken’’ in the statute the issue before the court. And yet no court has held that such a disagreement, ipso facto, makes a statute ambiguous.
The second problem with this formalism is that if one uses the traditional tools of statutory interpretation in Step 1, then Step 2 risks being little more than a redun- dancy. After all, how does one determine the reasonable- ness of an agency interpretation? Why, by measuring it against the traditional tools of statutory construction! So Step 1 and Step 2 analyses begin to look very much alike. You can see this in some Supreme Court opinions. You can see also it in the Tax Court’s opinion here. Its basic reason for finding that section 6015(f) is ‘‘clear ’’ in Step 1 is also its basic reason for finding that the regulation at issue is ‘‘impermissible’’ in Step 2: Congress put two-year limitations periods in subsections (b) and (c) but not in (f). The Tax Court basically used the same construction of the statute — interpreting the space between the pickets as part of the fencing — to analyze both steps of the Chevron analysis.
So although I believe the Tax Court’s conclusion is correct, I honestly don’t know how my reasons fit into the formal Chevron analysis. I’ll give a stab at where each reason would go, but I would ask any reader who really thinks he or she can figure it out to let me know.
2. Why the Tax Court was right. The Tax Court got it right for three somewhat overlapping reasons: statutory context, equity, and IRS administrative inconsistency. I will discuss each in turn.
a. Statutory context. One of my favorite quotes is that ‘‘text, without context, is pretext.’’ The same can be said of textual silence, such as the silence in section 6015(f). The government’s best argument on why it can create a two-year SOL limitation period focuses on con- trasting the silence in (f) with the language in (f) and boils down to the following double-negative argument: ‘‘Con- gress gave us broad powers in (f) to implement the subsection and did not say in (f) that we couldn’t create a limitation period for (f) requests.’’ This argument con- vinced at least four Tax Court judges that the regulation was valid.
The government’s argument takes the silence out of statutory context — it removes the space from the statutory fence — and not just the context of section 6015 as a whole, which is what the Tax Court considered. The broad powers that Congress gave the IRS, in subsection (f) itself, to prescribe the procedures for using its discre- tion to grant (f) relief are powers that are granted within a much broader statutory context than just section 6015. As I explained in Part A, an important part of that context is that determinations made for (f) relief are collection decisions whereas determinations for (b) and (c) relief are liability decisions.
Once one sees the difference between (b) and (c) determinations on one hand, and (f) determinations on the other hand, one sees how a two-year SOL for (f) determinations violates the statutory context. The limita- tions periods in (b) and (c) are in harmony with the types of limitations imposed on taxpayers and the IRS for revisiting liability decisions. For example, section 6511 requires a taxpayer to submit a claim for refund within two years of making a payment or within three years of filing a return. Likewise, requests for spousal relief under subsections (b) and (c) are requests to revisit a liability determination, and the two-year period is akin to that section 6511 two-year period for refund claims in that both key off of either an actual payment or an enforced payment (such as an offset).
Further, the traditional reasons for statutes of limita- tions apply to requests for relief under section 6015(b) and (c): Memories grow stale, evidence is lost, and all of this means that the past facts necessary for adjudication become less and less discernible as time marches on.
In contrast, the regulation’s two-year SOL for (f) requests is out of harmony with the corresponding collection limitations periods. The IRS has, generally, 10 years to collect assessed taxes. During that 10-year pe- riod, delinquent taxpayers are presumed to be ‘‘won’t pays.’’ However, Congress has created several statutory relief valves that give taxpayers the opportunity to convince the IRS that they are really ‘‘can’t pays.’’ Those relief valves are found in section 7122 (OICs), section 6159 (installment agreements), and section 6015(f) (equi- table relief). Also, the IRS has administratively created other relief valves, notably the CNC decision but also others, like offset refund bypass.
The term ‘‘can’t pay,’’ of course, masks a whole host of policy choices that are made by Congress or the IRS but are generally found in IRS guidance about when it will grant relief from collections. For section 6015(f) deter- minations, the guidance comes in Rev. Proc. 2003-61 which, as I describe above, says that determining eligi- bility for (f) relief necessarily involves a determination of present economic status, present compliance status, and present health status. Those factors may not be determi- native, but they are necessary to consider.
Unlike liability decisions, this ‘‘can’t pay’’ determina- tion is not based on past facts but requires at least three present status determinations, any of which can change at any time during the 10-year collection period. So it is a non sequitur for the Department of Justice to argue on appeal that the two-year SOL is necessary for (f) deter- minations because ‘‘the longer the permissible period is for requesting relief, the more likely it is that memories will fade and documents will be lost that would have shed better light on the matter.’’ That argument simply makes no sense.
The DOJ also repeats an additional argument first found in the Tax Court dissent. It argues that ‘‘allowing a taxpayer to qualify belatedly for relief under subsection (f) would tend to undercut the statutes of limitations’’ for subsections (b) and (c). It explains that ‘‘if section 6015(f) is to be interpreted to apply to a taxpayer who would have qualified for relief under section 6015(b), but simply did not make her election in time, then the statute of limitations [for section 6015(b) relief] would be nulli- fied.’’
This argument might be true if subsection (f) were a liability decision. But subsection (f) is not simply (b) redux. That is, qualifying for relief under (b) is neither necessary nor sufficient to qualify for relief under (f). In the DOJ example, qualifying for (b) relief is not sufficient; not only must the conditions for (b) relief be satisfied, but the requesting spouse must also show economic hard- ship. And, as I discussed above, a subsection (f) suppli- cant who does not qualify for (b) liability relief — either because of too much knowledge or because of being responsible for the items giving rise to the liability — might nonetheless qualify for (f) relief. The DOJ appears to believe this a horrible result, noting that ‘‘the practical effect of the Tax Court’s decision is that a taxpayer is potentially eligible for relief under section 6015(f) as long as the statute of limitations on collections remains open.’’ Yep. That’s the idea, folks. At bottom, subsection (f) is a decision about the fairness of demanding pay- ment. It is no more horrible than allowing taxpayers to apply for OICs at any time during the collection period.
The proper limitation period for subsection (f) relief requests is thus the corresponding collection period. Imposing a two-year SOL for subsection (f) requests cuts taxpayers off from showing that they are ‘‘can’t pays’’ for the remainder of the 10-year collection period. Lantz’s briefs to the Seventh Circuit make this point well. They put the argument in terms of symmetry. Because (b) and (c) are limited to requests for relief from deficiencies, they are always available to spouses in all instances when the IRS seeks to impose this additional liability on what has already been reported. But because (f) relief involves underpayments, it needs to be available as long as the collections period is open. To do otherwise violates the statutory context.
b. Equity. Generally, courts recognize that ‘‘it is not a feasible judicial undertaking to achieve global equity in taxation * * * And if it were a feasible judicial undertak- ing, it still would not be a proper one, equity in taxation being a political rather than a jural concept.’’ The question is what happens when Congress makes the political decision to statutorily command the IRS to do equity in a given class of cases? How far may the IRS go to limit or avoid the unpleasant tax administration con- sequences of that command? That is part of the Lantz puzzle.
In its explicit statutory reference to equity, section 6015 is unique among the procedural statutes in the code. No other provision explicitly incorporates equitable prin- ciples to relieve taxpayers either from liability (part of (b) determinations) or from payment (the (f) determination). For example, Congress commonly allows taxpayers to avoid liability for penalties if they can show some ‘‘reasonable cause’’ for the transgression and also show they acted in ‘‘good faith.’’ Yet these defenses are not equitable in nature, and the courts have had no hesitation in following the strict statutory contours before allowing relief.
Section 7122 is the only other place I can think of where notions of equity come into play, but that is not because Congress enacted them into law. Section 7122 allows the IRS to compromise tax liabilities and nowhere uses the term ‘‘equity’’ or ‘‘equitable.’’ However, the RRA ’98 conference committee report was pretty darned clear that the taxwriters expected the IRS put some type of equitable bases for compromises in the regulations.
And so reg. section 301.7122-1 creates a category of cases in which the IRS will compromise a tax liability on the basis of ‘‘effective tax administration’’ (ETA). The concept is defined in the implementing regulation which makes the chief basis for an ETA compromise the taxpayer’s economic hardship (very similar to the economic hard- ship determination used for granting subsection (f) re- lief). Also, the regulation allows taxpayers to be relieved from enforced collection if they can prove up some “compelling public policy or equity consideration” that “collection of the full liability would undermine public confidence that the tax laws are being administered in a fair and equitable manner.”’ ’[ 59]
Courts have given the IRS wide latitude to craft the particulars of what it will consider for an ETA compromise. As the Tax Court put it: We do not discern in section 7122 an intent of Congress to override application of specific provi- sions of the tax laws in every instance in which the liability is perceived to be unfair or inequitable. * * * The terms of section 7122, the regulations adopted under it, and the Internal Revenue Manual are consistent with the experience and expertise of IRS personnel in evaluating financial circumstances.
Terms such as ‘‘promotion of effective tax administration’’, ‘‘special circumstances’’, and ‘‘compelling public policy or equity consi- derations’’ have a narrower meaning than that urged by petitioners, and the explanations of those terms in the regulations and in the Internal Rev- enue Manual are not unreasonable.
But notice that although the Tax Court allowed the IRS great flexibility (that is, discretion) in how to shape to contours of ETA compromises, the IRS did not there seek to impose an SOL shorter than the collection period SOL. Only in implementing section 6015(f) has the IRS sought to create a categorical timing requirement.
By setting a strict limitations period to administer a command to do equity, the regulation at issue goes beyond what the statute permits. In the long history of equity jurisprudence, statutes of limitation have no place. That is because statutes of limitation do not ‘‘discriminate between the just and the unjust claim, or the avoidable and unavoidable delay. They come into the law not through the judicial process but through legislation. They represent a public policy about the privilege to litigate.’’
Here, Congress has made the explicitly expressed policy choice that the IRS use equitable principles, to consider all the facts and circumstances. That is why its silence in subsection (f) is so important. Congress put in a statutory SOL for subsection (b), which is a mix of legal rules leavened with an equity command, and for subsection (c), which is purely a set of legal rules, but not for subsection (f), which is pure equity.
The equity counterpart to statutes of limitation is the doctrine of laches, which involves a case-by-case deter- mination. The equitable origins and character of section 6015(f) relief point to laches as the proper doctrinal guide to cutting off belated efforts to request spousal relief. Laches is generally used to dismiss plaintiffs from main- taining suits. It requires a finding that the person seeking relief had no good reason to delay the suit and the delay would prejudice to the party opposing relief. As ap- plied to spousal relief, the nonrequesting spouse is the natural ‘‘party opposing relief,’’ and so the IRS would evaluate the prejudice to the nonrequesting spouse, not simply the prejudice to the federal fisc. That is also consistent with the conference committee example of appropriate relief: Where the requesting spouse reason- ably expected the nonrequesting spouse to pay the re- ported tax but the nonrequesting spouse instead used the money for other purposes.
Understandably, the IRS does not want to perform laches analysis for every section 6015(f) request. IRS employees think in terms of bulk processing, and they design programs accordingly. This is evident from almost every national taxpayer advocate (NTA) annual report — the 2008 report is especially illuminating. There, the NTA demonstrates the IRS tendency to overcentralize opera- tions in a never-ending quest for efficiencies. Those with a masochistic streak can also read a recent article I wrote that also demonstrates this trend, starting with World War II.
In general, it is a well-settled principle of administra- tive law that agencies can create and apply bright-line rules in situations in which the agency must adjudicate in bulk. The Social Security ‘‘grid regulations’’ are perhaps the most notorious example. The DOJ seized on this principle to argue that the regulation at issue was ‘‘justi- fiable on grounds of ease of administration.’’ And when the Seventh Circuit asked, during oral argument, what was the ‘‘the grave concern that causes [the government] to want to appeal,’’ the DOJ attorney responded, accu- rately enough, ‘‘ease of administration.’’
The government’s claim of administrative imperatives is open to doubt. First, although the operations are performed in the Cincinnati Centralized Innocent Spouse Operations, it is not clear that the number of requests for relief each year rises to the level of ‘‘bulk.’’ There are probably about 50,000 requests each year, and it is reasonable to think that most of those are made within two years of the first collection action against the request- ing spouse.[69 ] The need to eliminate all of these claims by using a rule because of administrative burden is a stretch. For example, the IRS Office of Appeals received more than 125,000 cases of all types during FY2009 and closed more than 112,000 cases in that time period. In contrast, the IRS must process more than 130 million taxpayer returns each year. Now that’s bulk processing, and regu- lations that eliminate classes of returns from being ac- cepted make sense. A claim that processing 50,000 requests creates the same kind of efficiency needs is much weaker.
Most importantly, administrative reasonableness is not the test of whether a regulation is a permissible interpretation of a statute. There are many reasonable ways the IRS can administer the code. Even accepting the claim of administrative ease, the IRS must choose to act consistently with its statutory authority. Ours is not a system of prerogative regulations, nor does the code contain any ‘‘Henry VIII clauses.’’ If choosing to impose a categorical bar to requests for subsection (f) relief is beyond the statutory pale, it does not matter how admin- istratively desirable such a rule is. The statutory fences do not permit it.
c. Inconsistent IRS administration. The third reason supporting the Tax Court’s decision has to do with how the IRS itself has interpreted section 6015(f). It is the IRS that has interpreted the statute to require the same kind of present-time status determinations used to grant tax- payers relief from collection in other circumstances. Yet in those other circumstances, the IRS does not seek to impose an SOL on those taxpayers but considers requests for relief from collection throughout the entire period during which the IRS may collect the tax.
This inconsistency was discernable at oral argument before the Seventh Circuit, although I don’t think anyone caught it. One of the judges asked government counsel to address a hypothetical situation in which the requesting spouse had been in a coma and had only awoken after the regulation’s two-year SOL had run. The judge was evidently concerned to hear what relief the IRS could provide such a person. Eventually, the government attor- ney got to the right answer: There are other safety valves during the collection process when a taxpayer can get a ‘‘can’t pay’’ determination. Such a taxpayer could be put into CNC status, or might have a case for an ETA OIC. Both types of relief are used to relieve taxpayers who are suffering economic hardship and/or severe medical problems. In the Lantz case itself, the IRS Office of Appeals had ordered CNC status for Lantz’s husband Dr. Chentnik.
What no one apparently pursued during oral argu- ment was the reasons for the inconsistency. That is, if the IRS makes status determinations about taxpayer health and hardship throughout the entire period during which it may collect the tax for OIC and CNC purposes, why not for subsection (f) purposes? Both sets of decisions use the same present status criteria. It appears inconsistent for the IRS to impose a two-year cutoff for such determi- nations for subsection (f) relief. Perhaps if Congress had mandated such a result, that would explain the inconsis- tency, but Congress did not. Congress was silent, both as to the scope of what constitutes equity and as to the period in which the taxpayer could seek such spousal relief.
The inconsistency is of the IRS’s own making. The IRS could have written a regulation that interpreted the scope of equitable relief narrowly, along the lines of what the conference committee report suggested: Subsection (f) relief would be available only in those situations in which the requesting spouse had good reason to believe — at the time the liability arose — that the nonrequesting spouse would pay it. If the IRS had taken that route, then just as the Tax Court has had no problem in upholding the OIC regulations defining the scope of the ETA OIC, so the Tax Court would most likely have no problem upholding the IRS interpretation over what it means forit to be ‘‘inequitable to hold the individual liable for any unpaid tax.’’ By taking a narrow view of the scope of equity in (f), the IRS would have thus limited the equitable determination to past facts, which would more strongly support the use of a categorical limitation period.
Instead of a narrow interpretation of subsection (f), however, the IRS chose, in Rev. Proc. 2003-61, to use equitable factors that have nothing to do with the re- questing spouse’s reliance on the nonrequesting spouse’s promise to pay the liability (whether the liability was self-reported or arose from a deficiency). It chose, indeed, to use the same factors that it uses in making other collection decisions, such as releasing levies, granting ETA OICs and classifying accounts as CNC. Having done so, there appears to be no good reason for imposing a two-year SOL for the one situation and not the others. That is to say, the SOL appears to be arbitrary.
It is no answer to say that the two-year SOL is consistent with the two-year SOLs for subsection (b) and (c) relief. Those are false comparisons because they are liability decisions. One must compare apples to apples. A subsection (f) relief determination is a collection decision, just like the other ones that use economic hardship, health status, and current compliance as tests for grant- ing relief from the payment of taxes admittedly owed.
It is also no answer to say that the two-year period is reasonable because of bulk processing needs or ‘‘ease of administration.’’ It is the inconsistency that is arbitrary, not the regulation standing in isolation. It would be equally reasonable — in the sense of easing the admin- istrative burden — for the IRS to limit all the safety valves that taxpayers can use to come and convince someone that they are ‘‘can’t pays.’’ For example, it would be just as reasonable for the IRS to require taxpayers to submit OICs within two years of the first collection action. But the IRS has not done so. By doing so for subsection (f) determinations and not for similar collection decisions made in other contexts, the IRS acts inconsistently and thus, without some good reason for the inconsistency, arbitrarily.
D. The Seventh Circuit Got It Wrong
The Seventh Circuit’s panel opinion is authored by Judge Posner, the Learned Hand of our ear. But even Judge Hand was wrong sometimes. (75) And Judge Posner’s opinion contains several serious errors. I will first summarize the opinion, then critique it. One aspect of the opinion I liked was that Judge Posner buries his single reference to Chevron in a string cite and does not even try to use its two-step doctrinal structure. Instead, Judge Posner decides that his job is to uphold this Treasury regulation unless it was unreasonable. (76) The opinion then proceeds to dismiss the various reasons given by the Tax Court and the taxpayer on why the regulation was unreasonable.
First, Judge Posner dismisses the Tax Court’s main idea that the statutory omission of a limitations period precluded the IRS from creating one. Judge Posner in- stead buys into the government’s double-negative argu- ment (‘‘Congress didn’t say we couldn’t’’) that I address in Part C.2.a. Second, Judge Posner rejects the taxpayer’s idea that the regulation was unreasonable because it created an asymmetry with the IRS collection period (similar to my ‘‘statutory context’’ analysis above). Third, he rejects the idea that the regulation was unreasonable because it was inflexible, partly because ‘‘neither party suggests that laches might be an adequate substitute for a fixed deadline,’’ but mostly because he buys the gov- ernment’s argument that a deadline for (f) of longer than two years would undermine the statutory limitation periods in (b) and (c). The opinion then finishes up with some miscellaneous thoughts, which do not appear to have independent importance but instead bolster the earlier thinking. For example, Judge Posner says that we must also not overlook the introductory phrase ‘‘under procedures.’’ That thought is part and parcel of the government’s double-negative argument.
Judge Posner’s opinion contains three errors. They are, in order of appearance (but not importance): (1) he misreads the taxpayer’s symmetry argument (an error that eviscerates his rationale for rejecting it); (2) he ignores the critical difference between (f) determinations on one hand and (b) and (c) determinations on the other; and (3) he unthinkingly applies the general administra- tive rule permitting agencies to use rulemaking to in- crease adjudicatory efficiencies without applying that rule to the facts of the case before him. I shall briefly discuss each error in turn.
1. Misreads taxpayer’s argument. Judge Posner says the taxpayer argued that ‘‘because section 6502… imposes a 10-year deadline on the government’s right to collect taxes… there is a time limit on claims for relief under section 6015(f).’’ He then rejects this argument for two reasons. First, he says ‘‘this argument confuses an exter- nal circumstancethat as a practical matter creates a time limit with a statute of limitations.’’ Second, he says ‘‘more important, the 10-year limit in section 6502 is not a constraint on taxpayer action. It’s the period within which the IRS must act to collect a tax… if it does act within this period, section 6502 imposes no time limit on the taxpayer’s response.’’
Judge Posner fights a straw man here. The taxpayer never argued that section 6502 imposed a limitation period on section 6015(f) claims. The taxpayer did argue that the regulation was unreasonable because it violated the statutory context found in a fair reading of subtitle F, basically the same argument that I make in Part C.2.a above, although without all the gory detail about the difference between collection decisions and liability de- cisions. Judge Posner should not be faulted for not following the argument, but it certainly is one reason why one might think the Tax Court better situated to deal with these kinds of cases. Even smart people may be very confused by the technicalities of a system they have never participated in to any great degree. That should at least make Tax Court opinions worth more than their face value to a generalist Court of Appeals.
In the language I quote above, Judge Posner appears to believe that there must be some time limit shorter than the collection period for taxpayers to be relieved of their payment obligations when it would be ‘‘inequitable’’ (the statutory language, remember) to collect from them. That’s a darned strange position to take, and Judge Posner does not here explain what is wrong, either as a matter of law or policy, for a taxpayer to be able to ask for (f) relief during the entirety of the collection period.
It might be that Judge Posner thinks every action permitted by the code should have a limit. However, empirically we know that not every code provision has a time limit. For example, once the IRS serves a summons, there is no limit on the period in which it can petition a court to enforce the summons. The leading case for that proposition is none other than another opinion authored by Judge Posner. Somewhat similarly, the limitations period for filing a refund suit does not start until the IRS disallows a taxpayer’s administrative claim (or else the taxpayer signs a waiver of disallowance). So if the IRS does not send a Notice of Disallowance, the period to file suit is, in practical effect, unlimited. That has bite because a taxpayer may make an informal claim for refund that is never formally disallowed.
It also might be that Judge Posner only means that (f) relief, specifically, should have a limit. Later in the opinion he says that to allow taxpayers (f) relief after the period has run for (b) or (c) relief would undermine those provisions. I explain that error below.
More important, Judge Posner misconstrues the judi- cial task here. It is not to find out what limits might be buried or implied in section 6015(f) to prevent taxpayers from seeking relief from collection. It is to decide whether a regulation that creates a categorical imposition of a two-year SOL for taxpayer’s seeking equitable relief from collection is unreasonable. Just because this regulation is invalid does not prevent the government from writing another regulation that imposes, say, a limitation period measured by the section 6502 collection period. That would be a sensible regulation.
Judge Posner is, in effect, demanding to see a limita- tions period for what is, in effect, an equitable defense. It is like seeking the proper limitations period for equitable recoupment. In the absence of the regulation, the answer would be that taxpayers are permitted to raise this equitable defense so long as the IRS can come after them. I confess total confusion on just why Judge Posner finds it problematic that ‘‘within [the 10-year collection pe- riod], section 6502 imposes no time limit on the taxpay- er’s response.’’
2. Ignored the difference between (f) and (b), (c). The central flaw in Judge Posner’s opinion is that he equates (f) relief with (b) and (c) relief, based on a startling misreading of Rev. Proc. 2003-61. Here’s his analysis:
An applicant who manages to satisfy both criteria in subsection (b) is thus bound to satisfy the criterion in (f). So, on the Tax Court’s view, the two-year deadline imposed by subsection (b) drops away; anyone eligible for relief under (b) is eligible for relief under (f) no matter when she applies.
This reading is what undergird’s Judge Posner’s con- clusion that laches cannot possibly be a reason why the regulation is unreasonable:
In short, if there is no deadline in subsection (f), the two year deadlines in subsections (b) and (c) will be set largely at naught because the substantive crite- ria of those sections are virtually the same as those of (f).
That is absolutely wrong, for two reasons. First, a taxpayer who satisfies both criteria in subsection (b) is not ‘‘bound to’’ satisfy the criteria in (f). The substantive criteria of (b) and (c) are not ‘‘virtually the same’’ as (f). Importantly, Rev. Proc. 2003-61, section 4.02 and section 4.03 incorporate some of the requirements for (b) and (c) relief, but those requirements are neither necessary nor sufficient for (f) relief. Thus, a requesting spouse who meets the threshold requirements of section 4.01 will be evaluated for relief either under section 4.02 or section 4.03 of the revenue procedure. Under section 4.02, the requesting spouse must show a current status of eco- nomic hardship in addition to the (c) requirement of being separated or divorced and the (b) requirement of having had no reason to know of the item giving rise to the understatement or underpayment. Under section 4.03, the requesting spouse is evaluated on economic hardship and also on the other present status factors, all of which may (in admittedly rare circumstances) combine to override even the spouse’s knowledge.
The second reason Judge Posner is wrong to equate (f) with (b) and (c) is the obvious point that (f) not only applies to deficiencies, but also to underpayments. It makes no sense to claim that the underpayment relief undercuts (b) and (c) in any way. And, as I showed above, it is mostly underpayments that the IRS is seeking to collect.
Finally, Judge Posner’s instinct to consider laches is absolutely right. But his reasons for rejecting laches gets things completely backwards. Even as applied to defi- ciency cases, (f) is a safety valve. The whole point of a safety valve is to allow relief when the literal require- ments of the rules in (b) and (c) — including the two-year SOL — are not met. Unlike those subsections, subsec- tion (f) is a pure equity play, and laches is the proper doctrine for the IRS when it believes a taxpayer’s request for relief comes too late to fairly evaluate.
3. Misapplied general administrative law rule. I always warn my students that rhetorical questions are a sign of weakness. ‘‘Never ask a rhetorical question,’’ I say, ‘‘somebody just might answer you.’’ Judge Posner’s opinion falls into that rhetorical trap. Part of the govern- ment’s double-negative argument is that Congress granted the IRS a boatload of discretion to write the
‘‘procedures’’ so that it ‘‘may’’ grant equitable relief. Judge Posner seizes on this to ask:
Since the government can refuse to grant equitable relief to someone who meets the statutory criteria and applies within two year of the first collection action, why can’t it decide to deny relief to a class of applicants defined as those who waited too long?
The short answer to Judge Posner’s question is: judi- cial review. While it is true the IRS can deny relief to someone who meets all the criteria, it is equally true that such person has a statutory right to petition the Tax Court for a review of that decision. Forcing the IRS to deny relief on a case-by-case basis means the taxpayer can invoke the traditional adversarial check on an agency’s abuse of its discretion. The regulation not only cuts off the taxpayer’s ability to seek relief, but also cuts off the Tax Court’s ability to review that reason for rejecting the request. Without the regulation, IRS decisions about timeliness are reviewable as they are made, on a case-by- case basis. Indeed, the Tax Court has many times re- versed IRS denials of (f) relief. As I have demonstrated in painful detail elsewhere, Congress gave a significant push toward requiring adversarial process in tax admin- istration in RRA ’98. Section 6015(f) is part of that push.
The longer answer to Judge Posner’s question is that the statutory context, principles of equity, and the IRS’s own internal decisions on how to administer this statute demonstrate that the regulation is an unreasonable implementation of the statute. I give that answer in Part C.
Most important, Judge Posner again simply asks the wrong question. The Tax Court did not hold that the IR cannot write any kind of limiting regulation, so the generality of Judge Posner’s question is misleading. Indeed, the Tax Court opinion explicitly reserves the question of whether other types of limiting regulations might be appropriate. The general administrative law rule that agencies can write regulations to make their adjudications more efficient is a fine rule as far as it goes. The proper question is whether that general rule of administrative law applies in this case. Judge Posner’s rhetorical question masks his analytical weakness. He simply concludes that ‘‘one would expect Congress to leave it up to the Treasury to establish deadlines opti- mized to the substantive criteria.’’ Oh, sure, but one would also expect the courts to police the Treasury’s exercise of that discretion and make sure that the dead- lines had some reasonable relationship to the statutory context, and to the substantive criteria created by the IRS itself. Here, for the reasons I explain in Part C.2, these regulations fail on both counts.
Statutes, like fences, mark out the boundaries for agency action. We generally think of courts as interpret- ing statutes, and we generally think of agencies as implementing them. Whereas courts interpret statutes ex post, when an agency is challenged and there is a dispute, agencies interpret statutes ex ante, in the context of designing or modifying agency programs or actions. That is why administrative law junkies have repeatedly ob- served that ‘‘statutory interpretation is frequently a poor descriptor of the practice of agencies or the logic of administrative action.’’ Put another way, while courts evaluate agency action in light of the statute, agencies evaluate the statute in light of administrative impera- tives.
In this case, it is not surprising that the IRS would make choices that best suit its centralized innocent spouse program and make the program more efficient. Only after making the regulatory choice would the IRS check to see whether the choice was permissible under the statute. Here, the form of that check was notice-and- comment rulemaking and internal review. But that check was not an adversarial check or a neutral check, even when made by neutral IRS Office of Chief Counsel attorneys acting in good faith and with great skill. That is why we need courts: to perform an adversarial check on agency action.
From either court or agency perspective, the Tax Court’s conclusion was correct: The IRS’s understandable attempt to create a categorical two-year limitation period for section 6015(f) requests contravenes the most reason- able interpretation of the statute and is an unreasonable implementation of the statute.
 Max Knight (translator), The Gallows Songs, Christian Mor- genstern’s Galgenlieder, University of California Press (1964).
 You may assume that all statutory citations refer to the IRC, 26 U.S.C. section 1 et seq. unless I say otherwise in the cite. Section 66(c) contains parallel p r ovisions to section 6015(b) and (f) for spouses living in community p r operty states and my analysis of section 6015(b) and (f) should apply equally to section 66(c).
 Y es, I know that, technicall y , this r egulation was issued by the Department of the T r easury and not by the IRS. But the thinking behind the r egulation, its drafting, and its execution all come f r om the IRS, with T r easury performing a limited (although important) oversight r ole, so please indulge me th r oughout this r eport when I attribute this r egulation (and others) to the IRS.
 132 T .C. 131 (Ap r . 7, 2009), Doc 2009-7979 , 2009 TNT 65-8 .
 After all, the T ax Court operates nationall y . In CC-2010-005 (Ma r . 12, 2010), Doc 2010-5710 , 2010 TNT 51- 1 1 , the O f fice of Chief Counsel designated this issue for litigation and said it ‘ ‘will not settle or concede this issue. ’ ’ The notice inst r ucts attorneys to send section 6015(f) to the IRS for a determination on the merits (made at the Cincinnati Centralized Innocent Spouse Operations). Cases that the IRS r ejects on the merits a r e to be litigated. Howeve r , cases in which the IRS says the r equesting spouse should get r elief must be coo r dinated with the National O f fice to decide ‘ ‘the best course of action. ’ ’ Gosh, I wonder what that can mean if the O f fice of Chief Counsel will not ‘‘settle or concede the issue?’’
 This is a very brief summation. For those who want more detail, please see my prior articles, available at http://ssrn.com/ author=364489.
 Internal Revenue Service Restructuring and Reform Act of 1998, P .L. 105-206, 1 12 Stat. 722.
 See Bryan T . Camp, ‘ ‘The Unhappy Marriage of Law and Equity in Joint Return Liabilit y , ’ ’ T ax Notes , Sept. 12, 2005, p. 1307, Doc 2005-18027 , or 2005 TNT 176-31 (exploring the tension between competing tax policies in context of r eviewing the statutory history of various p r ovisions r elated to joint r eturns f r om 1913 to p r esent); Bryan T . Camp, ‘ ‘Between a Rock and a Ha r d Place, ’ ’ T ax Notes , July 18, 2005, p. 359, Doc 2005-14239 , or 2005 TNT 138-30 (critiquing the T ax Court for taking jurisdiction on stand-alone 6015(f) petitions as over r eaching statutory authority to achieve equity Cong r ess fixed the jurisdictional p r oblem the next year). Cong r ess has since fixed section 6015(e) to give the T ax Court jurisdiction over stand-alone (f) petitions. Both articles a r e also available at http://ssrn.com/ author=364489.
 In the old parlance this is a ‘ ‘ r emedial ’ ’ statute, although that term is not a particularly p r ecise term and carries centuries of baggage. Howeve r , laws that p r ovide a r emedy for the r ed r ess of injuries a r e p r etty squa r ely within anyone’s definition. A good overview of the law can be found in Norman J. Singer and J.D. Shambie Singe r , 3 Sutherland Statutory Construction section 60:2 (264-298) and section 60:5 (307-3 1 1) (7th ed. 2008).
 T o satisfy the first element, the r equesting spouse must show that ‘ ‘he or she did not kno w , and had no r eason to know ’ ’ of the understatement. Section 6015(b)(1)(B). T o satisfy the second element the r equesting spouse must show that ‘ ‘it is inequitable ’ ’ to be held r esponsible for the understatement. Section 6015(b)(1)(D). The equity test he r e basically involves a determination of what, if an y , benefits the r equesting spouse derived f r om the understatement. Hayman v . Commissioner , 992 F .2d 1256, 1262 (9th Ci r . 1993) ( ‘ ‘Although no longer a specific r equi r ement of the statute, the existence of a significant benefit to the spouse claiming r elief nevertheless is a material factor to
[1 1] Note that although the ‘ ‘significant benefit ’ ’ does not have to have happened befo r e the tax r eturn at issue was filed, it still is a determination of something that either happened or did not happen befo r e the trial.
 Reg. section 1.6015-3(c)(3) (emphasis supplied).
 Reg. section 1.6015-4.
 All th r ee p r ovisions a r e written in terms of r elieving a taxpayer’s liabilit y . And certainly in my prior article, supra note 8, I assumed that all th r ee avenues of r elief we r e liability decisions. W riting this r eport gave me an opportunity to think the matter th r ough mo r e ca r efully and, for the r easons I give he r e, I think the better view is that (f) is much mo r e like a collection determination than it is a liability determination. This has consequences for how the T ax Court ought to r eview IRS determinations as well. The T ax Court generally r eviews IRS liability decisions de novo. See, e.g., section 6214. It generally r eviews IRS collection decisions for abuse of disc r etion . See, e.g., section 6330. The T ax Court has long st r uggled with the p r oper way to r eview IRS spousal r elief decisions. See, e.g., Porter v . Commissioner , 130 T .C. 1 15 (2008), Doc 2008-10827 , 2008 TNT 96-12 . I believe that di f fe r entiating between (b) and (c) determinations on one hand, and (f) determinations on the othe r , is a good way to r esolve the p r oblem of r evie w .
 H. Rept. 105-599 at 254.
 See Bryan T . Camp, ‘ ‘ T ax Administration as Inquisitorial P r ocess and the Partial Paradigm Shift in the IRS Rest r ucturing and Reform Act of 1998, ’ ’ 56 Fla. L. Rev . 1, 1 15, note 589 (2004).
 Rev. Proc. 2003-61 section 4.03(2)(a)(ii), cross-referencing section 4.02(1)(c).
 Id. at section 4.03(2)(a)(vi).
 Id . at section 4.03(2)(b)(ii).
 Id . at section 4.03(2)(a)(iv).
 Reg. section 301.7122-1(b)(3)(i).
 Reg. section 301.7122-1(c)(3)(ii), (3)(iv), Example 1.
 Reg. section 1.6015-2(d).
 I take all my facts from the parties’ appellate briefs and from the Tax Court opinion.
 The exact same language is used in section 6015(b)(1)(E) and (c)(3)(B).
 Even the most APA-happy lawyers could find no fault in how these regs were promulgated. Cf. Intermountain Ins. Serv. of Vail, LLC v. Commissioner, 134 T.C. No. 11 (May 6, 2010), Doc 2010-10163, 2010 TNT 88-12 (where two Judges found tempo- rary regulations invalid for not conforming to the Administra- tive Procedure Act’s notice and comment requirements).
 132 T.C. 131, 150. The majority was 11 of the active judges, along with Senior Judge Haines, who the reporter also listed as being in the majority. I am not sure why Judge Haines was listed.
 Id. at 139.
 131 T .C. at 140-141.
 121 T .C. at 148 (emphasis supplied).
 Id. at 153-155.
 131 T.C. at 155-159.
 Id. at 159. The T ax Court majority goes th r ough the history of the r egulations implementing section 66. I do not consider them he r e because the IRS basically t r eats r equests for r elief made under section 66(c) the same as r equests made under section 6015. If imposition of a two-year SOL in the section 6015 r egulation is indeed invalid, I cannot imagine the IRS hanging on to the section 66(c) r egulation.
 Id. 150-152. Judge Halpern basically went to the r egulatory toolshed, looked a r ound, and said: ‘ ‘if taxpayers cannot use a ladder to climb over the SOL barrier, here’s a dandy hammer they can use. ’ ’ Specificall y , Judge Halpern was smitten by the 301.9100-series of regulations. These are regulations designed to give the IRS flexibility in allowing taxpayers extensions of time to make any of the myriad substantive elections they are allowed to make under the code, such as elections to be treated as a resident alien under section 7701(b) or elections to use certain inventory accounting methods under section 474 or elections to forgo the benefits of section 911, and many more. Judge Halpern thought this series of regulations would apply to spousal relief because reg. section 301.9100-1(b) says that ‘‘election includes an application for relief in respect of tax.’’ Although the IRS told the Tax Court that it would not apply those regulations, Judge Halpern dismissed that as ‘‘no more than a litigating position.’’ Judge Halpern did not opine on how a taxpayer was supposed to actually proceed under regulations that the IRS refuses to apply to subsection (f). One presumes he expects taxpayers to follow the procedures in the 9100 regs. Even he admits that most of those procedures ‘‘are of doubtful application to putative innocent spouses.’’ That would be because the IRS ‘‘litigating position’’ is actually the truth: The series 9100 regs are simply not designed nor appropriate for the purpose. It is difficult to see how taxpayers are supposed to know to use the 9100 reg. hammer to climb over the SOL erected by the section 6015 regulations. More importantly, reg. section 301.9100-3(e)(5) says that requests for relief under the series 9100 regs are requests for letter rulings and must follow the revenue procedures for letter rulings. Judge Halpern’s position begs the question of whether the Tax Court has jurisdiction to review an adverse letter ruling simply because the letter ruling applies to section 6015.
 Chev r on USA Inc. v . NRDC, Inc ., 467 U.S. 837 (1984).
 Ann Graham, ‘ ‘Sea r ching for Chev r on in Muddy W atters , ’ ’ 60 Admin. L. Rev . 230, 271 (2008). Essentiall y , P r of. Graham shows that Chev r on is a loser’s a r gument; it is almost always invoked only by the dissents.
 131 T .C. 137-138. The T ax Court majority takes this formulation almost verbatim f r om Chev r on , with one inte r esting, and p r esumably deliberate, omission: in Chev r on the Sup r eme Court said ‘ ‘Rathe r , if the statute is silent or ambiguous with r espect to
the specific issue, the question for the court is whether the agency’s answer is based on a permissible const r uction of the statute. ’ ’ 467 U.S. at 843 (emphasis supplied).
 467 U.S. at 843, note 9. Some might disag r ee with the Chev r on emphasis on ‘ ‘intent of Cong r ess. ’ ’ Justice Holmes once w r ote: ‘ ‘ W e do not inqui r e what the legislatu r e meant; we ask only what the statute means. ’ ’ Oliver W endell Holmes, ‘ ‘The Theory of Legal Interp r etation, ’ ’ 12 Har v . L. Rev . 417, 419 (1899).
While some ‘ ‘textualists ’ ’ might give primacy to the text while ‘ ‘purposivists ’ ’ give primacy to the motives of those who w r ote the text, the two app r oaches have much mo r e in common than their ext r eme versions suggest, as demonstrated in the following very inte r esting article: John F . Manning, ‘ ‘What Divides
T extualists F r om Purposivists, ’ ’ 106 Colum. L. Rev . 70 (2006).
 Karl N. Llewellyn, ‘ ‘ Remarks on the Theory of Appellate Decision and the Rules or Canons About How Statutes A r e to Be Const r ued, ’ ’ 3 V and. L. Rev . 395 (1950).
 Richa r d A. Posne r , ‘ ‘Statutory Interp r etation — in the Class r oom and in the Court r oom, ’ ’ 50 U. Chi. L. Rev . 800, 806 (1983).
 See, e.g., Maislin Industries v. Primary Steel, 497 U.S. 116 (1990) (finding the specific statute at issue ambiguous (Step 1) but nonetheless finding the agency interpretation contrary to ‘‘the statutory scheme as a whole’’ (Step 2)) 497 U.S. at 131. Some courts have restricted their Step 1 analysis by using only certain tools of statutory interpretation, saving the other tools for their Step 2 analysis. See, e.g., Ball, Ball & Brosamer v. Reich, 24 F.3d 1447, 1451 (D.C. Cir. 1994) (using ‘‘plain meaning’’ analysis for Step 1 and evaluating Step 2 reasonableness by using the ‘‘language, legislative histories, and policies of the statute’’).
 The r e a r e, of course, other app r oaches. One article worth your time makes a very nice a r gument that Step 2 analysis is r eally an arbitrariness r evie w . See Ronald M. Levin, ‘ ‘The Anatomy of Chev r on: Step T wo Reconside r ed, ’ ’ 72 Chi.-Kent L. Rev . 1253 (1997).
 This is p r obably a Step 1 a r gument, but who r eally knows? Or ca r es?
 Jesse Jackson, quoted in Sheldon R. Gawiser and G. Evans W itt, A Journalist’s Guide to Public Opinion Polls (1994), p. 11 1. Jackson himself most likely was bor r owing f r om D r . Donald A. Carson who r eportedly has written in his book Exegetical Fallacies (1984): ‘ ‘My father used to tell me that a text without a
context becomes a p r etext for a p r oof text. ’ ’ See http:// en.wikipedia.o r g/wiki/D._A._Carson. I have no doubt the phrase has even longer antecedents.
 See Internal Revenue Manual 188.8.131.52.2 (May 20, 2008) (‘‘Offset Refund Bypass Procedures’’). The offset refund bypass is basically a way for taxpayers undergoing financial hardship to stop the otherwise automatic offsets of their tax refunds by convincing the IRS to manually override the machines. As with other relief valves, this one requires taxpayers to convince someone in the IRS that they are ‘‘can’t pays,’’ at least to the extent of getting their overpayments actually refunded to them instead of applied to their tax delinquencies.
 Again, a good example is found in IRM 184.108.40.206 which details the criteria and procedures for IRS personnel to use when evaluating a request for an offset refund bypass.
 Brief for the Appellant at 46.
 Id .
 Reply Brief for the Appellant at 1 1.
 Re v . P r oc. 2003-61, section 4.02(1)(c).
 Brief for the Appellant at 33, n.10.
 Again, a good example is found in IRM 220.127.116.11 which details the criteria and procedures for IRS personnel to use when evaluating a request for an offset refund bypass.
 Brief for the Appellant at 46.
 Id .
 Reply Brief for the Appellant at 1 1.
 Re v . P r oc. 2003-61, section 4.02(1)(c).
 Brief for the Appellant at 33, n.10.
 Lantz has been capably represented throughout the litiga- tion by Rob Nadler of the Legal Aid Society of Middle Tennessee and Paul Kohlhoff from the Valpariso University Law Clinic.
 This is probably also a Step 1 argument. But I bet many would say it’s a Step 2.
 Kenseth v. Commissioner, 259 F.3d 881, 885 (7th Cir. 2001), Doc 2001-21203, 2001 TNT 154-9 (citations and quotation omit- ted).
 Sea r ching the code on LEXIS for the terms ‘ ‘equitable ’ ’ or ‘ ‘inequitable ’ ’ I found only th r ee other statutes in subtitle F (P r ocedu r e and Administration) using the terms: sections 6305 and 6402 which p r ohibit courts f r om hearing certain subjects, whether raised in ‘ ‘legal or equitable ’ ’ forms; and section 65 1 1 (allowing T ax Court jurisdiction over ‘ ‘equitable r ecoupment ’ ’ claims). Also, Cong r ess w r ote an o f f-code p r ovision that is
placed in then notes following section 7804 that uses the term. It is section 1204 of the Internal Revenue Service Rest r ucturing and Reform Act of 1998, P .L. 105-206, 1 12 Stat. 722, which r equi r es that IRS employees be evaluated, in part, on their ‘ ‘fair
and equitable t r eatment of taxpayers. ’ ’
 See, e.g., sections 6664(c) ( ‘ ‘Reasonable Cause Exception for Underpayments ’ ’); 6694(a)(3) ( ‘ ‘ T ax Return P r epa r er Penalty ’ ’).
 See, e.g., Chaplin v . Commissioner , T .C. Memo. 2007-58, Doc 2007-6198 , 2007 TNT 49-16 ( r ejecting taxpayer’s a r gument that imposing section 6662 penalties was unfair because the adjust-
ments had r esulted in a huge AMT hit).
 H. Rep. 105-599 (June 24, 1998) at 289.
 Reg. section 301.7122-1(a)(3).
 Speltz v . Commissioner , 124 T .C. 165.
 Chase Securities Corp. v . Donaldson , 325 U.S. 304, 314 (1945).
 See, e.g., AMTRAK v. Morgan, 536 U.S. 101 (2002); Bensky v. Powell, 391 F.3d 894 (7th Cir. 2005).
 National T axpayer Advocate, ‘ ‘2008 Annual Report to Cong r ess, ’ ’ at 260-274, Doc 2009-241 , 2009 TNT 4-21 .
 Bryan T . Camp, ‘ ‘Theory and Practice in T ax Administration, ’ ’ 29 V i r ginia T ax Review 101 (forthcoming), available at
 See, e.g., National Pet r oleum Refiners Ass’n v . FTC , 482 F .2d 672 (D.C. Ci r . 1973) (Federal T rade Commission can define the parameters of acts which constitute ‘ ‘unfair methods ’ ’ by r egulation in inte r ests of administrative e f ficiency).
 See Heckler v . Campbell , 461 U.S. 458 (1983) (SSA can define the types and numbers of jobs that exist in the national economy by r egulation for use in disability determinations in inte r ests of administrative consistency).
 Brief for the Appellant at 45.
 Oral a r gument available at http://www.ca7.uscourts.gov/ tmp/YB1FFNCU.mp3/.
 The IRS Data Books do not show how many claims for spousal relief are received each year, but the National Taxpayer Advocate’s 2005 Annual Report says there were some 56,000 Form 8857 dispositions in FY 2003, 57,000 in FY 2004, and 48,000 in FY 2005. I assume those numbers have remained constant.
 IRS 2009 Data Book, T able 21, Doc 2010-5330 , 2010 TNT 48-18 .
 See generally Ronald M. Levin, ‘ ‘The Anatomy of Chev r on : Step T wo Reconside r ed, ’ ’ 72 Chi.-Kent L. Rev . 1253, 1260 (1997) (a r guing that the Court’s step two standa r d is vague and ‘ ‘seems to ve r ge on internal incohe r ence ’ ’). Under the formalities of Chev r on , the r easonableness inquiry is part of the Step 2 analysis.
 See generally Noga Morag-Levine, ‘ ‘Agency Statutory Inte r p r etation and the Rule of Common La w , ’ ’ 2009 Mich. St. L. Rev . 51, 52 (2009). I highly r ecommend this short and informative history of agency r ulemaking. ‘ ‘Henry VIII clauses ’ ’ started with King Henry VIII (as you might have guessed). Parliament enacted a ‘ ‘Statute of P r oclamations ’ ’ which basically said the King could issue any p r oclamations he darned well pleased ‘ ‘concerning the advancement of his commonwealth and good quiet of his people. ’ ’ Id . at 57, n. 30. This b r oad language allowed the King to make all kinds of laws, insulated from the argument that they we r e made in de r ogation of Parliamentary powe r . The statute also was a back-door affirmation by Parliament that it was the source of legislative authority. In the rise of the British administrative state in the 1880s, Parliament used similar clauses to authorize agencies to modify the law when necessary, which would then be put to Parliament for post hoc approval.
 If one r eally wants to put this in Chev r on terms, one might sa y , as Richa r d Murphy suggested to me, that the r efe r ence to equity in (f) takes it out of the r ealm of statutory silence — so no need for the Tax Court’s ‘‘audible silence’’ formulation. That is, by expressly and unusually invoking equity in (f), Congress said out loud not to use SOLs.
 I think this is a Step 2 a r gument. Furthe r , it’s a possible app r oach to Step 2 that makes the step distinct f r om Step 1 because this a r gument evaluates what the agency has done in terms of administration, not in terms of statutory command.
 See, e.g., Universal Camera Corp. v. NLRB, 340 U.S. 474 (1951), on remand at 190 F.2d 429, 431 (2d Cir. 1951) (Frank, J., concur- ring). (‘‘Recognizing, as only a singularly stupid man would not, Judge Hand’s superior wisdom, intelligence and learning, I seldom disagree with him, and then with serious misgivings.’’)
 Opinion at 4. I think by this Judge Posner means that it might be either an unreasonable read of the statute or an
unreasonable implementation. Judge Posner says he is giving no deference to the Tax Court, but that the panel would reach the same result even with deference. I do think opinions of the Tax Court deserve much more respect, because of its unique role and placement in tax administration. This would be a good subject for another article. You have been warned.
 Opinion at 9-11; the quote is on p. 9.
 Opinion at 6 (emphasis added).
 Id .
 United States v. Administrative Enterprises Inc., 46 F.3d 670 (7th Cir. 1995), Doc 95-1909, 95 TNT 27-35.
 See, e.g., Provident Life & Cas. Ins. Co. v. Ginther, 2000 U.S. Dist. LEXIS 5168 (W.D.N.Y. 2000) (allowing defense of equitable recoupment as long as the recoupment related to the conduct that formed the basis of the plaintiff’s complaint against the defendant).
 Opinion at 10, citing to the r evenue p r ocedu r e (Judge Posner emphasized ‘ ‘both ’ ’; I emphasized ‘ ‘bound to ’ ’).
 I am indebted to Lavar Taylor for this thought.
 Opinion at 12.
 I am indebted to Bob Nadler for this thought.
 This is what Judge Halpern was most likely getting at in his concurrence, that taxpayers could somehow leverage section
9100 r egs to get court r eview even after the two-year period. I explain the difficulties with that idea above.
 See Bryan T. Camp, ‘‘Tax Administration As Inquisitorial Process and The Partial Paradigm Shift in the IRS Restructuring and Reform Act of 1998,’’ 56 Fla. L. Rev. 1, 115, note 589 (2004).
 Noga Morag-Levine, ‘‘Agency Statutory Interpretation and the Rule of Common Law,’’ 2009 Mich. St. L. Rev. 51, 52 (2009).
Previously published by the Texas Tech University School of Law, 2010
A. Statutory Background
B. T ilting at W indmills in Lantz
C. The T ax Court Got It Right
D. The Seventh Circuit Got It W rong
One time there was a picket fence with space to gaze f r om hence to thence. An a r chitect who saw this sight app r oached it suddenly one night, r emoved the spaces f r om the fence, and built of them a r esidence. The picket fence stood the r e dumbfounded with pickets wholly unsur r ounded, a view so loathsome and obscene, the Senate had to intervene.
— From The Picket Fence by Christian Morgenstern