INTERPRETING AMBIGUOUS TAX STATUTES

Abstract

In this essay, I explore the question of who should determine what an ambiguous tax statute means, the courts or the Department of Treasury. The answer to that question is based on two administrative law doctrines: Chevron and Brand X. Here, I explain why Chevron and Brand X violate the Administrative Procedure Act and are unworkable. Then, using a provision in the tax code, I propose that we return to a standard that is both consistent with the APA and easier to implement: Skidmore.

I. Introduction

Imagine it is April 15. Taxes are due. You completed your tax return last night and head to the post office to mail it. (Yes, in the United States many taxpayers still mail tax returns). You chat with the postmaster. She will remember your visit. You eagerly await your refund check; it never comes. Instead, you get a notice of deficiency with penalties. You owe penalties both for failing to file and for failing to pay your taxes by the deadline. ¹ You learn that the Internal Revenue Service (the IRS) never received your tax return because of a postal error. Now, the deadline has passed, and the IRS refuses to consider the postmaster’s affidavit that she remembers you mailing the return by the deadline.

While some courts have said the postmaster should be able to help you prove mailing, the IRS disagrees. Their interpretation of an ambiguous tax statute conflicts. When Congress writes an ambiguous tax statute, who should decide what that language means, the courts or the Department of the Treasury (hereafter, the Treasury)?

An administrative law doctrine known as the Chevron ² doctrine answers this question with a typical law school answer: it depends. If the statutory language is clear, then the Treasury and courts have no power to interpret the language differently. Congressional intent controls. However, if the statutory language is ambiguous, then both the Treasury and the courts have the power to interpret that language.

But what if a court interprets that language first? Is the Treasury bound by the court’s preexisting interpretation? Another administrative law doctrine known as the Brand X ³ doctrine answers this question: surprisingly, no. The Supreme Court decided that agencies, including the Treasury, have the power to reject judicial interpretations of ambiguous statutory language with which the agency disagrees. In the conflict between agency flexibility and certainty, the Court chose agency flexibility.

Chevron and Brand X are not inevitable. Both are simply standards of review that courts apply to determine whether an agency’s interpretation of an ambiguous statute is valid. Using a provision in the tax code for context, this essay explores two problems with these doctrines and recommends an alternative.

I proceed as follows. In Section II, I provide background and context. First, I describe the conflict between the courts and the Treasury in one narrow area: the delivery of tax documents. I explain the rules used to determine when a mailed tax document is delivered to the IRS: (1) the oldest rule, the physical-delivery rule; (2) the middle rule, the common law mailbox rule; and (3) the latest rule, the statutory mailbox rule. I then explain how the Treasury and some courts interpreted the statutory mailbox rule differently. Some courts interpreted the statutory mailbox rule in a taxpayer-friendly way, while the Treasury interpreted the statutory mailbox rule in a less taxpayer-friendly way. As noted above, when an agency’s interpretation conflicts with a preexisting judicial interpretation, the decision of which interpretation controls is based on the Chevron and Brand X doctrines. Hence, these doctrines are described in Section III.

Next, in Section IV, I identify two issues with the Chevron and Brand X doctrines. Specifically, they violate the Administrative Procedure Act (APA) and are difficult to apply individually and collectively. ⁴ After describing these two issues with both doctrines, in Section V, I recommend a return to a prior doctrine, known as Skidmore ⁵ deference, and apply that doctrine to the opening hypothetical. Finally, in Section VI, I conclude.

II. The Mailbox Rules

The U.S. Tax Code runs thousands of pages. ⁶ The implementing regulations and case law run to the tens of thousands of pages. ⁷ Hence, this essay will focus on one very narrow section of the code: the development of the rules governing the delivery of tax documents.

A. The Physical-Delivery Rule versus the Common Law Mailbox Rule

Prior to 1954, to be considered “filed,” tax documents had to be physically delivered to the IRS before any applicable deadline. ⁸ While mail delivery counted, it counted only if the IRS physically received the mailed document before the deadline. However, because documents could be delayed or lost in the mail through no fault of the taxpayer, this physical-delivery rule left taxpayers vulnerable to postal service errors. ⁹

To mitigate the adverse consequences of the physical-delivery rule, courts developed the common law mailbox rule. ¹⁰ Under the common law mailbox rule, extrinsic evidence (like an affidavit from the hypothetical postmaster) that a taxpayer properly and timely mailed a tax document raised a rebuttable presumption that the IRS received the document within the time a normal mailing would take. ¹¹ So, for example, a taxpayer could provide proof that a tax return was mailed on a certain day by offering witness testimony. ¹² Such testimony, when credible, raised a rebuttable presumption that the document was physically delivered to the IRS in the time that the document would ordinarily take to be mailed. While the IRS could rebut the presumption, doing so successfully was difficult.

B. The Statutory Mailbox Rule

In 1954, shortly after the courts developed the common law mailbox rule, Congress passed 26 U.S.C. § 7502 to mitigate the harshness of the physical-delivery rule. Section § 7502 is known as the statutory mailbox rule.

The statutory mailbox rule has two relevant subsections. First, § 7502(a)(1) applies when taxpayers mail a tax document using regular mail. This subsection provides that if the IRS receives a mailed tax document after the deadline, the document will nonetheless be deemed to have been delivered on the date that the document was postmarked if that date precedes the deadline. The statute provides:

If any return, claim, statement, or other document required to be filed, or any payment required to be made, within a prescribed period or on or before a prescribed date under authority of any provision of the internal revenue laws is, after such period or such date, delivered by United States mail to the agency, officer, or office with which such return, claim, statement, or other document is required to be filed, or to which such payment is required to be made, the date of the United States postmark stamped on the cover in which such return, claim, statement, or other document, or payment, is mailed shall be deemed to be the date of delivery or the date of payment, as the case may be. ¹³

Thus, this subsection has two prerequisites: first, the IRS must receive the tax document at some point, and second, the document must be postmarked by the filing deadline.

Second, a different subsection, § 7502(c), applies when taxpayers mail a tax document using certified or registered mail. This section provides that taxpayers may use the registered or certified receipt to prove timely mailing regardless of whether and when the IRS received the document. Thus, this subsection has only one prerequisite: a registration showing the document was mailed timely. ¹⁴ Actual delivery is irrelevant.

The statutory mailbox rule carves out exceptions to the harsh physical-delivery rule. Delivery is not required when the taxpayer uses certified or registered mail. But the statutory mailbox rule does not address what happens when a taxpayer mails a document by ordinary mail and the IRS never receives it. In this situation, the postmark would not be available to serve as proof of mailing. Thus, the statute has a gap.

Should the common law mailbox rule fill this gap? Under the common law mailbox rule, a taxpayer could provide evidence of timely mailing using extrinsic evidence. Courts soon had to grapple with whether taxpayers could still rely on extrinsic evidence to prove timely filing where the taxpayer mailed a tax document using ordinary mail, but the IRS never received it. The courts had to decide whether § 7502(a) abrogated the common law mailbox rule by providing the only avenue by which taxpayers could prove timely delivery when they used regular mail or whether § 7502 augmented the common law mailbox rule. Section 7502 could provide a safe harbor for those cases where the IRS received the mailing late or the taxpayer used registered or certified mail and still leave the common law mailbox rule in place for those cases when the IRS never received the mailing. In short, courts had to decide whether the statutory mailbox rule supplanted the common law mailbox rule or supplemented it.

The circuit courts split on this issue. The Second and the Sixth Circuits held that § 7502 supplanted the common law mailbox rule. ¹⁵ In contrast, the Third, Eighth, and Ninth Circuits held that § 7502 supplemented the common law mailbox rule. ¹⁶ This circuit split caused inconsistency and potential unfairness: taxpayers in the Second and Sixth Circuits were treated differently from those in the Third, Eighth, and Ninth Circuits.

Citing the circuit split, the Treasury amended its relevant regulation, Regulation § 301.7502-1(e), in 2011. With its amendment, the Treasury clarified that the statutory mailbox rule supplanted the common law mailbox rule. Taxpayers who used ordinary mail could no longer prove timely mailing by using extrinsic evidence. If the IRS did not receive the document, the taxpayer was out of luck. The regulation as amended provides:

Other than direct proof of actual delivery, proof of proper use of registered or certified mail, and proof of proper use of a duly designated [private delivery service], are the exclusive means to establish prima facie evidence of delivery of a document to the agency, officer, or office with which the document is required to be filed. No other evidence of a postmark or of mailing will be prima facie evidence of delivery or raise a presumption that the document was delivered. ¹⁷

When the Treasury applied its new regulation, taxpayers challenged it as an invalid interpretation of the statutory mailbox rule. For the courts in the circuits that had held that the statutory mailbox rule supplemented the common law mailbox rule before 2011, the issue was whether they had to accept Treasury’s regulation or reject it because it conflicted with their precedent. If the judiciary has the constitutional authority to interpret statutory language and say what the law is, which interpretation controlled: the Treasury’s or the court’s interpretation? ¹⁸

III. The Chevron and Brand X Doctrines

Resolution of this issue involves two doctrines: Chevron and Brand X. This section explains these two doctrines and why the circuit courts must apply the Treasury’s regulation instead of their own precedents.

A. Chevron

Before 1984, courts applied the Skidmore doctrine to determine whether courts should defer to agency interpretations of ambiguous statutes. ¹⁹ Under Skidmore, agencies earned deference with thorough, careful, and well-reasoned decision-making. ²⁰ Among other factors, courts considered the consistency of an agency’s interpretation when deciding whether to defer to it. Under Skidmore, the status quo was generally maintained. Those wishing to deregulate needed a new doctrine. And they got one.

In 1984, the Supreme Court replaced Skidmore in Chevron U.S.A., Inc. v. National Resources Defense Council, Inc. ²¹ In that case, the Court decided whether courts or agencies had the power to interpret ambiguous, or unclear, statutory language. ²² Chevron involved a question about the meaning of a phrase in the Clean Air Act. ²³ The Act did not define the term. ²⁴ Under the Carter Administration, the Environmental Protection Agency (EPA) promulgated a regulation interpreting the phrase in way that prioritized environmental interests over business interests. ²⁵

In 1980, President Reagan was elected to office after promoting a deregulatory agenda. Under his leadership, the EPA reinterpreted the Carter regulation to favor business growth over protection of the environment. ²⁶ Environmentalists challenged the new interpretation. The issue before the D.C. Circuit Court was whether the EPA’s new interpretation was valid. Applying a de novo standard of review, the D.C. Circuit rejected the EPA’s interpretation. ²⁷

The Supreme Court reversed, upholding the EPA’s business-friendly interpretation. ²⁸ In doing so, the Court rejected the de novo standard of review that the courts had been using to evaluate the validity of agency interpretations of statutory language and created a new deference framework, colloquially known as the Chevron two-step. Pursuant to Chevron’s first step, a court determines “whether Congress has directly spoken to the precise question at issue.” ²⁹ At this step, a court “employ[s] traditional tools of statutory construction” to determine whether congressional intent about the meaning of language is clear or whether there is a gap or ambiguity for the agency to fill. ³⁰ The traditional tools include an examination of the text of the statute, its legislative history, and its purpose. ³¹ Under this first step, courts do not defer to agency interpretations at all. Rather, “[t]he judiciary is the final authority on issues of statutory construction.” ³² This step is de novo review. Had the Court stopped here, Chevron would not have ushered in an administrative law revolution.

But the Supreme Court did not stop at step one. Because it concluded in that case that congressional intent was not clear at step one, the Court continued and created step two. Under step two, the Supreme Court said, a court must accept any “permissible,” or “reasonable,” agency interpretation, even if the court believes a different interpretation would be better. ³³ Deferring to the agency’s interpretation at step two is known as Chevron deference. Unlike the de novo review at step one, this step is very deferential. To be fair, the Chevron analysis has become significantly more complicated than this short description suggests, but it is sufficient for this essay.

The Court offered three reasons to justify its decision to defer to reasonable agency interpretations at step two: agency expertise, implied congressional delegation, and democratic theory. Each justification was grounded in legal positivism, the view that law is a social construct rather than a moral one. ³⁴

First, the Court reasoned that agency personnel are experts in their field, while judges are not. ³⁵ Congress delegates to agencies to implement law in a particular area because of this expertise. For example, scientists and analysts working for the Food and Drug Administration are more knowledgeable about food safety and drug effectiveness than are judges. Similarly, economists and accountants are more knowledgeable about the effects of tax policies than are judges. Because agencies employ specialists, agencies are in a better position to implement legally enacted public policies. Courts lack this expertise; hence, Congress would likely want to delegate interpretive power to agencies. And delegation is constitutional under the delegation doctrine. Under that doctrine, Congress can delegate to the executive branch the power to fill in the details of broadly enacted policy. ³⁶ So long as Congress decides what the policy will be, agencies can lawfully implement those policies. As Cass Sunstein has said:

[S]ometimes [statutory] interpretation is not simply a matter of uncovering legislative will, but also involves extratextual considerations of various kinds, including judgments about how a statute is best or most sensibly implemented. Chevron reflects a salutary understanding that these judgments of policy and principle should be made by [agency] administrators rather than judges. ³⁷

Second, the Court reasoned that Congress simply cannot legislate every detail in a comprehensive regulatory scheme. Even though today’s legislation is complex and detailed, ³⁸ gaps, ambiguities, and even inconsistencies are inevitable. An agency must fill in and resolve these gaps, ambiguities, and inconsistencies to implement lawfully enacted congressional policy. In Chevron, the Court presumed that by leaving these gaps and ambiguities, Congress impliedly delegated the authority to the agencies, not the courts, to resolve them. ³⁹ And there is some truth to this presumption. ⁴⁰ It makes sense that Congress would prefer those with business or finance degrees, rather than those with legal degrees, to resolve ambiguities in the tax code.

Third, the Court reasoned that administrative officials, unlike federal judges, have a political constituency to which they are accountable. “[F]ederal judges—who have no constituency—have a duty to respect legitimate policy choices made by those who do.” ⁴¹ Those who have a constituency include legislators and the executive. Every two to four years, Americans elect a president and legislators. Those legislators and president set and implement new policies. The Chevron doctrine respects this democratic process.

In Chevron, the Court answered the question of what standard of review courts should use to evaluate the validity of agency interpretations of statutory language when there were no preexisting judicial interpretations of that language. The remaining question was what standard of review was appropriate when a court interprets statutory language before an agency does. Can the agency reject the court’s preexisting interpretation? The short answer is: it depends, as I will explain next.

B. Brand X

The Supreme Court answered that question in National Cable & Telecommunications Ass’n v. Brand X Internet Services. ⁴² In this case, the Court explained how to determine the validity of an agency’s interpretation when it contradicts an existing judicial opinion (a preexisting opinion). Recall that Chevron requires a two-step analysis: First, determine whether Congress has spoken to the precise issue before the court using the traditional tools of statutory interpretation (text, legislative history, and purpose). If so, that interpretation controls. But if not, see if the agency’s interpretation is reasonable, even if it is not the best interpretation. In Brand X, the Court explained that a preexisting judicial interpretation of statutory language reached at Chevron’s first step precludes an agency from later interpreting that language differently. ⁴³ When statutory language is clear, an agency has no power to interpret that language differently than Congress intended.

However, if a court determines at step one that Congress did not have a specific intent regarding the meaning of the statutory language and defers to the agency’s reasonable interpretation of the language at Chevron step two, then the agency may change its interpretation later. A prior judicial interpretation resolved at Chevron’s second step does not bind an agency because, under Chevron, agencies have the power to interpret ambiguous statutory language, not courts.

Brand X’s two-pronged approach flows from Chevron. A preexisting judicial interpretation indicates that either congressional intent was clear or that congressional intent was unclear, and the agency’s interpretation was reasonable. If congressional intent is clear, then Congress has spoken, and the agency and the courts must interpret the statutory language consistently with congressional intent. However, if congressional intent is not clear, then the agency may interpret the language differently so long as its interpretation is reasonable. More troubling is that the agency can change its interpretation later so long as the new interpretation is also reasonable.

IV. Problems with Chevron and Brand X

While Brand X seemed to make sense given Chevron’s two-step process, both doctrines violate the APA and are extremely difficult to apply. This section explores these two criticisms.

A. Chevron and Brand X Violate the APA

These doctrines are contrary to the text of § 706 of the APA. ⁴⁴ The APA is the law that governs the procedures that all agencies must follow when regulating. As background, before the APA’s enactment, concerns had arisen about the adjudicatory and other practices of administrative agencies. Each agency followed its own procedures; consistency was missing. Consequently, in 1939, President Roosevelt asked the attorney general to appoint a committee to investigate and suggest procedural reforms. ⁴⁵ The committee issued its final report in 1941; then World War II intervened. After the War, Congress finally passed the APA in 1946. The APA represented a compromise between New Dealers, enthusiastic about expanding agency power, and conservatives, who viewed such power as a veil for tyranny. ⁴⁶

Section 706 of the APA identifies the standards of review for courts to use when reviewing agency actions. That section provides, “the reviewing court shall [1] decide all relevant questions of law, [2] interpret constitutional and statutory provisions, and [3] determine the meaning or applicability of the terms of an agency action.” ⁴⁷ Admittedly, § 706 does not identify the standard of review for the courts to apply when “decid[ing] all relevant questions of law [and] interpret[ing] statutory provisions.” However, there are four reasons this relatively short sentence demands that courts independently interpret statutory provisions, not defer to reasonable agency interpretations.

First, deference is not mentioned. If Congress wanted courts to review agency interpretations of statutes deferentially, Congress could have and should have said so. It did not.

Second, the APA failed to include a specific standard of review for agency interpretations of statutes. Prior to the APA’s enactment, courts used the de novo standard of review to resolve these issues. ⁴⁸ When a legislature enacts a new law, the legislature is presumed to know the existing law and to fit the new law into that legal structure. ⁴⁹ The new law’s failure to include any standard suggests that Congress intended courts to use the standard they were already using: de novo review.

Third, § 706 requires courts to “decide all relevant questions of law.” It does not direct courts to review questions of law an agency has already decided. Further, § 706 directs courts to decide “all” questions of law, not just those questions an agency has not decided. Questions about the meaning of statutory language are questions of law. ⁵⁰ Deciding questions of law is what courts do. It is their expertise. As in any case involving statutory interpretation and as with any litigant or amicus curiae, agencies could suggest interpretations for the court to consider, but it is up to the courts to determine whether the interpretation is consistent with congressional intent.

Fourth, § 706 requires courts to interpret “constitutional and statutory provisions.” These two types of provisions are linked, suggesting that the same standard of review applies to both. Courts interpret constitutional provisions de novo. ⁵¹ (Can you imagine a court deferring to an agency’s interpretation of the Equal Protection Clause or the Second Amendment!) Thus, courts should interpret statutory provisions using the same standard of review: de novo review.

But wait, there is more. In the two subsections of § 706 following the short sentence quoted above, the APA identifies two actions a court may take when reviewing an agency action. A court may (1) “compel agency action unlawfully withheld or unreasonably delayed,” and (2) “hold unlawful and set aside agency action, findings, and conclusions found to be” arbitrary and capricious, contrary to the constitution, ultra vires, procedurally defective, and unsupported by substantial evidence. ⁵² The standards in the second phrase are explicitly identified and are deferential. In contrast, the language quoted contains no explicit standard. Nowhere does § 706 direct courts to defer to reasonable agency interpretations of statutory language or to ignore their own precedents.

For these reasons, the text of the APA requires courts to use de novo review when deciding the meaning of statutory language even when an agency has interpreted that language. “To many modern readers, the most reasonable reading of the APA is that judges must interpret” statutory language de novo. ⁵³ However, the Supreme Court has never addressed these arguments. Chevron did not even cite § 706. ⁵⁴

But knowing what the text says is not enough. What about the legislative history and statutory purpose? Do they offer insights? The purpose of the APA was to make agencies’ procedures more consistent and uniform. This purpose offers little, if any, guidance regarding the proper standard of review for a court to apply when reviewing agency action. ⁵⁵

The legislative history is similarly unhelpful. In the committee report accompanying the draft, APA waffled regarding the appropriate standard of review:

Even on questions of law, [independent judicial] judgment seems not to be compelled. The question of statutory interpretation might be approached by the court de novo and given the answer which the court thinks to be the “right interpretation.” Or the court might approach it, somewhat as a question of fact, to ascertain, not the “right interpretation,” but only whether the administrative interpretation has substantial support. Certain standards of interpretation guide in that direction. Thus, where the statute is reasonably susceptible of more than one interpretation, the court may accept that of the administrative body. Again, the administrative interpretation is to be given weight–not merely as the opinion of some men or even of a lower tribunal, but as the opinion of the body especially familiar with the problems dealt with by the statute and burdened with the duty of enforcing it. ⁵⁶

However, this passage “overstates the extent to which deference to administrative interpretations was contemplated by the APA as evidenced by [the APA’s] text and underlying purposes, both of which argue in favor of independent review.” ⁵⁷ In short, the APA’s unhelpful purpose and ambiguous legislative history cannot trump the statute’s clear text.

In sum, Chevron and Brand X, which is based on Chevron, likely violate the APA. While the Supreme Court has yet to address this issue, some conservative members of Congress have tried for years to legislatively overrule Chevron but, so far, they have been unsuccessful. ⁵⁸

B. Chevron and Brand X are Unworkable

Even if the doctrines did not violate the APA, they should be jettisoned for another reason: they have become exceedingly complex and difficult to apply. This section explores this very real criticism.

1. Chevron’s Application Has Become Uncertain and Unworkable.

Chevron’s two-step approach was easy to apply when it was first decided. A court had to determine whether the statutory language was clear by considering the text, legislative history, and purpose. If the language was clear, the meaning of that clear language controlled. If the language was unclear, or ambiguous, courts would turn to the agency’s interpretation, which was almost always reasonable, even if it was not the best interpretation. Empirical research confirmed that appellate courts deferred to agencies’ interpretations at Chevron’s step two almost routinely. ⁵⁹ Indeed, before he left the bench, former Justice Kennedy chastised the lower courts for their cursory, overly deferential Chevron analysis. ⁶⁰

In the thirty-seven years since Chevron was decided, however, the waters muddied. First, courts, including the Supreme Court, have not approached the two-step process consistently. At times, courts have considered only the statute’s text at step one; at other times courts have considered the statute’s text, legislative history, and purpose. Which process is correct remains unclear even today.

Second, in Chevron, the Court did not explain what makes an agency interpretation reasonable. Should courts continue to examine the traditional tools of interpretation at step two or something else? Some courts have applied arbitrary and capricious review (a standard of review used to evaluate agency determinations of fact) at this step; other courts have stayed with a more traditional statutory interpretation analysis. However, when courts apply the traditional approach and consider the statute’s text, legislative history, and purpose at step one, what is left for the court to consider at step two? The steps become redundant.

In Chevron, the Court did not explain its new second step. Rather, the Court concluded that the EPA’s interpretation of language in the Clean Air Act was a “reasonable accommodation of manifestly competing interests and [was] entitled to deference.” ⁶¹ The Court reasoned that deference was appropriate because “the regulatory scheme [was] technical and complex, the agency considered the matter in a detailed and reasoned fashion, and the decision involve[d] reconciling conflicting policies.” ⁶² Perhaps courts should defer at step two whenever a regulatory scheme is technical, complex, and requires that conflicting policies be reconciled, so long as the agency considered the matter in a detailed and reasoned fashion. ⁶³ But again, almost forty years later, we still do not know.

Third, the Court has added multiple steps to its two-step dance, including step zero and the major questions doctrine (step one and one-half?). Sometimes, the Court has told us, Chevron is not applicable. But determining just when Chevron applies and when it does not apply is not an easy exercise. Pursuant to step zero, an agency must use a procedure having “the force of law” to earn Chevron analysis. ⁶⁴ An agency uses “force of law” procedures when “Congress . . . delegate[s] legislative power to the agency and . . . the agency . . . exercises that power in promulgating the rule.” ⁶⁵ Presumably, formal rulemaking, notice-and-comment rulemaking, and formal adjudication involve force of law procedures, while informal procedures generally, but not always, do not. ⁶⁶ Again, the Court has not been clear. And, pursuant to the major questions doctrine, agencies are not entitled to Chevron analysis when the issue before the Court is one of significant importance to the economy, such as health care and tobacco. ⁶⁷ But what makes one issue significant while another is insignificant? The Court has provided little guidance.

Fourth, what is Chevron? Under the Trump Administration, agencies tried to waive Chevron. ⁶⁸ For example, in 2019, the Bureau of Alcohol, Tobacco, Firearms and Explosives argued that the D.C. Circuit should apply de novo review regarding an interpretation it promulgated using force of law procedures, claiming “Chevron plays no role in this case.” ⁶⁹ The D.C. Circuit disagreed. ⁷⁰ The Supreme Court refused to hear the appeal. ⁷¹

If Chevron is a standard of review, as I and others believe, or a canon of construction for interpreting statutory language, as others claim, then Chevron likely is not waivable. ⁷² But if Chevron is a judge-made rule of decision, ⁷³ then Chevron may be waivable. The circuit courts are split on this issue, and the Supreme Court has not yet addressed it. Hence, whether agencies can waive Chevron has yet to be resolved.

For these reasons, the Chevron doctrine has become incredibly and unnecessarily complex and unwieldy.

2. Brand X’s Application is Similarly Uncertain and Unworkable.

Similarly, Brand X is as, if not more, difficult to apply. You will recall that Brand X is based on Chevron; if the prior case resolved the interpretation at Chevron step one, the agency cannot change the interpretation. If, instead, the prior case resolved the interpretation at Chevron step two, the agency can change the interpretation. Seems simple enough when the prior opinion clearly identifies which of the two steps was decisive; spoiler alert, courts are not always clear.

Regardless, what happens when the judicial decision was issued before Chevron was decided? This question is not hypothetical. In Home Concrete & Supply LLC v. United States, ⁷⁴ the Supreme Court applied Brand X to a judicial opinion rendered in 1958, long before the Court decided Chevron. Determining how to apply Brand X split the Court. Concurring, former Justice Scalia colorfully encapsulated the difficulties with applying Brand X to a preexisting judicial interpretation that predated Chevron:

In cases decided pre-Brand X, the Court had no inkling that it must utter the magic words “ambiguous” or “unambiguous” in order to (poof!) expand or abridge executive power, and (poof!) enable or disable administrative contradiction of the Supreme Court. Indeed, the Court was unaware of even the utility (much less the necessity) of making the ambiguous/nonambiguous determination in cases decided pre-Chevron, before that opinion made the so-called “Step 1” determination of ambiguity vel non a customary (though hardly mandatory) part of judicial-review analysis. For many of those earlier cases, therefore, it will be incredibly difficult to determine whether the decision purported to be giving meaning to an ambiguous, or rather an unambiguous, statute. ⁷⁵

Alternatively, what happens when the first court did not apply Chevron to evaluate the agency’s interpretation? Again, this question is not hypothetical. In Baldwin v. United States, ⁷⁶ the Ninth Circuit faced this conundrum when it addressed whether the statutory mailbox rule supplemented or supplanted the common law mailbox rule.

Before Baldwin, the Ninth Circuit had held that the statutory mailbox rule supplemented the common law mailbox rule in Anderson v. United States. ⁷⁷ Because the Treasury had not yet issued its regulation, the Ninth Circuit applied de novo review to resolve the issue. The court reasoned that the § 7502 did not expressly indicate that registered and certified mail were the only ways to prove mailing when delivery failed. ⁷⁸ Further, the Court said that “‘[a]bsent a clear manifestation of contrary intent, a newly-enacted statute is presumed to be harmonious with existing law and its judicial construction.’” ⁷⁹ The common law mailbox rule existed when Congress enacted § 7502, thus, absent congressional intent otherwise, that statute supplemented the common law mailbox rule rather than legislatively overruling it.

However, in 2011, the Treasury promulgated its new regulation (§ 301.7502-1(e)(2)), which provided the exact opposite: according to the regulation, § 7502 supplanted the common law mailbox rule. Then, the Treasury applied its new regulation to the Baldwins. The Baldwins are movie producers who produced “Ray.” In 2005, the Baldwins timely filed a joint tax return, paying $170,951. ⁸⁰ In 2007, they experienced a net operating loss of more than $2.5 million due to several movie right options terminating. ⁸¹ They wanted to use some of their 2007 loss to offset their 2005 tax liability. So, they prepared an amended 2005 tax return, seeking a refund of $167,633. ⁸² The amended return was due by October 15, 2011. ⁸³

The Baldwin’s employee mailed the amended return on June 21, 2011, long before it was due; ⁸⁴ however, the IRS never received it. ⁸⁵ Accordingly, the IRS denied the refund claim as untimely. ⁸⁶ The Baldwins sued. ⁸⁷ They offered to prove the mailing date through the testimony of their employees. The government moved to dismiss the case, arguing that the Baldwin’s extrinsic evidence of timely mailing was inadmissible under its new regulation. The Baldwins responded that the regulation was inapplicable because it conflicted with the Ninth Circuit’s decision in Anderson. ⁸⁸ Agreeing with the Baldwins, the trial court denied the government’s motion.

The case proceeded to trial. The Baldwins testified that they signed the amended return after receiving it from their tax preparer on June 21, 2011. They gave it to their employee, Nicholas Ruta. Ruta testified that he gave it to another employee, Ryan Wuerfel. Wuerfel testified that he placed it in a green and white stamped envelope addressed to the IRS and put it in the mail at the post office in Hartford, Connecticut on that same day. ⁸⁹ Thus, the amended return should have arrived at the IRS service center well before the October 15, 2011 deadline. The IRS offered no evidence to contradict this testimony. The trial court credited the witnesses and cited Anderson to conclude that the Baldwins’ amended return was timely mailed even if never received. ⁹⁰ The government appealed to the Ninth Circuit.

Thus, the Ninth Circuit had to decide whether its holding in Anderson or the Treasury’s new, contrary regulation applied to the Baldwin’s situation. While the regulation won, the court’s analysis was abysmal. First, and without reference to Brand X or Anderson, the court applied Chevron as though there were no prior judicial interpretation. ⁹¹ Moreover, the court applied Chevron in one, very short paragraph. Although Chevron directs courts to apply “all the traditional tools of statutory interpretation,” ⁹² the Ninth Circuit considered just one such tool: § 7502’s text. The court noted that the statute was “conspicuously silent” regarding whether a taxpayer who sends a document through regular mail could rely on the common law mailbox rule to establish timely delivery. ⁹³ After this cursory analysis, the court concluded that Congress had not directly spoken to this issue. ⁹⁴ The court then applied Chevron’s second step and concluded that the Treasury’s regulation reasonably interpreted the statute simply because the circuit courts had split on this issue. ⁹⁵

It is unclear why the Ninth Circuit applied Chevron. Chevron would be the correct standard of review were there no prior judicial interpretation, because the Treasury issued its interpretation using notice-and-comment rulemaking. But the Anderson decision existed, and Brand X seems to require a different analysis when a prior judicial interpretation exists. When an agency’s interpretation of a statute conflicts with a preexisting judicial interpretation of that same statute, the appropriate standard is Brand X. Applying Brand X, the Ninth Circuit should have determined whether Anderson resolved the issue at Chevron’s first or second step.

There is a credible argument that Brand X should not apply when the prior judicial decision did not involve Chevron, which I will not explore here because the Ninth Circuit did not address this point even in passing. Instead, the court applied Chevron, then moved to Brand X, where the court should have started its analysis.

Moreover, the court’s Brand X analysis fares little better. You will recall that in Brand X, the Supreme Court directed lower courts to determine whether the prior interpretation was resolved at step one or two of Chevron. If the decision were resolved at step one of Chevron, that meant that the prior court would have concluded that Congress intended one interpretation. If the decision were resolved at step two, then the prior court would have concluded that Congress preferred for the agency to interpret the statute and for the court to defer to the agency’s interpretation, assuming it was reasonable.

To be fair, the Baldwin court could not apply Brand X in this way because the agency’s interpretation at issue in Anderson was not entitled to Chevron analysis under Chevron step zero. Brand X directs the subsequent court—here, the Ninth Circuit in Baldwin—to determine at what step of the Chevron two-step procedure the prior court—here, the Ninth Circuit in Anderson—resolved the case. The difficulty here is that because the Treasury had not interpreted the statute using force of law procedures, the Ninth Circuit in Anderson did not apply Chevron.

So, what did the Ninth Circuit do in Baldwin? Ostensibly applying Brand X, the Ninth Circuit reasoned that the Treasury did not have to follow its decision in Anderson because the Anderson decision had not said that the court’s “interpretation of the statute was the only reasonable interpretation.” ⁹⁶ That was the extent of the analysis. In other words, the Ninth Circuit in Baldwin tried to apply Brand X and find the two steps of Chevron in the Anderson opinion, even though that decision never mentioned Chevron. But the court was unsuccessful; “the only reasonable interpretation” is neither the Brand X standard nor the Chevron standard. At best, “the only . . . interpretation” sounds like it could be Chevron’s first step, while “reasonable” sounds like it could be Chevron’s second step.

Thus, the Ninth Circuit’s Brand X analysis in Baldwin was faulty. First, the court applied Chevron to the Treasury’s interpretation and in so doing considered only the text at step one, not all the traditional tools of statutory interpretation. Had the court considered legislative history and especially purpose, the court may have reached a different result. ⁹⁷ Regardless, after incorrectly applying Chevron, the Ninth Circuit then misapplied Brand X and curtly concluded that Anderson was not controlling. ⁹⁸ The Baldwins sought review from the Supreme Court; however, review was denied, and we are left with the Ninth Circuit’s messy decision. ⁹⁹

If Brand X and Chevron are impossible to apply, what should replace them? The courts should return to the standard the courts used before the Supreme Court developed the Chevron two-step: Skidmore deference.

V. Skidmore Deference to the Rescue

Forty years before deciding Chevron, in Skidmore v. Swift & Co., the Supreme Court held that agency interpretations should be given deference only when they are persuasive. ¹⁰⁰ If an agency can persuade a court that its interpretation of a statute is the appropriate one, whether or not the court has already interpreted the statute differently, then the court should respect the agency’s interpretation. After all, many statutes, like the tax code, are highly complex. Congress entrusts agencies like the Treasury to create a comprehensive regulatory regime to implement them, and Congress expects courts to consider agency interpretations when it drafts these statutes. ¹⁰¹

Skidmore involved the appropriate level of deference for a court to give an interpretive rule, a guidance document that does not go through notice-and-comment rulemaking procedures. ¹⁰² The issue in that case was whether certain employees of Swift & Co. were entitled to overtime pay under the Fair Labor Standards Act of 1938 (FLSA). The employees were paid for the work they performed during the day but were not paid overtime for their “in-active duty,” or on-call time, during which they had to remain on company premises even when not working. ¹⁰³ The Department of Labor issued an “interpretive bulletin,” which offered the Department’s interpretation of FLSA on this issue. ¹⁰⁴ Both lower courts ignored the bulletin entirely, deciding the issue de novo, and held that no overtime pay was warranted. The Supreme Court reversed and remanded, directing the Fifth Circuit to consider, but not defer to, the agency’s interpretation. ¹⁰⁵

In so doing, the Court clarified that the agency had “accumulated a considerable experience in the problems of ascertaining working time in employments involving periods of inactivity and a knowledge of the customs prevailing in reference to their solution.” ¹⁰⁶ The Court noted that the agency’s interpretation should not be ignored, because the agency had expertise in this area. ¹⁰⁷ Further, the weight to give an agency’s interpretation should depend on “all those factors which give it power to persuade, if lacking power to control.” ¹⁰⁸ According to the Court, the factors giving an agency’s interpretation “power-to-persuade” included: (1) the consistency in the agency’s interpretation over time, (2) the thoroughness of the agency’s consideration, and (3) the soundness of the agency’s reasoning. ¹⁰⁹ In other words, the more thoroughly considered and well-reasoned an agency interpretation is, the more deference a court should give the interpretation.

Like it had in Chevron, the Court reasoned that deference was appropriate because agencies are experts in their field and are familiar with the industry customs surrounding certain issues. Hence, agency expertise can inform a court’s interpretation. Under Skidmore’s “power-to-persuade” test, agencies are akin to judicial advisors offering expertise in an area of judicial uncertainty; however, courts retain the interpretive power. Thus, a court reviewing an agency’s interpretation would be free to use the traditional tools of interpretation. Skidmore’s “power-to-persuade” test simply added an agency’s interpretation as another source for the court to consider during the interpretive process.

Skidmore is preferable to the other doctrines for two reasons. First, unlike Chevron and Brand X, Skidmore does not violate the APA. Skidmore leaves the interpretive power to the courts. As Professor Peter Strauss described:

What is “exclusively a judicial function” does not exclude agency views. Once a question of statutory interpretation has been put before a court, it is for the court to resolve the question of meaning. Among the matters indispensable for it to consider, however, are the meanings attributed to it by prior (administrative) interpreters, their stability, and the possibly superior body of information and more embracive responsibilities that underlay them. They may be entitled to great “weight” on the judicial scales. ¹¹⁰

And, as noted earlier, ¹¹¹ the Attorney General’s Committee on Administrative Procedure explained that agency interpretations were relevant because of agency expertise and statutory complexity; ¹¹² however, the committee was clear that agency interpretations were to be given consideration, not controlling power. Skidmore is consistent with this view.

Second, Skidmore is much simpler to apply. There is no two-step process, even when a court has already interpreted the same language. Instead, courts consider a variety of factors and weigh them appropriately. This balancing approach forces agencies to explain why their interpretations are accurate. The more consistent, thorough, and well-considered their interpretations, the more likely a court will defer to them, even if the court originally thought the statutory language meant something different. Why? Because agencies have expertise in the subject that courts lack. In short, under Chevron, deference is automatic; under Skidmore, deference is earned.

Let’s return to the initial hypothetical in which our imaginary taxpayer sent her tax return, but the IRS never received it due to a postal error. ¹¹³ By now, you should realize that the hypothetical was based on Baldwin. How would a court approach this issue if the appropriate standard were Skidmore rather than Chevron and Brand X? The court would consider the traditional tools of interpretation: text, legislative history, and purpose. But the court would contemporaneously consider the persuasiveness of both the court’s prior interpretation and the agency’s contradictory interpretation.

So, let’s apply that process. The text of § 7502(a)(1) does not answer whether the taxpayer could offer extrinsic evidence in our hypothetical situation; instead, it addresses other situations, specifically, late delivery by regular mail and late or no delivery by certified or registered mail. On the one hand, the linguistic canon expressio unius suggests that Congress only intended to remedy the specific situations listed and no others. Expressio unius directs that the expression of one thing means the exclusion of that which was not expressed. Here, what was expressed in § 7502(a)(1) is what should happen when the taxpayer mails a document by regular mail and it arrives late, and what was expressed in § 7502(c) is what should happen when the taxpayer mails a document by certified or regular mail, and it never arrives. Not expressed was what should happen when a taxpayer mails a document by regular mail and the IRS never receives it. Thus, because our hypothetical situation was not addressed, expressio unius tells us that Congress likely did not intend § 7502 to address it.

Further, § 7502(c) would be unnecessary if a taxpayer could establish delivery with evidence other than a registered or certified receipt. The rule against surplusage presumes that every word, phrase, and sentence in a statute has meaning. Congress identified how to establish delivery when tax documents are never received. This section would be surplusage if taxpayers could establish delivery ways other than those receipts, such as by testimony.

On the other hand, if Congress wanted to legislatively overrule the common law mailbox rule, then Congress could have said expressly that taxpayers could no longer offer extrinsic evidence regarding the mailing of documents through regular mail when the IRS never receives those documents. Congress is presumed to know the law when it legislates, so if Congress intended to legislatively overrule the common law mailbox rule, Congress should have said so clearly; Congress did not. Arguably then, the common law mailbox rule survived. Thus, the text and linguistic canons do not resolve the statutory ambiguity.

Neither does the legislative history. It is largely silent regarding whether physical receipt of regularly mailed documents is required under § 7502. On the one hand, both the Senate and House reports provide, “This new section applies in the case where documents . . . are mailed to the proper office within the time prescribed by the internal-revenue laws, as indicated by the postmark on the envelope, and are received by that office after such time has expired. In such case, the document is deemed timely filed.” ¹¹⁴ On the other hand, the reports say nothing about what happens when tax documents are mailed by regular mail and never received. Congress knew the common law mailbox rule, yet the reports do not indicate whether § 7502 displaced this rule. If Congress intended to displace that rule, the reports would likely have said so; they did not. Thus, the legislative history does not resolve the ambiguity either.

Finally, the statutory purpose provides the most guidance. Congress enacted the statutory mailbox rule, in part, to mitigate the harsh consequences of Treasury’s physical-delivery rule because that rule left the taxpayer at the mercy of the postal service. ¹¹⁵ Given that Congress intended § 7502 to benefit taxpayers, § 7502 could be read as a safe harbor. Read in this way, § 7502 would create two exceptions to the physical delivery rule that the taxpayer could rely on without question; however, the statute would not bar a taxpayer from relying on other, judicially crafted exceptions when the statutory exceptions did not apply. In those cases, the extrinsic evidence would establish a presumption that the documents were received, but the Treasury could rebut that presumption. The Treasury could not rebut receipt in the situations § 7502 covered.

The agency’s resurrection of the physical-mailbox rule for situations where the taxpayer uses regular mail contravenes this statutory purpose. Indeed, it resurrects the mischief Congress was trying to address. Interpreting § 7502 consistently with its purpose, a court should conclude that the statute supplemented the common law mailbox rule. In other words, Anderson was rightly decided.

Were we deciding this issue on a clean slate, we would end our analysis here. But we are not deciding on a clean slate. We have two more factors to consider: both a prior judicial interpretation and the Treasury’s interpretation. Let’s examine the prior court interpretation first.

Assuming our taxpayer is in the Ninth Circuit, as the Baldwins were, then the court would critically examine Anderson, which rejected the government’s argument that § 7502(c) was the only exception to the statutory mailbox rule of section (a)(1) and that § 7502 supplanted the common law mailbox rule. ¹¹⁶ So, how persuasive was the Ninth Circuit’s reasoning in Anderson? Not very. The court first cursorily examined the text of the statute and concluded that the text did not resolve the issue. ¹¹⁷ Agreed. Then, quoting the reasoning of the Eighth Circuit in Estate of Wood v. C.I.S., ¹¹⁸ the Ninth Circuit said, “‘[a]bsent a clear manifestation of contrary intent, a newly enacted statute is presumed to be harmonious with existing law and its judicial construction.’” ¹¹⁹ The Ninth Circuit considered neither the statute’s legislative history nor its purpose and simply concluded, “We agree with the Eighth Circuit.” That was it. Thus, the Ninth Circuit reasoned first that the text of the statute itself did not show congressional intent to supplant the common law mailbox rule. The court applied the harmony canon second to conclude that Congress wanted the common law mailbox rule to continue. ¹²⁰ Not the most persuasive judicial reasoning.

Finally, we are left to consider the persuasiveness of the Treasury’s interpretation. Under Skidmore, agency interpretations are persuasive when they are consistent over time, thoroughly considered, and soundly reasoned. ¹²¹ First, the Treasury’s interpretation of § 7502 has been extremely consistent both before Congress enacted the statute and after. The Treasury has never wavered that physical delivery is required. ¹²² Indeed, the Treasury’s unrelenting consistency created the conflict!

Second, the Treasury thoroughly considered this issue and did so using notice-and-comment rulemaking. The Treasury first proposed its interpretation in a notice of proposed rulemaking published in the Federal Register in 2004. A public hearing was held on January 11, 2005. ¹²³ However, the Treasury did not promulgate the regulation; instead, it issued a new round of notice-and-comment rulemaking in 2011. During this rulemaking, four commenters “expressed concern that the proposed regulations limited the proof needed to satisfy the timely mailing/timely filing rule of section 7502(a).” The Treasury responded that the “final regulations do not limit the use of U.S. Mail, [rather they] clarify the prima facie evidence of delivery rule of section 7502(c).” ¹²⁴

Two other commenters expressed concern that certified and registered mail were expensive and inconvenient compared to mailing documents using first class mail. Presumably, these commenters were also encouraging the Treasury to maintain the common law mailbox rule, although they did not say so explicitly. The Treasury responded that it had no power as it interpreted § 7502 to allow an additional method by which taxpayers could prove delivery. ¹²⁵ Thus, the Treasury considered this issue and explained why it disagreed with those opposed to its interpretation.

Third, and arguably, the Treasury’s interpretation was soundly reasoned. The Treasury explained that it chose certainty and efficiency. Even though some taxpayers might experience hardship under this rule, certainty and bright lines often limit expensive adjudication. The Treasury noted that if Congress disagrees with that choice, Congress can amend § 7502 to allow taxpayers to prove delivery using extrinsic factors. ¹²⁶ Hence, applying Skidmore, a court would likely defer to the Treasury’s interpretation in its regulation.

Perhaps it should be unsurprising that application of Skidmore to our hypothetical scenario demonstrates that the holding in Baldwin was ultimately correct. The Treasury’s regulation should get deference because it is persuasive, not because the Treasury issued it. But our analysis also shows that the court’s reasoning in Baldwin was unnecessarily convoluted and wrong. A lower court in the Ninth Circuit will have a hard time when it next encounters this same analysis.

Returning to the initial question that began this essay—who should interpret ambiguous statutory language in the tax code—the answer is the judiciary, but with the Treasury’s help. When the Treasury consistently interprets ambiguity in the code, taxpayers can rely on that interpretation and conform their behavior accordingly. Skidmore values and rewards agency consistency. Instead, Brand X and Chevron allow the Treasury, and all agencies, to change interpretations each time a new president takes office, affecting certainty for the regulated. Further, Skidmore values and rewards thorough and well-reasoned decision-making. When the Treasury thoroughly considers how an ambiguity should be resolved by carefully considering public comments and suggestions, taxpayers have more confidence that their voices were heard, even if not agreed with. Finally, when the Treasury offers sound reasoning for its decisions, taxpayers better understand and are more willing to follow the agency’s choice, even if they would prefer a different rule.

VI. Conclusion

This essay explores the weaknesses of two administrative law doctrines, Chevron and Brand X, within the context of the tax code and the statutory mailbox delivery rule. The essay explains that these doctrines likely violate the APA and are exceedingly difficult to apply. Further, it recommends returning to Skidmore’s persuasiveness test and uses a hypothetical to demonstrate why Skidmore enables the Treasury to issue tax regulations, even when they conflict with preexisting judicial interpretations.