chevron-down Created with Sketch Beta.

The Tax Lawyer

The Tax Lawyer: Winter 2025

The Tax Separation of Powers

Susannah Camic Tahk

The Tax Separation of Powers
Brus_Rus via Getty Images

Jump to:

Abstract

Traditional accounts of the separation of powers start from a familiar premise: within the federal government, different branches perform different functions. Put simply, the legislature writes the law and the judiciary interprets it. Using original data from the first large-scale comparative study of the tax judiciary and legislature, this Article finds that the separation of powers in tax law diverges from this standard account. In tax, the branches not only perform different functions, but they also deal with different subjects. This Article empirically documents how some sections of the Internal Revenue Code belong to the Tax Court and others belong to Congress. Those sections differ significantly in their substance. This Article proposes a theoretical framework for understanding this previously unnoticed substantive separation between the Tax Court’s “judicial Code” and Congress’s “legislative Code.”

The Article makes three key contributions to the existing literature. First, its large-scale comparative study provides the first account of the separation of powers in tax law. Second, the Article theorizes the separation and examines potential reasons for it, offering explanations in terms of textual factors and factors of political economy. Third, the Article draws out what its novel account brings to the broader literature on separation of powers. Specifically, the idea that separation of powers can be about substance as well as function upends many of the literature’s foundational assumptions and introduces a crucial new dimension to the traditional analysis. The Article also considers normative implications of its findings for both tax law and broader questions about distribution of government power.

I. Introduction

Who does what in federal tax law? That is the question this study set out to address. According to traditional accounts of the separation of powers, the legislature should be making tax law, and the federal courts, particularly the Tax Court, should be interpreting it. But what, as a practical matter, does that look like? Broadly speaking, what kind of decisions is the Tax Court making and what kinds belong to Congress?

Traditional theories of the federal separation of powers suggest one answer: Congress makes tax law and the federal courts, especially the Tax Court, interpret it. In other words, the separation of powers between Congress and the judiciary should be about different functions. However, through the first comparative study of the tax judiciary and the tax legislature, this Article reveals that the separation of powers in federal tax law confounds traditional expectations. The separation of powers in tax law is about substance, not just about function. While standard separation-of-powers analysis might have predicted that the two branches take up different tasks with respect to the same law, in fact, comparing the two branches’ work reveals a stark substantive separation. In tax, this Article shows, the judiciary is regularly working on one set of Code provisions and the legislature on another. This Article theorizes this split by labeling the judiciary’s sections “the judicial Code” and the legislature’s “the legislative Code.”

This Article’s account of the tax separation of powers is based on the first comparative study of the tax judiciary and tax legislature. The research relies on two, novel, large hand-coded datasets on the workings of the Tax Court and Congress. These datasets examine, among other things, which sections of the Code each branch examines, and how often. While previous scholarship has examined the Tax Court and Congress’s tax activities, no previous work has compared the two. In so doing, this Article breaks new comparative ground.

This Article’s novel approach reveals a finding that is surprising in light of traditional understandings of the separation of powers. The judicial Code and the legislative Code are not just different in function, but different in the topics they cover. The judicial Code includes many of tax law’s foundational statutes, the ones that answer bedrock questions about what counts as income and what counts as deductible expense or loss. The judicial Code also, unsurprisingly, deals with matters of procedure, including those that particularly affect low-income taxpayers. In contrast, the legislative Code contains many narrowly tailored subsidies for specific industries, especially energy and real estate. Congress’s substantive tax work also concerns income support, retirement savings and disaster response.

Why are the judicial and legislative Codes substantively so separate? This Article offers two potential explanations. First, textual differences in language and structure may partially determine which sections are in the judicial group and which are in the legislative. Judicial Code sections, this Article’s analysis reveals, are often structured as standards rather than as rules. In addition, some of the most frequently cited judicial Code sections are, this Article argues, tax versions of what the statutory interpretation literature calls “common law statutes.” As such, they invite active judicial engagement and even policymaking.

Second, the two Codes have distinct political economies. Which actors control the tax agenda plays a major role in slotting Code sections into each of the two buckets. Specifically, the Tax Court has limited ability to control the judicial agenda; its judicial Code is limited to sections that arise in individual disputes between taxpayers and the Service. In contrast, interest groups have substantial control over the legislative agenda. For that reason, well-resourced private parties may be the ones able to determine which Code sections become part of the legislative Code.

This Article makes three major contributions. One, with its original large-scale comparative study of the Tax Court and Congress, this Article provides the first account of the separation of powers in tax law. Two, the Article theorizes and probes the potential reasons for the separation it empirically documents, offering explanations grounded in text and in political economy.

Three, the Article also draws out what its novel account brings to the broader literature on separation of powers. The Article’s account of the tax separation of powers challenges the dominant understanding of separation of powers. As Constitutional scholars Brian Richardson and Joshua Macey recently put it, the Constitution “fashions branches that are ‘omnicompetent’ as regards subject-matter but ‘unipowered’ as regards tools at their disposal.” Separation-of-powers analysis normally starts with that premise as both a matter of Constitutional text and practice. However, the Tax Court and Congress are not, it turns out, acting omnicompetent. For that reason, this Article intervenes in the separation-of-powers literature to show that context matters. Looking at an area of law that the separation-of-powers literature has never considered introduces to that literature a new dimension. In conversation with the existing accounts, the Article also considers the normative issues that the divide between judicial and legislative Codes raises.

In its analysis of the judicial and legislative Codes, this Article proceeds in three parts. Part I describes the research question and design and then presents its findings about the Tax Court and Congress. Part I also situates these findings within the broader literature on those two institutions. Part II then considers the separation between the judicial and legislative Codes. Part II.A explicates the separation, and Part II.B looks at potential explanations for it. Specifically, Part II.B.1 contemplates textual differences between the Codes and Part II.B.2 examines differences in political economy. Part III explores possible implications of the separation between Codes for the separation-of-powers literature: Part III.A as a descriptive matter and Part III.B as a normative one.

II. The Comparative Study

A. Study Question and Design

This Article started with the question of how does the separation of powers work in tax? The extensive literature on separation of powers tells a familiar story: the legislature makes the law and the judiciary interprets it. That story is one in which the different branches have different functions, and it largely comes from the field of Constitutional law. This Article wanted to explore both how well that story describes tax, and how well tax fits within that story.

The inquiry required looking at tax law empirically across institutions, taking an original comparative look at the federal tax judiciary and legislature. Doing that required data-gathering on each. To begin, the study collected and hand-coded data on a large sample (n=775 cases) of decisions of the federal Tax Court between 2018 and 2020. That amounted to 4,036 interpretations of 408 Code provisions. Then, the study looked at Congress. Congress amends tax law much less frequently than Tax Court adjudicates under it, so the study used a longer time frame for Congress, analyzing the full population of tax provisions that Congress amended between 2015 and 2022. That included 47 tax bills and 1,074 changes to 333 Code provisions.

Originating in 1924 as an administrative agency called the Board of Tax Appeals, the Tax Court became an Article I court in 1969. Scholars consider it a well-established court of law for purposes of the Appointments Clause of the U.S. Constitution, and it is a court of record under Article I. The Tax Court shares its jurisdiction with two other federal courts: federal district courts and the U.S. Court of Federal Claims. Perhaps the primary difference between the Tax Court and the other two is that the Tax Court allows taxpayers to bring their case before they actually pay the Services’s proposed tax deficiency. This may be the reason that the substantial bulk of federal tax litigation appears before the the Tax Court, where private parties face the federal government as it is represented by the Service, the repeat player in Tax Court cases. All types of litigants can appear before the Tax Court. These include individuals, corporations, and estates. While different researchers report somewhat different figures, the largest study shows that over 50 percent of taxpayers appear before the Court without counsel and that nearly 90 percent of small tax cases are filed by pro se litigants. Taxpayers can also choose to litigate their federal tax disputes in U.S. federal district court or the U.S. Court of Federal Claims. The current study also reviewed those dockets for the relevant time period, but the number of tax cases in each was relatively small and focused mostly on matters of non-tax civil procedure.

While other research, described below, has examined the Tax Court on its own, no study before this one has compared its work to Congress’s. To do that, the current study examined all the changes made in the bills in which Congress amended a section of the Internal Revenue Code from 2015–2022; I then sorted these amendments by the Code section involved, using the Joint Committee on Taxation’s Technical Explanations. The JCT publishes these “technical explanations” of all tax bills that Congress passes and labels each explanation by both the section of the relevant bill and the Code section(s) affected. The explanations describe prior law in the area and then summarize the changes. I also reviewed the statutes and the explanatory texts themselves.

B. The Judicial Code

1. Findings from the Current Study

Examining the numbered Code sections that the Tax Court interprets in each of its cases provided evidence of the specific substantive topics the court was taking up most often. The top five subjects that the Tax Court adjudicated most often during the 2018-2020 period were what items taxpayers can deduct as business expenses (section 162 plus section 274 and section 280A); what constitutes income (section 61); what counts as a deductible gift to charity (section 170); and what qualifies as a deductible loss (section 165).

During this period, some of the other Code provisions the Tax Court examined most frequently fell under what tax scholars and practitioners would call “tax practice and procedure.” These are “accuracy-based penalties” for incorrectly calculated tax liability, burden of proof in tax disputes, the notice requirements to levy taxes, and the procedural requirements for imposing tax penalties. Each of these procedural sections is about how to handle disputes between a taxpayer and the Service. As a result, it is not surprising that a court tasked with resolving tax disputes spends substantial time on these sections. The goal of this study, however, is to understand what tax issues occupy the Tax Court, so the account below focuses on its frequently interpreted substantive provisions.

To begin, take section 162. The substance of section 162 lies in its basic rule: allowing “as a deduction all the ordinary and necessary expenses paid or incurred during the taxable year in carrying on any trade or business . . . .” The statute goes on, in section 162(a), to make a short list of what it acknowledges is a nonexclusive list of deductible examples. These include salaries, travel, and rent. The statute also carves out a prescribed list of nondeductible items including bribes, lobbying expenses, fines and penalties, costs of reacquiring one’s own stock, and “excessive” employee compensation. The bulk of the statute’s language and the associated regulations set forth the parameters of the various nondeductible items.

Yet, most of the Tax Court cases concern the initial language of section 162, found in section 162(a), plus the short list of deductible items. A representative case is Rivera v. Commissioner. In it, the taxpayers, a lab assistant and a nurse, participated in a multilevel marketing scheme to sell telephone, internet, and related services. Former President Donald Trump and his family had promoted the scheme heavily on his show “The Celebrity Apprentice.” The Riveras deducted the fees to participate in the scheme and attend its conferences. The Service denied these deductions, along with other purported business deductions, and the case went to the Tax Court. The court found that the Riveras had failed to document some of the expenses sufficiently, but, citing another recent Tax Court decision allowing taxpayers to deduct multilevel marketing fees and convention costs, as well as accepting the Riveras’ own “credible” testimony that they “were required to pay such fees regularly in order to participate,” the court permitted the fee deduction.

To take another instance of a typical section 162 business case, Christensen v. Commissioner involved the oft-discussed rules about commuting expenses. This case concerned Dean Lee Christensen, a taxpayer who worked as a part-time professor for Piedmont Community College, “where he taught classes in general education, psychology, technology, and humanities.” At the time, Piedmont had two campuses that were 24.4 miles apart, one in Person, North Carolina, and the other in Caswell, North Carolina. No matter which campus Mr. Christensen was teaching at, he traveled to his office at the Person campus at the start of every workday to retrieve his class rosters. He also returned to the Person campus at the end of every day to lock the class rosters in his office, because he “felt it was important to protect the personal information of his students included on the rosters.” Mr. Christensen tried to deduct the cost of round-trip mileage of 48.4 miles for each day.

Under the relevant case law, employee expenses for travel from home to work are generally not deductible. The Service argued that Mr. Christensen had been commuting from his home to his two teaching positions for Piedmont. Mr. Christensen contended that he was not “commuting from his home but traveling from one teaching location to the other, which is not commuting.” By his account, he kept student records with personal information locked up at the Person campus, so he had to always travel there first to pick up the records before going to Caswell. However, the court held that because Mr. Christensen “was not required by Piedmont to lock up the student records at one of the campuses or otherwise to stop at one of the campuses before traveling to the other and it was his personal decision to do so . . . ,” the costs of travel between the campuses was commuting, and therefore not deductible.

Another common issue on the Tax Court’s docket concerns what items taxpayers must include in the category of income. On first blush, this looks like an easy question. Even a first-month tax student can state the rule found in the section 61(a) or in the foundational Glenshaw Glass case: gross income includes “all income from whatever source derived.” After going through a prescribed set of examples set forth in landmark cases from decades ago, which provide instances of items included in income, students may come to believe that what to include in income is not, in fact, a legally difficult question. However, the Tax Court regularly takes up cases on the topic. These cases span several sub-issues, including what counts as an excludable gift, what qualifies as a forgivable debt, and others.

For instance, in Novoselsky v. Commissioner, Chicago attorney David Novoselsky found ways to secure private financing for some of his class actions through litigation support agreements. The sources of the funds were plaintiffs in his cases, persons whose “interests were economically aligned with the interests of such plaintiffs,” lawyers with whom he had fee-sharing arrangements, or “individuals seeking a high return on a speculative investment.” Mr. Novoselsky was not obligated to repay the funds unless the litigation in question resulted in attorney’s fees and/or costs. Under these agreements Mr. Novoselsky received payments totaling $410,000 in 2009 and $1 million in 2011, which he said were loans. Courts take the position that, because loans come with a “genuine obligation to repay,” they do not constitute gross income. However, as the Tax Court noted in this case, “[w]here an obligation to pay arises only upon the occurrence of a future event, we have consistently held that a valid debt does not exist for Federal tax purposes.” Here, because Mr. Novoselsky “did not have an ‘unconditional obligation to pay a definite sum of money,’” the litigation support agreements were includible as income, and were not excludable from income as loans.

To take another example of a typical section 61 case, in Duncan v. Commissioner, Dann Duncan was an attorney specializing in workman’s compensation claims for members of police and fire departments. He successfully represented a police officer who won $360,000 in such a claim. The client had paid a $2,500 flat fee at the beginning of the litigation, and then, after the victory, sent Mr. Duncan a check for $30,000. Mr. Duncan did not include the latter amount in income on the grounds that it was nontaxable as a gift. The Service disagreed. Citing the landmark Duberstein case, the Tax Court looked to the client’s intent in making the payment, writing that “a gift made out of generosity or charity is not taxable . . . but one made out of gratitude, particularly gratitude stemming from a prior economic or commercial relationship, is.” Here, the Tax Court found that the client appeared to have intended the gift “as compensation for a job well done (akin to a tip or gratuity for exceptional services rendered) and much more directly so than in many cases in which we have similarly held.” As a result, Mr. Duncan had to include the $30,000 in income.

Another common substantive topic that the Tax Court took up in the studied period come as a pair of Code sections, section 274, which covers deductible entertainment expenses, and section 280A, which addresses business deductions specifically for personal and vacation homes. As discussed above, taxpayers may deduct expenses associated with a trade or business. Section 274 limits the extent to which taxpayers may deduct those business items if they relate to “activities that are generally considered to be entertainment, amusement, or recreation.” Unlike section 162, the foundational business-expense statute, section 274 is long, containing relatively detailed rules about particular activities that might fall into the categories of entertainment, amusement, or recreation. Its corresponding regulations are also hefty.

Becnel v. Commissioner is typical of Tax Court section 274 analysis. Damon Becnel was a “serial entrepreneur originally from Louisiana who moved east to Destin, Florida, dubbed by its boosters the ‘World’s Luckiest Fishing Village,’” from which he launched a ‘real estate empire. In December 2005, Becnel paid just over $2 million for the Britney Jean, a fishing yacht. The Tax Court wrote excitedly that, “[w]e find the boat a beautiful specimen of its species: It’s 67 feet long with four staterooms; a well-appointed salon, galley, and bar; a tuna tower; and a slew of electronic gadgets and built-in fishing equipment.” The Tax Court also found that the “Britney Jean is a fishing boat, and Becnel used her from day one almost exclusively for that purpose.” In fact, “[i]n 2006 she took part in ESPN 2’s “Billfish Xtreme Release League, a televised release-only fishing league at destination resorts,” and, after that, “in about four regional fishing tournaments a year between 2009 and 2011.” Mr. Becnel argued that the Britney Jean may have looked like a recreational fishing boat but that is was, in reality, “a marketing tool for his real-property businesses.” For that reason, he deducted the costs of the Britney Jean as business expenses. The Service denied the deductions.

In its reasoning, the Tax Court applied section 274 to find that the costs of the Britney Jean were prohibited entertainment expenses under multiple prongs of the statute’s scheme. For one, the section 274 regulations specifically ban “any activity which is of a type generally considered to constitute entertainment, amusement, or recreation, such as entertaining on fishing trips” According to the court, “that’s basically all the Britney Jean was used for.” Then, the yacht itself was an entertainment “facility,” the costs of which, unlike those of entertainment “activities,” are not deductible at all. Further, section 274 itself requires taxpayers to show that any deductible business-entertainment expenses are “directly related to the active conduct of [a] business,” and the section imposes specific documentation requirements.” Playfully, the Tax Court wrote that “Becnel runs aground on both: Other than his testimony, which is not credible on this point, there’s no evidence of a proximate relationship between the yacht-related expenses and the active conduct of any of his businesses.” As a result, the Court denied the yacht deductions.

Another frequently adjudicated substantive provision is section 170. This provision governs deductible contributions to charity. Most of the section 170 cases in the sample dealt with a single type of tax transaction: the conservation easement donation. As tax trade publication Tax Notes has explained the deals:

You buy a piece of undeveloped land for, say, $1 million [and] sell partnership shares in it for $4 million to various folks wanting a large tax deduction. Then you donate a conservation easement on the land and tell the IRS that the easement is worth $16 million. Net result: Every investor gets a $4 charitable tax deduction for each $1 they paid, saving them roughly $1.50 in taxes.

While conservation easements have been on the IRS’s radar screen for decades, late in 2016, the IRS officially categorized them as “listed transactions,” meaning that taxpayers who structured deals with these easements needed to report that to the IRS. That change resulted in “hundreds of transactions being reported to the IRS, quickly followed by hundreds of Tax Court cases, relatively few of which have reached a decision.” At the 2023 nomination hearings of a proposed Service chief counsel, Senator Bill Cassidy observed; “It is rumored that the IRS counsel’s office has a directive to never settle and instead litigate 100 percent of conservation easement cases. If this is true, conservation easement cases have the potential to completely overwhelm the Tax Court for the next decade, with a reported volume of cases upwards of 1,000.” Even if that estimate is excessive, conservation easement cases do heavily populate the Tax Court docket.

While taxpayers generally may not take charitable deductions for donating partial interests in property, that rule does not apply to “qualified conservation contributions” of a qualified real property interest to qualified organizations. for “conservation purposes.” Treasury regulations guide how to value conservation easements for charitable deduction purposes, and preferential rules apply to determine the size of the deduction. As will be discussed below, during the studied period, Congress was also involved in changing these rules.

In court, conservation easement cases generally present one or both of the following issues: donor intent and valuation. For instance, in Wendell Falls Development, LLC v. Commissioner, a real estate developer did a paradigmatic conservation easement transaction. Courts have consistently held that a charitable contribution deduction is not available “where a taxpayer intends to obtain a direct or indirect benefit in the form of enhancement in the value or utility of the taxpayer’s remaining land or otherwise to benefit the taxpayer.” The Service denied the charitable contribution on the grounds that Wendell Falls intended exactly this. Additionally, the Service argued that the conservation easement that Wendell Falls had donated to a land-trust charity was not worth the several million dollars that Wendall Falls claimed, but was in fact appropriately valued at zero. The Tax Court observed that giving the easement to the land trust ensured that surrounding acres would become a park. The park would be fortuitously located next to one of Wendell Falls’s residential developments. For that reason, the Tax Court held that Wendell Falls could expect a benefit from contributing the easement. Further, the court found that the “highest and best use” of the surrounding land was as a park, with or without the easement. If the easement made no difference to the value of the surrounding land, it could not be worth anything, let alone the millions of dollars that Wendell Falls had deducted. The Tax Court denied the deduction.

Not all of the Tax Court’s section 170 cases in my sample were about conservation easements. Section 170 has a variety of complex requirements that the Tax Court interprets. For instance, in Chrem v. Commissioner, Mark and Esther Chrem owned stock in a Hong Kong corporation that a related company had proposed to buy. After the Chrems sold some of their shares to that company, they donated the rest to a charitable organization, valuing it at $4,500/share. The related company then purchased the shares back from the charity for that price. Among other things, the Service claimed that the Chrems had failed to obtain “qualified appraisals” of the stock under sections 170(f)(11)(C) and (D). On a motion for summary judgment, the Court considered the Services’s arguments that the appraisal the Chrems secured was not addressed to them, did not include a date, and did not state that it was “prepared for income tax purposes” as the regulations require In addition, the appraisal did not value the specific stock that the Chrems contributed; rather, the appraisal valued the whole corporation. The Chrems claimed that “this [was] a case of ‘no harm, no foul’” because the buyer had agreed to the price and the shares were no less valuable than any other shares in the company. For those reasons, the Chrems maintained that they did substantially comply with the statutory appraisal mandate. Further, the statute excuses taxpayers from complying with the mandate if the failure “is due to reasonable cause.” The Chrems argued that their experienced accountant reviewed the appraisal and told them they did not even need to attach it to their tax returns, a circumstance that was a reasonable cause for failing to comply with the appraisal rules. The Tax Court held that these questions presented issues of material fact and denied the Services’s motion for summary judgment.

The substantive topic the Tax Court adjudicated next most frequently came from section 165, which allows taxpayers to deduct “any loss sustained during the taxable year and not compensated for by insurance or otherwise.” The statute then lists various types of losses that would fall into this category, one of which is loss from “theft.” Typical of a section 165 case is Littlejohn v. Commissioner. Here, Charles Littlejohn, a lawyer, and his wife Maxine bought 85.5 acres off Prefumo Canyon Road in San Luis Obispo, California. The property included a four-story house in the style of a French chateau, which had six bathrooms, a pool, a spa, a wine cellar, and a detached four-car garage tiled with travertine marble [and] serving as a sundeck. After the Littlejohns bought the house, the garage ceiling collapsed. This began a long dispute with the sellers of the house and their contractor (hired to effect repairs), which eventually wound up in civil court. The contractor, Mr. King, did not respond to the civil complaint and a California superior court awarded Mrs. Littlejohn $150,000 as a default judgment. The judgment stated that Mr. King “was a willing member of a conspiracy” by “engaging in a fraud involving an omission and/or cover-up of material defects of the garage and travertine deck.”

The Littlejohns tried to deduct some of the property’s value as a theft loss, because “they were defrauded in the purchase of the Prefumo property” in a variety of ways. However, the Tax Court denied the deduction on the grounds that the Littlejohns had “not cited, nor have we found, any case holding there to be a crime of theft” under California state law “based on an alleged failure to disclose defects in the course of a commercial real estate transaction.” Further, observed the court, California law defines a theft by false pretenses as requiring that “any [relevant] misrepresentations were made ‘knowingly and designedly.’” After extensive review of the factual record, the Tax Court found that the Littlejohns could not show that anyone “made misrepresentations” either affirmatively or by failing to disclose defects. Although the Littlejohns tried to use language from their default judgment as evidence of misrepresentations, the Tax Court wrote that the judgment was not “probative evidence as to any crime involving fraud,” and, even if it was evidence of fraud, it was only about the garage, the cost of which they had already recovered through insurance.

As described above, the judicial Code also includes sections about the Tax Court and administrative procedure. This Article generally does not discuss those sections, because it adds little to the analysis to note that the judiciary frequently takes up the topic of its own rules. However, one of these procedural sections does reveal a relatively unexpected Tax Court focus: the administrative rules for taxpayers experiencing financial hardship. This is section 6320. Under this section, the IRS must give written notice of the filing of a federal tax lien to the person whose property is subject to the lien, not more than five days after the agency files. That notice of federal tax lien must include, among other requirements, notice of the right to a Collection Due Process (“CDP”) hearing. Certain rules apply to a CDP hearing, among them a requirement that IRS verify that certain procedures have been followed, rules regarding issues that the parties can raise and rules regarding how the Service determines the amount collected. The Service issues a “notice of determination” of tax liability after a CDP hearing, which the taxpayer may appeal to the Tax Court.

CDP cases often involve pro se taxpayers or taxpayers who otherwise lack resources. A paradigmatic case is Loveland v. Commissioner. James Loveland was a retired boilermaker and Tina Loveland a teacher. During a short period of time, the Lovelands lost their home to foreclosure, Mr. Loveland developed a heart condition, and Mrs. Loveland was diagnosed with breast cancer. During what the Tax Court agreed was a “tumultuous time,” the Lovelands stopped paying their taxes. The Service thereupon began proceedings to levy the Lovelands’ income, and the couple, citing insufficient funds for the full amount, proposed to the collections officer a partial payment (an “offer-in-compromise”), which the officer rejected. The Lovelands then tried to work out an installment plan with the Service; as negotiations on this moved along, they also tried to take out a second mortgage on their house to pay off some of the tax debt. However, as part of the collection effort, the Service placed a lien on the house, preventing the Lovelands from getting the second mortgage. At this point, the Lovelands appealed the lien via the CDP process and again proposed an offer-in-compromise as well as a revised installment plan. The Service’s internal appeals officer declined to consider either proposal and the case went to Tax Court.

Siding with the Lovelands in full, the court first ruled that the Service had abused its discretion when the appeals office did not consider either the offer-in-compromise or the installment plan. The court acknowledged that the CDP statutory framework precludes the appeals office from looking at a taxpayer proposal if “the issue was raised and considered at a previous hearing . . . or in any other previous administrative or judicial proceeding.” The question before the court was, did the Lovelands’ conversation with the collections officer when they first raised the offer-in-compromise count as an “administrative proceeding”? Looking at statutory and regulatory context, the court found in the negative. As a result, the court held that the I.R.S. Independent Office of Appeals (I.R.S. Appeals) should in fact have reviewed the offer-in-compromise. Similarly, while the Tax Court had held in other cases that I.R.S. Appeals does not need to consider installment plan proposals if the taxpayers do not provide enough financial information, the court here found that the disclosures the Lovelands made were sufficient to have allowed the appeals office to consider the proposed installment plan. In addition, the court found that I.R.S. Appeals should explicitly have taken into account the Lovelands’ economic hardships.

The remainder of Tax Courts’ most commonly adjudicated substantive provisions of the Code are relatively similar to the ones described so far. Some are procedural, such as section 6213 (Tax Court petitions), section 7433 (unauthorized collection actions), section 6015 (relief for those unaware of their spouses’ dubious tax behavior), section 6212 (IRS notice requirements), section 7623 (whistleblowers among a few other things), section 6511 (limits on claiming credit for tax overpayments), section 6331 (more on levies and collections), section 6501 (statutes of limitations on collections), section 7421 (bans on suits to stop collections), section 6663 (civil fraud penalties), section 6103 (confidentiality of tax information), section 6673 (assignment of court costs), section 6402 (use of tax overpayments), section 6323 (priority of IRS liens), section 6654 (taxpayer failure to pay estimated taxes), section 6321 (IRS liens generally), section 7122 (amounts in compromise), section 6020 (dummy returns (such cases appearing 14 times in the dataset)), section 6532 (statutes of limitations), section 6672 (failure to pay taxes (such cases appearing 11 times in the dataset)), and section 6231 (notice in partnership proceedings). Some are largely corollaries to section 162 and provide additional statutory rules relating to the business-expense issue; these include section 262 (ban on deducting personal expenses as business ones), section 212 (deductible expenses for income-producing activities), section 469 (prohibition on some deductions for investment losses), and section 280A (limits on deducting personal property even if used in a business). The rest of the provisions appearing in 20 or more cases were section 72 (annuities and life insurance), section 1 (tax brackets), section 172 (business loss deductions), section 1401 (self-employment tax), section 163 (interest deductions), section 7701 (business-tax definitions), section 6214 (Tax Court jurisdiction), section 1366 (treatment of S corporation income), section 501 (tax-exempt organizations), section 67 (miscellaneous itemized deductions), section 446 (methods of accounting), section 152 (definition of dependents) and section 62 (definition of adjusted gross income).

Some further information from the dataset adds context about who is in Tax Court litigating the judicial Code. Many taxpayers appeared pro se (372/775) or were represented by small private firms (141/775). Unsurprisingly, the cases where the lawyer came from a large private firm (73/775) were those with significant amounts in controversy. The overwhelming majority (620/775) of taxpayers were individuals, with the rest businesses (119/775), and the government won more often than the taxpayer did (515/775).

2. The Tax Court Generally

The results of the current study are roughly in line with other research on the Tax Court. The number of empirical studies of the Tax Court is small. Most of those endeavors gather data on the Tax Court aimed at answering a particular research question. Only one other, a large quantitative study by political scientists Michael Bonnarito and Daniel Katz, along with accountant Jillian Isaacs-See, breaks down Tax Court activity by Code section. Its findings mirror this Article’s. Bommarito, Katz, and Isaacs-See also find that “the Tax Court pays a disproportionate amount of attention to [only] a small number of sections of the Code.” They find, as the current study does, that “[w]hile some of these citations are procedural, many are not[,]” and “[t]herefore, an analysis of Tax Court jurisprudence should likewise focus on these important sections.” Importantly the “list of the most-cited sections is striking for what it does not include—no sections addressing capital gains or losses, no sections addressing tax shelter reporting or complex transactions, and no sections addressing foreign investment.” Instead, “[t]he sections that are cited most frequently deal with the distinction between personal and business expenses, with family relationships, and with forms of income that are earned by individuals.” Further, the Bommarito team observes that “[g]iven the size of the Code, the percentage of Code sections that are cited often by the Tax Court is extremely small; a small number of sections are cited very often, and the network of citations is structured around a few central sections.” Notably, that becomes more true in their data over time.

The researchers draw two conclusions from this evidence. First, they determine that any further “analysis of Tax Court jurisprudence should . . . focus on these important sections.” Second, the researchers speculate that the Tax Court’s emphasis on a handful of key sections “could mean that the Code is effectively broken into two tax codes.” In other words, “a separation between the content in the Tax Court’s written opinions and the focus of tax legislation.” However, Bommarito, Katz, and Isaacs-See stop before actually examining Congress to find out which parts of the Code receive legislative attention.

Another study that analyzed the Tax Court docket substantively is legal scholar Bill Whitford’s probe of the Tax Court’s small-case procedure. “Small tax cases,” which are also included in my sample, are handled under simpler, less formal procedures than regular cases. Only certain disputes are eligible to be filed as small tax cases, and generally, the amount in controversy cannot exceed $50,000. Whitford’s data classify the cases by issue area rather than Code section. He finds that the small-case docket from 1979–82 consisted of, most commonly, business-expense deductions (44 percent of the docket), losses and bad-debt deductions (10 percent of the docket), income inclusions (8 percent of the docket), the dependency exemption (9 percent of the docket) and medical-expense and taxes-paid deductions (6 percent of the docket).

Other studies of Tax Court behavior use empirical data on the court to answer questions besides those dealing with the substantive makeup of its docket. For instance, in her foundational 1999 article, tax professor Leandra Lederman looks at the factors that make Tax Court cases likely to settle. Among other things, she observes that “cases that had been through the Appeals Office prior to being docketed in Tax Court were about four times more likely to go to trial than cases that had not been to the Appeals Office prior to docketing.” In addition, she finds that “higher stakes cases were much more likely to go to trial” and also that “taxpayers disproportionately win higher stakes case.” Professor Lederman’s other landmark Tax Court study, with Warren Hrung, concerns the effect of lawyers on taxpayer results. Here, she finds that attorneys obtain significantly better results in cases that go to trial than unrepresented taxpayers do. The magnitude of the effect goes up as attorney experience increases. However, she also finds that in Tax Court, attorneys do not obtain better outcomes in settled cases.

Earlier work on the Tax Court took up the question of potential government bias in that arena. Comparing Tax Court to federal district court, tax scholar Deborah Geier found, looking at a period in the 1980s, that “the government won or partially won an average of 70.5% of district court cases in the years indicated, and no particular trend in this figure is discernible” during the time period. In contrast, “the percentage of cases won or partially won by the government in the Tax Court averaged 90.4% for the same period—a full twenty-point difference—evidencing a decided pro-government trend in recent years.” From this, Lederman concludes that Tax Court judges lack independence from the executive branch. For this reason, she says, Congress should provide Tax Court judges with Article III tenure and salary guarantees. Using a different sample of cases from different years, however, tax professor James Maule finds the opposite: that “the Tax Court, as an institution, is not biased in favor of the IRS,” and “[if the data] demonstrate any bias, they support the contention of former Chief Counsel and Justice Department attorneys that the Tax Court is biased in favor of taxpayers.”

Following these broad inquiries about whether Tax Court bias exists, legal scholar Daniel Schneider asked whether some judges might display more bias than others. In his 2001 study, he analyzed judge characteristics to find that Tax Court judges who were female, were Democratic appointees, had a shorter tenure, were educated at less-elite law schools, or were not white could be associated with the taxpayer winning. Judges who came from private practice also could be associated with the taxpayer winning. In a separate study, he examined variation in approaches to judging on the court and found that “an elite college education for a district court judge was associated with his reliance on practical reasoning and his failure to defer to regulations” or other Service guidance.

Also on the topic of the Tax Court’s relationship to the Service, tax scholar Charles Borek examined cases in which a taxpayer alleges that the Service has “abused its discretion in some ways. Borek founds that “institutional (e.g., ‘corporate’) taxpayers regularly enjoy a diminished deferential standard” whereas, “individual taxpayers in collection cases are faced with a Tax Court willing to accord great deference” to the Service. More recently, also on the topic of judge attributes, accounting professors Bradley Lindsay, Sophie McDonnell and William Moser collected a sample of Tax Court cases indicating that “if the Tax Court judge presiding over the case was appointed by a Republican president, then the case is more likely to be resolved through a negotiated settlement.” In addition, in their study Tax Court judges with longer tenure on the bench were more likely to preside over a tax dispute resolved through a negotiated settlement.

While all of these studies illuminate key aspects of Tax Court behavior, only the Bonnarito, Katz, and Isaacs-See study (discussed above) takes up any substantive inquiry into what the tax judicial branch does. Because of this shortcoming, past studies leave unanswered the question of how separation of powers operates in the tax context. That is the gap into which this Article steps. To proceed in that inquiry, this Article next turns to what the tax legislative branch does.

C. The Legislative Code

1. Findings from the Current Study

During 2015–2022, Congress enacted tax legislation via 47 bills, all of which I examined to identify and analyze the substantive issues involved. During this period, the most legislated Code section was section 401, which covers private retirement accounts. Almost all of the changes to it were in 2023 legislation about retirement savings. In that bill, for instance, Congress required certain retirement plans to provide for automatic enrollment, allowed taxpayers to take a “saver’s credit” for amounts contributed to certain retirement plans, permitted employee payments of student loan debt to qualify for a matching payment from an employer into a retirement account, and increased the age by which certain tax-advantaged retirement plans have to start paying benefits. Congress also raised the limits for certain “catch-up payments” from employers to retirement plans and put in place new incentives for employee contributions.

The next-most most legislated provision was section 168, which covers depreciation. Businesses that have assets may deduct part of the cost of the asset over a set number of years, called either the “useful life” or the “recovery period.” These deductions are valuable to business taxpayers, whose goal is generally to deduct as much of the value of the asset as quicky as possible. Different types of property have different “class lives,” and the class life, in turn, determines the recovery period. Congress regularly amends section 168. For instance, before 2015, owners of fruit and nut plants had to deduct the costs of those plants evenly over seven or ten years, depending on whether the plant was a tree or vine, or something else In 2015, Congress changed the law to allow owners of “certain plants bearing fruits and nuts” to elect to deduct 50 percent of their costs up front, and then 10 percent for each of the following years. To take another example from the same bill, while the baseline statutory recovery period for nonresidential real property is 39 years, in 2015, Congress extended a temporary change to that statute to give the owners of “motorsports entertainment complexes” the shorter period of seven years. In another instance, in 2020 Congress allowed taxpayers to deduct “[q]ualified improvement property,” which is the interior portion of nonresidential real property, over 15 or 20 years instead of the 39 years that the IRS had been requiring.

Congress took legislative action an equal number of times in regard to section 72. This section is the tax preference for annuities, which are contracts, generally from insurance companies, that pay out a certain amount of money every year as long as the contract holder or “annuitant” is alive. Under these contracts, the portion of each payment received as determined under a statutory ratio is excluded from gross income until the entire investment has been recovered. Congress altered this provision often in the studied period. For instance, section 72 imposes a penalty on early withdrawals from an annuity; but in 2023 Congress exempted distributions made to corrections officers or forensic service employees both for long-term care costs and related to excess IRA contributions. As part of its early response to the COVID-19 pandemic, Congress also exempted from the early withdrawal tax any “coronavirus-related distribution” from qualified retirement plans and annuities, and, the year before, Congress did the same for distributions made following certain natural disasters. In 2015, the exemption from the early withdrawal tax was for federal law enforcement officers, firefighters, and air traffic controllers, and in 2016, it was for nuclear materials couriers, the US Capitol Police, the Supreme Court Police, and diplomatic security special agents.

After section 168 and section 72, the next-most legislated section (15 times) was section 24, the child tax credit, which is a credit for taxpayers with children, and one of the hallmarks of the federal government’s anti-poverty toolkit. The maximum child credit amount has increased over the period that studied, rising from $1,000 per child prior to 2018 to $2,000 per child from 2018 through 2020. The amount of child credit that taxpayers can receive as a tax refund is generally limited based on the amount by which the taxpayer’s income exceeds $2,500. As a result, “children of nonworking taxpayers receive no benefit from the credit and children of low-earning taxpayers receive only a partial benefit from the credit.” In addition, since 2017, the refundable portion of the credit has been capped at $1,400 per child, further limiting the potential benefits of the credit for the children of low-income taxpayers.

Congress’s activity around the child credit was perhaps its highest-profile tax project in the studied period. In March 2021, Congress passed the American Rescue Plan Act in response to the continuing COVID-19 pandemic. In that act, Congress temporarily increased the credit to $3,600 for children up to age 5 and $3,000 for children ages 6-17, and it expanded the credit to reach nonworking families with children for 2021. For that year, Congress also temporarily removed the limits on the credit’s refundability, turned the credit payable in advance of filing a tax return, and introduced some other provisions aimed at making the credit more widely available during the pandemic. In addition to the pandemic-related credit widening, Congress legislated section 24 several more times during the studied period. For instance, the 2017 Tax Cuts and Jobs Act temporarily increased the child tax credit to $2,000 per qualifying child and provided a $500 nonrefundable credit for non-child dependents. The bill also required that, in order to receive the credit for a qualifying child, a taxpayer must include the Social Security number of the child on the tax return claiming the credit. Congress also lowered the earned-income threshold for taking the credit to $2,000/year, having permanently lowered it to $3,000/year in the 2015 PATH Act.

The next most frequently legislated substantive provision during the studied period was the earned income tax credit, best known by its acronym the “EITC” (13 times). The EITC is among the largest anti-poverty programs in the United States today, allocating a refundable credit to low- and moderate-income taxpayers who have positive earnings from employment or other work. The credit amount depends on family size and income, with maximum benefits ranging from $519 for taxpayers without children to $6,431 for taxpayers with three or more children. The average EITC amount is currently approximately $2,500. Many of the child credit changes during the studied period had corresponding revisions in the EITC. For instance, the 2021 pandemic response bill temporarily reduced the minimum qualifying age for a taxpayer with no children from 25 to 19 in most cases, and it reduced the maximum qualifying age altogether. That bill also increased the percentages and maximum amounts of earned income taken into account to calculate the credit, as well as the maximum amounts of the credit itself.

Following the EITC, the next most common Code section on which Congress took action was the credit for production of renewable energy, section 45. This credit allows energy companies to take a credit of 1.5 cents (adjusted for inflation) for each kilowatt hour of electricity that the taxpayer sells to an unrelated person and produces at a qualified facility from wind, “closed-loop biomass,” “open-loop biomass,” geothermal energy, solar energy, small irrigation power, municipal solid waste, hydropower, or marine and hydrokinetic renewable energy. Different timing rules apply to different types of facilities. The credit is reduced under a statutory formula that decreases the credit in any year in which the “reference price” for electricity exceeds 8 cents (indexed for inflation) per kilowatt hour. The credit also applies to companies that produce refined coal, at an amount of $4.375 (indexed for inflation) per ton of refined coal sold by the taxpayer to an unrelated person. and is subject to phaseout as the reference price rises above a threshold level.

In the studied period, Congress revised and expanded the time horizon of various pieces of this statutory scheme. For instance, in 2021 as part of a reconciliation package, Congress pushed the credit’s expiration date out three years. Congress also increased the credit amount by 10 percent for power plants built out of steel, iron and manufactured components made in the U.S. Also as part of the Taxpayer Certainty and Disaster Relief Act of 2020, Congress created a special rule for property used in offshore wind facilities, allowing them a 30 percent investment credit, not subject to the normal schedule of cuts applied to wind facilities.” The Bipartisan Budget Act of 2018 allowed companies who qualified for the renewable-energy production credit to elect to take the Code’s general investment credit instead, if the investment credit amount turned out to be higher.

Following section 45 in number of Congressional amendments was the general business energy credit, section 48. Section 48 provides a credit for businesses that invest in “energy property.” “Energy property” includes certain solar and geothermal, fuel cell, microturbine, heat and power system combination, wind, ground water, waste energy recovery, energy storage, biogas and microgrid-control equipment. Congress has been active in revising this provision. For instance, in 2016, Congress extended an increased credit percentage for property using solar energy for heating and cooling. Then, in the 2022 reconciliation bill, Congress again extended expiring parts of the credit, with a particularly long extension for geothermal heat pump property. Congress also added a credit for qualified biogas property that includes, captures, or cleans gas that has a certain percentage of methane. Congress also put in place a credit for microgrid controllers, which control electrical systems of a statutorily defined size and operate within the electrical grid, widened the definition of fiber optic solar property to include electrochromic glass, and increased the allowable credit amounts for companies with domestically manufactured parts.

The next most often legislated section was the only section of the Code that appears frequently in both judicial and legislative Codes, section 170. Here, Congress took up precisely the same issue as the Tax Court: namely, charitable gifts of conservation easements. At the beginning of the studied period, Congress loosened the rules governing these donations, in fact making permanent some of the provisions that enabled the related deductions. After that, Congress attempted to tighten up the rules for these deals statutorily. Recall that one of the provisions allowing taxpayers to deduct easements is the one that treated them as “qualified conservation contributions.” In 2023, Congress took away that treatment in many instances. One of the instances where Congress still allowed easements to qualify was for them to be “certified historic structures.” However, Congress also imposed strict requirements on how to show the easement pertains to a “certified historic structure.” Congress made a few other changes to section 170 during the studied period, including increasing the deduction for gifts of food inventory and allowing non-itemizing taxpayers during the pandemic to take a small credit.

Congress amended the Code several hundred more times during the 2015-2022 period. Receiving approximately 10 changes each were section 423 (multiemployer plans), section 6426 (fuel credit), section 42 (low-income housing credit), and section 857 (real estate investment trusts). Then, the next level down, with 7–8 changes, were section 430 (single-employer defined-benefit pension plans), section 401 (pension plans more generally), section 6103 (tax return confidentiality), section 38 (general business credit), section 414 (definitions relating to pension plans), section 179D (energy efficient commercial buildings deduction), section 25C (energy efficient home improvement credit), section 403 (employee annuities), section 9502 (airport and airway trust fund), section 408 (individual retirement accounts), and 4081 (fuel excise tax). With 5-6 changes each were section 1445 (real property transfer tax), section 30C (alternative fuel vehicle refueling credit), section 529 (savings plans for tuition programs), section 3111 (Social Security tax rates), section 142 (tax-exempt facility bonds), section 4271, section 199, and section 6427.

2. Tax Legislation Generally

While the study in this Article provides a novel comparative view of Congress’s substantive tax activities over a period of time, other scholarly accounts of tax legislating offer context about how the process works and how different parties operate within it.

According to the Constitution, “[a]ll Bills for raising Revenue shall originate in the House of Representatives; but the Senate may propose or concur with Amendments as on other Bills.” As a result, tax bills originate in the House. As the main House tax-writing committee, House Ways and Means Committee begins the formal process by taking up proposed tax legislation, but ideas can arise from “other sources, including the Treasury Department, the Administration, the IRS, or outside interests (e.g., industry and trade associations) and their lobbyists.” The House Committee on Ways and Means membership consists of majority and minority members, and legislative hearings there usually start the process of considering legislation. After hearings, the Ways and Means Committee revises the legislation at a “markup” session, in which the committee “reach[es] tentative decisions on specific issues.”

Once the Ways and Means Committee reports out a bill, the full House considers and votes on it and, if the vote is successful, the bill moves to the Senate. The Senate also considers tax bills through a committee, here, the Senate Finance Committee. Like Ways and Means, the Senate Finance Committee conducts hearings, holds markup sessions, and may make amendments. The full Senate then debates the bill, and if the version adopted is different than the one that the House had reported out, the bill goes to back to the House and a Conference Committee with appointees from both the House and the Senate, reconciles the two drafts. The reconciled Conference Committee version of the bill is then reported back to House and Senate, together with a Conference Committee report that explains the committee decisions. If both Houses approve the Conference Committee version, the president then may sign the bill.

Research has documented the relative ease with which certain types of tax legislation pass. The House Ways & Means Committee is relatively effective at getting tax legislation through Congress, in part because the Committee has traditionally had powerful leadership and cultivated credibility. In addition, many tax bills enjoy certain formal procedural protections, and the Ways & Means Committee, as compared to other House committees, has operated relatively free from the control of a small handful of powerful interest groups; while multiple interest groups are generally able to give input to Ways & Means, no particular subset has tended to dominate.

Even so, Oei and Osofsky have written, the tax legislative process “has moved from an expertise-driven model to one with more amorphous control, with more lobbyist and industry group participation.” Interest groups are able to tinker with a Code that can otherwise be difficult to change. Oei and Osofsky find that “the nature of the tax legislative process means that agreement can often be reached to make small tweaks, but that a broader reformulation (even if the purpose was just clarification) would require more extensive legislative buy-in, which is costly and more difficult to obtain.” Oei and Osofsky document a “statutory inertia” in which items already in the Code are hard to change, quoting an interviewee who said, “[t]here’s a lot of inertia . . . in the tax world and . . . in Congress in general. . . . [I]t takes a lot to overcome that. And it usually ends up not happening.”

Along similar lines, political scientists Alexander Hertel-Fernandez and Theda Skocpol, tracking the share of Congressional tax hearings that refer to small business, have described the increasing role of small business in the tax legislative process and “the tightening linkage between small businesses and taxes” Another political scientist, Amy McKay, has identified the phenomenon of “microlegislation,” using as her case study the Affordable Care Act, which was structured as a tax bill. She finds that “legislative favors to special interests are intentionally hidden in low-visibility microlegislation” that lobbyists will try to get passed. In fact, she observes, in many cases that the language in the amendment is the same as language as in the group’s written comment.” McKay’s data show that the predicted probability that a group is successful in getting its requested microlegislation introduced in a senator’s amendment is more than four times greater for groups that contribute to committee members relative to groups that lobby the same legislators on the same issue at the same time but do not contribute. Along similar lines, in the 1980s, journalists Donald Bartlett and James Steele combed the 880-page 1986 Tax Reform Act and identified about 650 pieces of microlegislation for specific firms or individuals.

Most recently, economists Brian Kelleher Richter, Krislert Samphantharak and Jeffrey Timmons looked at the relationship between companies’ lobbying and their effective tax rates. These researchers estimate that firms that increase their lobbying expenditures by 1% in one year reduce their effective tax rates in the range of 0.5 to 1.6 percentage points the following year. According to the researchers’ data, once the initial lobbying investment is taken into account, for each additional $1 spent on lobbying the average company received somewhere in the range of $6 to $20 of tax benefits, which makes “lobbying look astonishingly profitable. These scholars tell the following story about tax lobbying, which they call the “World Series of lobbying.”

One recent piece of legislation that provides tax benefits to a specific firm can be found in Section 704 of the “Jobs Creation Act of 2004.” The law allows for accelerated tax depreciation of specific types of construction expenses on any “motorsport entertainment complex” placed in service after October 22, 2004, and before December 31, 2007. The clause could in principle apply to all three of the major firms that own and operate NASCAR facilities—International SpeedwayCorp. (ISC), SpeedwayMotorsports Inc., and DoverMotorsports Inc.; however, given the limited three-year window and the large site planning and municipal approval lead time necessary, it is unlikely that the law could provide substantial benefits to any firm that was not already planning on construction projects. Only ISC touted major construction plans during the 2005–2007 time frame in their 2004 Annual Report. ISC was also the only firm that lobbied in the appropriate time frame: the company started lobbying two years prior to the enactment of the favorable legislation, increasing its initial lobbying expenditure of $180,000 in 2003 to $200,000 in 2004, before beginning to receive tax benefits in 2005.14 ISC hired, as its sole registered lobbyist, Williams and Jensen. The lobbying firm, according to its web site, has “the primary mission of advancing the tax policy interests of clients” and claims to have a “results-oriented approach, proven by outcomes” including “creating new tax code provisions to help finance a client’s project” by “securing special effective dates and exemptions when Congress adopts tax law changes.” Given the timing of the legislation and its narrow application, Williams and Jensen appears to have secured a tax break for ISC in this instance.

III. Separation of Codes

The above analyses of what the federal Tax Court does and what the Congress does reveal two different tax codes: the judicial and the legislative. A separation of powers that maps relatively cleanly onto substantive areas of law is different from traditional understandings of separation of powers. Writings on the federal separation of powers conceives of the branches splitting and dividing functions and activities between themselves, but not also as splitting topics. For this reason, recognizing how the separation of powers works in tax upends many longstanding assumptions about how that separation works more generally. In the following Part, I will first turn the above descriptions of the judicial and legislative Codes into a more general account of tax and separation of powers. I will then offer two different hypotheses about why the judicial and legislative Codes are so separate substantively. Next, I will put this analysis of tax and separation of powers in conversation with the conventional approach to the separation of powers to show what the study of tax adds. Finally, I will explore normative implications of this new understanding.

A. Differences Between Judicial and Legislative Codes

The accounts above of the judicial and legislative Codes reveals a substantive separation between the two. The Code provisions that most frequently occupy Tax Court and the ones that occupy Congress are relatively distinct from each other.

As described above, the tax judiciary spends the bulk of its time on sections defining income, business expenses and charitable gifts. Going down the list of judicial Code sections shows that Tax Court also issues relatively frequent decisions on investment, business and other kinds of losses. Moving further down the list reveals sections dealing with charitable organizations, self-employment tax, personal deductions and some definitions and issues around retirement savings. The judicial Code consists almost entirely of provisions that apply broadly and facially pertain to anyone filing a tax return.

In the studied period, Congress never turned its attention to most of those judicial sections. So, what was it doing instead? During the studied period, the legislature was primarily taking up our substantive topics in the legislative Code. The first consists of industry-specific benefits; the others deal with income support for low- and middle-income families, natural or semi-natural disasters, and retirement savings. Most of the legislative Code falls into this first bucket. In contrast, Tax Court rarely deals with any of these topics. In the 775 cases examined in this study, the Tax Court looked at the child credit only 17 times and the earned income tax credit only 16 times. The court examined the energy production credit only 14 times.

The top five sections of the judicial Code and the legislative Code do not overlap. Neither do the top ten. The top twenty only have one section in common. Widening the lens, the judicial Code had approximately 408 sections in it during the studied period, and the legislative Code had 333. The branches took up many of their sections more than once. Tax Court interpreted its 408 sections a total of 4,036 times and Congress amended its 333 sections a total of 1,074 times. There were 57 sections that both branches looked at more than once. These tended to be sections that neither the judiciary nor the legislature took up very often.

The following graphic shows the overlap, with each Code section weighted by how many times each body addressed it.

Figure 1

Figure 1

The judicial Code provisions appear in the shaded region next to the y-axis and the legislative Code provisions appear in the shaded region next to the x-axis. As described in the results above, only one section is so commonly examined by both that it does not belong in either bucket, and that is section 170, on charitable gifts.

Comparing the sections at the forefront of the judicial and legislative Codes also demonstrates that not only are the two branches concerned (generally) with different Code sections, but also that those Code sections pertain to distinct areas of tax law. Setting aside procedural matters, the top five sections of the judicial Code address basic issues such as what counts as income, what counts as a business expense, and what counts as a charitable gift. In contrast, the top five sections of the legislative Code deal with depreciation for specific industries (and perhaps companies), income support, retirement savings, and energy companies. Widening the lens, the top ten provisions of the judicial Code includes the same topics as were in the top five, and the top ten of the legislative Code also encompasses real estate businesses and charitable gifts. The only topic that both the Tax Court and Congress both took up with any frequency was charitable gifts, namely, the problem of conservation easements. Zooming out again to the top twenty sections of the judicial Code also covers loss deductions, in addition to the subjects already in the top ten. In the top twenty of the legislative Code is also tax return confidentiality in addition to the topics included in the top ten.

Actors involved in the political process appear to be at least slightly aware of the separation between judicial and legislative Codes. In the study cited above, law professors Shuyi Oei and Leigh Osofsky interviewed Congressional staffers engaged in the tax legislative process and asked about the extent to which Congress considers possible court interpretation down the line. Staffers reported that “this was not a dominant focus.” The “potential future interpretations of statutory language” appeared not to be at the forefront of interviewees’ minds. One interviewee even stated directly when asked about audience, “Not courts. Forget the courts. The courts don’t matter.” While the existing scholarship on tax has not previously observed the separation between the judicial and legislative Codes, this quote suggests that, perhaps at the margins at least, legislative actors understand that Tax Court will not, in fact, examine any of the laws they are writing anytime soon.

Before elaborating on this argument, a few caveats are in order. First and foremost, this split between the judicial and legislative Codes does not say anything about the third branch of government. But, in fact, executive agencies are enormously important in tax, as an abundant body of literature has shown. While dialogue with that literature will be essential to developing a fully-formed theory of tax and separation of powers, that dialogue is beyond the scope of this Article.

A second caveat is that the separation between the judicial and legislative Codes does not imply that the Tax Court is interpreting statutes that Congress did not write. If the current study expanded its time horizon to the full history of the full Code, every section of the judicial Code is also in the legislative. However, this analysis is interested in what the tax judiciary and legislature are doing routinely, on a regular week-to-week basis. The question is not whether Congress has ever looked at a tax statute because, by necessity, Congress has written all of them. The open questions concern which statutes Congress returns to frequently, and to what extent those statutes overlap with the ones the Tax Court reconsiders often. The answer is to the latter question is that there is very little overlap in this regard.

Third, the split between the topics at the fore of the judicial and legislative Codes also does not imply anything about how the tax judiciary and the legislature concretely divide tasks. As will be discussed at length below, under traditional separation-of-powers analysis, the judiciary and legislature vary by functions rather than substantive issues. In the language of the classic formulation: the legislature makes the law, the judiciary interprets it; the law is the same, but the two branches approach it differently, according to an accepted division between their respective functions, not different matters of substance. Now, to be sure, Tax Court and Congress do allocate tasks more or less along functional lines and the activities associated with them. Therefore, the difference in the routine activities of members of the tax judiciary and members of legislature probably explains some of the substantive split between the judicial and legislative Codes, but by no means all of it. The next subpart turns to potential explanations for the substantive difference.

B. Possible Explanations

1. Textual Differences

The analysis so far has demonstrated a substantive separation between the Code that the Tax Court interprets and the Code that Congress crafts and amends. This next subpart considers possible reasons for this separation. Why do courts concentrate on some tax statutes and legislatures others?

The first set of potential explanations has to do with the language and structure of the statutory texts themselves. The way a statute is written could predestine it to its fate as judicial or legislative. Perhaps easily observable characteristics like the statute’s length or its number of subsections can go a long way toward predicting into which of the two Codes a particular section will fall. Or, the answer may depend on less quantifiable factors, such a statute’s diction, sentence structure, or its overall architecture.

In his foundational article on “easy cases,” law professor Frederick Schauer argued that Constitutional law scholars have disproportionately attended to the handful of cases that make it to court. Those cases present difficult legal questions; they are the “hard cases.” However, Constitutional law in fact exercises much of its power in the “easy cases,” whose answers courts need not resolve. The division, he argues “between what constitutional theorists have focused on and what they have largely ignored is a product of what has and has not been before the courts.” Why do some Constitutional provisions wind up in court? It is the “often fortuitous way in which some clauses have been drafted in different styles from others.” Constitutional litigation’s preoccupation with “certain substantive areas is importantly although certainly not exclusively, a product of linguistic design, in which relatively precise language forestalls litigation with respect even to matters of great moment, while relatively vague language encourages it.”

Perhaps a similar analysis applies in tax. The split between substantive areas may also be a product of linguistic design. One way to think about the differences in language between the Codes would be to think about the oft-discussed line between standards and rules. As law professor Louis Kaplow observed in his canonical piece on the divide between “rules” and “standards,” the two differ by the “extent to which efforts to give content to the law are undertaken before or after individuals act.” Or, as tax scholar Emily Cauble puts it in the tax context, “[a] ‘rule’ specifies, clearly and in advance, the tax consequences resulting from various activities,” whereas “a ‘standard’ provides only limited textual guidance to taxpayers before they act, deferring definitive determinations of tax consequences to after-the-fact analysis by the IRS and courts.”

In fact, Kaplow’s article actually suggests that whether a law is a rule or a standard relates to whether the law is the general province of courts or legislatures. He observes that some laws are rules and others are standards in part because of the “differences between legislatures, which promulgate many laws, and courts, which often apply them.” If legislatures write rules and courts apply standards, perhaps the rules/standards distinction is all we need to understand why some Code sections are legislative and others judicial. Maybe any section written as a standard just has to be part of the judicial Code.

Some of the judicial Code sections are in fact standards. For instance, tax scholars Alice Abreu and Richard Greenstein have convincingly argued that section 61 “has the appearance of a rule but is actually a standard.” This is true, they argue, because neither the statute nor the foundational case law specifies “what counts as an ‘accession to wealth.’” As a standard, according to this analysis, the question of what is income generates “a large ‘borderland’—a predominance of controversial cases.” Those cases can wind up in Tax Court and place section 61 squarely in issue.

To take another example, the Supreme Court itself has described at least some parts of the business-expense statutes, which belong to the judicial Code, as standards. Interpreting the distinction that sections 162 and 212 make between “business” and “personal” expenses, Justice Cardozo once wrote that “[o]ne struggles in vain for any verbal formula that will supply a ready touchstone. The standard set up by the statute is not a rule of law; it is rather a way of life. Life in all its fullness must supply the answer to the riddle” The frequency with which business-expense cases wind up in Tax Court suggest that the IRS and taxpayers often disagree about the answers that “life in all its fullness” is supplying.

That said, the rules/standards distinction takes us only so far in understanding into which category a particular statute falls. Not all judicial Code sections are standards, and not all standards are judicial. For instance, the section 165 theft loss statute is more like a rule. The extensive discussion in Littlejohn and similar cases are about how to define the word “theft.” The Court’s analysis is about parsing the relevant state law, not about balancing facts and circumstances. The section 170 statutory scheme has a number of rules; the Chrem case described above shows how the Tax Court can devote significant effort to determining whether the taxpayer has complied with one of them.

On the other hand, corporate tax law, found in “subchapter C” of the Tax Code, is full of standards, but the Tax Court does not spend time on them. For instance, tax scholar Richard Kovacs has correctly observed that, “[t]he ‘Subchapter C’ income tax provisions of the Internal Revenue Code applicable to corporations and their shareholders contain a troublesome handful of indefinite standards. He gives the example of section 385, which distinguishes corporate debt instruments that are known as debt from corporate equity ones. The statute articulates a nonexhaustive list of factors to determine whether a financial instrument is debt and, in so doing, “sets forth a true facts and circumstances standard.” Yet section 385 is nowhere in the judicial Code, and subchapter C is generally also missing. Similarly, tax professor David Weisbach has observed that many of the Code’s provisions aimed at preventing abusive tax avoidance take the form of standards. As he describes, a “typical anti-abuse rule allows the government (and only the government) to override the literal words of a statute or regulation” and requires a ‘reasonable’ tax result if the taxpayer enters into or structures a transaction with a principal purpose of reducing tax liabilities in a manner contrary to the purposes of the statute or regulation, even if the transaction otherwise literally complies with the rules.” The anti-abuse standards are also largely absent from the judicial Code.

The provisions of the legislative Code do suggest that, in recent years, Congress appears to be trying to write tax statutes as rules as opposed to standards. The examples discussed above do seem to be specifying, “clearly and in advance, the tax consequences resulting from various activities.” Congress generally may decide to write rules rather than standards as an interbranch power grab to keep its favorite subjects out of the hands of Tax Court. Yet, the conservation easement counterexample is an illustrative counterexample. Congress legislated this rules-based statutory framework several times during the studied period. Nonetheless, the Tax Court has dozens of conservation easement cases. Statutory ambiguities, in other words, can arise in rules and in standards, even if more in one than the other, and the Tax Court spends time on both, even though the Congressional preference in tax today appears to be for rules.

A different way to think about potential language differences between the judicial and legislative Codes is that the judicial Code may consist of what scholars have called “common law statutes.” As Constitutional scholar Chas Tyler has recently written,”[c]ommon law statutes” are a “special kind of federal statute,” written in “sweeping and general terms.” According to Tyler, “[c]ommon law statutes include some of the most significant and frequently litigated statutes on the books,” including the Sherman Act, the Civil Rights Act of 1871, the Labor Management Relations Act and the Federal Employers’ Liability Act. Uniting common law statutes are their broad terms, terms that courts have, by Congressional design, lots of room to interpret. As law professor Margaret Lemos has observed: “Most everyone agrees that common law statutes represent delegations to courts.”

Some sections of the judicial Code do seem to resemble the famous common law statutes in their sweep and level of generality, leaving the judiciary plenty of room to interpret. For example, section 61 with its broad and somewhat circular definition of income as “income from whatever source derived,” followed by a relatively short nonexhaustive list of examples, leaves substantial room for courts to flesh out the foundational concept of income. In section 162, the key words, “ordinary” and “necessary,” are relatively vague, and courts have a wide berth to decide, for example, how “necessary” it is to attend a Trump-branded seminar or race a yacht.

That said, the common law concept does not explain entirely the separation of Codes. Not all judicial Code sections are written in sweeping and general terms. For instance, while section 274 does start with a general rule, prohibiting deductions for “ for any item with respect to an activity which is of a type generally considered to constitute entertainment, amusement, or recreation, or [w]ith respect to a facility used in connection with an activity referred to in subparagraph (A),” none of those terms seem particularly sweeping or general, and, more importantly, the relevant statute goes on for dozens of subprovisions about what exactly these terms mean. Section 165 also has a long list (and a relatively specific list) of deductible losses. While the “abuse of discretion” language at work in the section 6320 analysis in the Loveland case seems fairly vague, that language is from prior case law, not the statute, which, like most of the procedural statutes, is quite precise in language.

In addition, not everyone agrees on what a common law statute even is. If a common law statute is one that delegates authority to courts, how do we think, Lemos, asks, about a federal statute like, Title VII, a “‘normal’ statute that most would agree lies outside the privileged common law category[?]” As Lemos observes, Title VII bars employment discrimination on the basis of race, sex, religion, and national origin, but that language itself leaves plenty of work for courts to do. For example, she writes, the statute does not say what it means by “discrimination” or to what extent “‘employer[s]’ are liable for the actions of supervisors or other employees.” In short, it is “silent or ambiguous” in significant respects. As a result, Lemos argues, even though most scholars do not call it a common law statute, Title VII also constitutes “a delegation to courts.” While Title VII does not delegate as much to the courts as paradigmatic common law statutes like the Sherman Act, “the differences are not categorical differences; they are differences in degree.”

Yet, the common law statute concept does seem to apply to the judicial Code, at least to some extent. Insofar as the judicial Code sections are in fact common law statutes, they become subject to the ongoing scholarly debate about the extent courts should interpret them differently than other statutes. One entrant in that debate is the “delegation model.” The delegation model holds that, because common law statutes delegate substantially unrestrained lawmaking power to courts, judges should therefore decide cases arising under common law statutes in a “free-wheeling and policy-driven manner.” A second model, the “incorporation model,” posits that courts should treat common law statutes as if they incorporate any common law rules in force when the relevant statute was enacted.” Tyler recently proposed a third model under which “judges should decide cases arising under common law statutes by applying rules representing a consensus among American courts today—i.e., rules that jurisdictions generally have in common.” While scholars have occasionally looked at methods of statutory interpretation in the Tax Court, no one has examined the question of whether certain tax statutes are common law statutes, whether Tax Court judges are in fact using any of the three models and which, if any, the court should use. These are questions beyond the scope of this Article, but interesting ones for future research. In any event, the common law framework does seem to be a partially useful one for understanding the split between the judicial and legislative Codes.

In sum, textual differences between the judicial and legislative Codes do have some power in explaining their separation. Yet, the two map incompletely onto various theories of textual statutory difference. Further, because Congress writes statutory text, the features of that text do not do much to explain why Congress is taking up the topics it selects. The next subpart, on political economy, will talk about that question in more detail.

2. Political Economy Differences

Another possible explanation for the separation between the judicial and legislative Codes concerns not how they are written, but for whom. In other words, the separation of the judicial and legislative Codes is also a matter of factors that social scientists treat under the header of “political economy.” The judicial Code involves different players than the legislative Code, and those actors differ in their economic resources and political wherewithal.

As described above, the judicial Code largely takes up the issues of individual taxpayers and small businesses with limited political organization. The cases described above give a flavor of the kinds of taxpayers who normally wind up in Tax Court, from the low-income to the affluent small business owner.

In contrast, most of the legislative Code focuses narrowly on a particular industry, typically one with vast economic resources and significant political organization. Of the top 20 provisions of the legislative Code, only five of them, section 401 (retirement savings), section 72 (annuities), section 24 (child tax credit), section 32 (earned income tax credit), and section 6103 (tax return confidentiality), apply broadly across the taxpaying population. The rest of them, the depreciation provisions, the energy credits and the real estate subsidies all involve small, technical changes or additions to statutes affecting only a small group of very wealthy taxpayers, or even, perhaps, as in the motorsports example discussed above, a single taxpayer. Even in the case of section 170, the charitable-giving rules, the pieces of the statute that Congress has recently been busy amending deal almost entirely with the rules for conservation easements, which presumably have nothing to do with most taxpayers.

The fact that Congress often passes such narrowly tailored tax legislation fits with the research reviewed above on the general workings of the tax legislative process. That research consistently describes a process with high interest-group involvement. The current study adds to that body of research a new layer: not only are interest groups often successful at getting what they want in the tax legislative process, but they also appear to be crowding out other things. Based on the relatively high number of narrowly crafted sections in the legislative Code, Congress appears to be devoting the bulk of its tax-legislative resources and energies to a small group of industry taxpayers.

Along similar lines, Constitutional law scholar Aziz Huq has begun to tease out the connection between Constitutional structure and political economy. In a recent article, he drew a link between the separation of powers and control of the Congressional agenda. Huq highlights how “governments are typically confronted with a great number of real, tangible issues at any one moment but can attend to them only one at a time,” a situation that requires creation of a “list of subjects or problems to which governmental officials, and people outside of government closely associated with those officials, are paying some serious attention at any given time.” Unsurprisingly, “the government’s agenda is typically an object of interest-group contestation” Interest groups “mobilize to elevate novel issues onto the government’s agenda, frame those issues so as to maximize their comparative advantage and engage in ‘negative blocking’ of disfavored issues. Crucially, interest groups also provide a “legislative subsidy” in the form of information about their pet issues Political-science literature has further documented how those interest groups that Huq notes that, as a matter of Constitutional law,”[a]genda-control rules [are what] directly concern the who and the how of state power.

What becomes part of the legislative Code, as opposed to the judicial Code, depends in part on agenda control. As a practical matter, Congress hears directly from interest groups, whereas the Tax Court does not. In her work on Constitutional structure and the distribution of wealth, Constitutional law scholar Kate Andrias describes the mechanism by which interest groups generally drive legislative agendas:

Corporate and trade association lobbyists provide information in the legislative process by call[ing] attention to issues, furnish[ing] evidence about how problems are being experienced on the ground, and provid[ing] expertise about the anticipated consequences-both substantive and political of proposed solutions. On less salient and more complex issues, well-funded business groups are often the only real lobby.

As “the only real lobby,” interest groups (along with Treasury and the Service) have some ability to decide what becomes part of the legislative Code.

External stakeholders may have less direct influence over the content of the judicial Code. While executive agencies, namely, Treasury and the Service, are certainly heavily involved in shaping the Tax Court docket, the data presented above suggests that the sorts of private parties that benefit from the legislative Code do not take their issues to court. It is true that my sample of Tax Court cases does include a some with multi-million dollar amounts in controversy and expensive litigation teams. However, the judicial Code does not include many industry-specific provisions. Unlike Congress, the Tax Court is not often dealing with depreciation schedules or energy or real estate. Perhaps that is because those statutes are written so precisely as to give rise to almost no disputes with the Service. Perhaps, too, that is because the audit rate is so low that the Service rarely evaluates taxpayer compliance in these areas. Or, perhaps large business interests do not want to take their issues to a forum where the Service may win. The judicial Code is a risky place to be. When well-resourced companies have a tax problem, they may make the strategic choice to devote those resources to lobbying rather than to litigation. For a taxpayer with choices between branches, the legislative Code may be preferred.

IV. Separation of Powers

A. Descriptive Implications

This account of the judicial and legislative Codes tells a story about separation of powers in tax. It is a very different story than the one that the enormous literature about separation of powers usually tells. The scholars who have contributed to that literature argue among themselves over many important points, but they generally agree that the separation of powers is about the separation of functions, not substantive topics. As political philosopher Jeremy Waldron puts it, “[t]he separation of powers counsels a qualitative separation of the different functions of government-legislation, adjudication, and executive administration.” Comparative constitutionalist Lawrence Claus sums up the consensus view:

[The Constitution] displayed and implemented two strategies for distributing power that have ostensibly operated and been hailed as hallmarks of American government ever since, and that have been widely emulated elsewhere. To distribute power horizontally within the new national government, the document purported to separate power by kind - legislative, executive, judicial.

Law professor Victoria Nourse has similarly observed that “[s]tandard understandings of the separation of powers begin with the concept of function.” Put another way, as cited above, Constitutional law scholars Richardson and Macey describe how the Constitution “fashions branches that are ‘omnicompetent’ as regards subject-matter but ‘unipowered’ as regards tools at their disposal.” The Constitution’s authors “purported to invest three institutions they created with differing kinds of power;” Articles I, II, and III begin “not by naming institutions, but by naming kinds of power, and then created institutions to exercise those powers.” For the drafters, “[t]he received wisdom declared separation by kind as the way to achieve liberty and the rule of law.” As a result, they made a “deliberate choice not to enumerate a clear separation of powers based on a taxonomical account of each branch’s substantive responsibilities.”

Yet, in the tax context, the branches do split along substantive lines. The separation between the judicial and legislative Codes is not merely about kind but also about substance. Empirically documenting, as the current study has done, the existence of such a phenomenon is a new contribution to the vast separation-of-powers literature. Even if the tax branches are still performing separate functions, the result is a separation in topics. This result is not one that scholars have previously considered.

In a welcome development, a recent strand in separation-of-powers scholarship has begun to move to make descriptive claims about what the three branches are doing. These accounts, however, keep their eyes trained on the function each branch performs, rather than on any substantive distinctions. For instance, constitutional scholars David Fontana and Aziq Huq have examined how the ways in which branches allocate power (or functions) among themselves “evinces the lingering influence of institutional [i.e. intra-branch] loyalty of the kind Madison anticipated, particularly in the executive and judicial contexts. Officials may variously support an increase or a decrease in the power of their institution but often enough be motivated by a loyalty to the best interests of their institution.” In their article “Parties Not Powers,” constitutional scholars Daryl Levinson and Richard Pildes re-envisioned “the law and theory of separation of powers by viewing it through the lens of party competition.” For them, “[the] practical distinction between party-divided and party-unified government rivals in significance, and often dominates, the constitutional distinction between the branches in predicting and explaining interbranch political dynamics.”

However, almost all of the extensive recent theorizing in legal scholarship about the separation of powers has been rooted in high-profile Supreme Court cases regarding a recurring handful of topics. On this basis, Constitutional law and administrative law scholars have developed descriptive accounts of how separation of powers works and should work. Yet, these accounts are limited in scope, generally taking only one area of law into account. Conversely, scholars from within particular legal subfields may look at how law and policy operate across the branches, but they fail to include bigger-picture separation-of-powers themes into their ensuing law and policy debates. Writing of the Supreme Court, but equally applicable to the separation-of-powers literature generally, criminal law scholar Rachel Barkow has observed that tax scholarship on the Court “approaches separation of powers questions with little regard for the substantive contexts in which they arise.”

Focusing a comparative lens on tax law reveals a different separation-of- powers landscape. That new landscape is particularly important because its contours reveal political and economic dimensions that scholars in the field have only recently attempted to include. As Daryl Levinson wrote in his 2015 Foreword to the Harvard Law Review, it is one thing to separate power among branches but “[i]t is quite another to ensure that power is divided between Democrats and Republicans, the rich and the poor, or racial or religious majorities and minorities, or to prevent one such group from tyrannizing the other.”

However, Levinson also laments that mapping political and economic dynamics onto the three branches of government can be difficult. Unfortunately, he writes, ‘[o]ver long time horizons, it really is hard to come up with reliable generalizations about differences” between, for example, levels of interest group power “at the highest levels of constitutional structure.” Now, “[c]onceptually, decisional power can be located by ascertaining the policy preferences of the relevant institutional actors and, in cases where these preferences conflict, observing actual policy outcomes.” Locating decisional power is exactly what identifying the differences between the judicial and legislative Codes allows us to do. Pinpointing what is in the judicial Code and what is in the legislative Code offers an entrée into who is pulling the levers in the different branches and on which substantive issues that pulling process exerts its impact.

B. Normative Implications

So far, this Article has documented the differences between the judicial and legislative Codes. This final subpart asks: to what extent is the divide problematic from a normative point of view? The separation-of-powers literature discussed in the previous subpart provides a lens through which to begin considering that normative question. Now that this Article has answered its motivating comparative question about which branch of government does what about the Code, the further question arises as to how well my results fit with—or diverge from—accepted theories about how different branches should divide their powers?

To begin to address this issue, it is useful to follow scholars who classify normative scholarship on the separation of powers into the two dominant camps of “formalism” and “functionalism.” For formalists, to evaluate a separation-of- powers arrangement requires “identify[ing] the type of power being exercised and . . . make certain that that power is exercised by an official residing in the appropriate governmental institution.” Or, as formalist scholar Thomas Nachbar puts it, the normative question on matters of the separation of powers should be “how to attribute power to the institutional forms defined in the Constitution rather than a substantive evaluation of an act.”

Turning to the functionalist camp, it is helpful to bear in mind that definitions of functionalism vary. In this subpart, it refers not to a description of the “functions” of the different branches as this Article has thus far used the term function. Instead, as regards functionalism as a normative school of legal thought, the expression has to do with a set of approaches to evaluating arrangements of interbranch power. Law professor Victoria Nourse offers one version of this approach. Functionalists, she writes, “do not see the question so much as one of legitimacy or authority as one of pragmatism.” The normative question for functionalists is whether our government should constitute a “‘workable’ whole.” Constitutional scholar Elizabeth Magill describes functionalism as an approach that “emphasizes that the departments of government have shared and overlapping powers, as well as separate powers.” The key issue becomes “whether an institutional arrangement upsets the overall balance among those branches by permitting one of them to compromise the ‘core’ function of another.”

A number of scholars, including Magill herself, have recently begun to look beyond or around the formalism/functionalism divide to argue for altogether different standards, some of which emphasize what proponents call “non-domination,” rooted in ideals of equality. For instance, recently, two pairs of prominent scholars, Nikolas Bowie and Daphna Renan, on the one hand, and the earlier-cited Richardson and Macey on the other hand, have made such proposals. Both teams look to the values underlying the Constitutional account of the separation of powers. Writing recently in the Yale Law Journal, Bowie and Renan argue that: “the normative values underlying the separation of powers in eighteenth century England—a definition of freedom as nondomination, protected by a rule of law enacted by political equals—found fertile soil in the new American republic.” These values demand what these scholars call a “republican separation of powers” that “advances the values of political equality, nondomination, and the rule of law.” For them, the favored “republican separation of powers” should encompass an ideal of political equality where “disputes about the construction of the American state are resolved not by one special group of people but by ongoing political contestation among representatives of a community of political equals.”

For their part, Richardson and Macey label their plan “a theory of separation of powers based on a principle of “anti-domination,” based on “the capacity of the branches to mutually check one another-whether or not it is exercised in a given case.” These scholars argue that “[t]he Constitution’s separation of powers is thus a highly formal idea, but its substantive content is modest: the separation of powers is breached only where a branch’s actions would allow it to tyrannize or to dominate another.” The separation of powers exists to “prevent the agglomeration of power in one branch-in other words, that no one branch dominates the machinery of government.”

Applying one or more of these separation-of-powers theories may help to assess the substantive split between judicial and legislative Codes normatively. It seems unlikely that anything about that split regarding tax law poses an actual legal problem under the Constitution. Yet, the different approaches that scholars have used to evaluate general separation-of-powers issues as legal matters also provide useful tools for normatively assessing the tax separation of powers.

Take the formalist approach. Even formalists themselves may agree that “[t]he problem of distinguishing the three functions of government has long been, and continues to be, one of the most intractable puzzles in constitutional law.” Nevertheless, formalists think that in separation-of-powers cases, the Supreme Court should determine whether the challenged branch action falls within the definition of that branch’s constitutionally derived powers of executive, legislative, or judicial.

One might evaluate the tax separation of powers using the formalist’s implied embrace of the notion that different branches have distinct strengths and weaknesses. According to this perspective, one strength the legislature would have as the democratically elected branch is that of weighing competing values and making difficult moral choices among them, or, in Magill’s words, specif[ying] “the basic norms that govern behavior.” Consistent with this view, the research reported in this Article shows that much of the legislative Code does in fact take up big questions involving competing values. For instance, the legislative Code deals with tax credits, including the ones for working, having children, producing various kinds of energy, and building housing. Having a tax credit for a particular activity reflects the value judgment that the activity is something that the legislative branch wants to subsidize. Which externalities and moral concerns might call for tax credits are topics that, by formalist thinking, perhaps do belong in legislative hands.

Even so, as my data show, it is too simple to say that the legislative Code includes all the provisions that involve competing values and that the judicial Code does not have any. Issues like what is income, what is a business, what is a gift, and what is a charity are also value-laden and create powerful government incentives. For instance, as tax scholars Bill Whitford and Beverly Moran write in their race-based critique of the Code, the Tax Court’s above-cited gift analysis reminds the reader that “a transfer with no obligation to repay constitutes a ‘gift’ for tax purposes only if it results from the donor’s detached and disinterested generosity.” Whitford and Moran observe that “the net result of the ‘detached and disinterested generosity’ requirement is that gifts from strangers (such as prizes and awards) are usually taxed,” whereas, in contrast, gifts from family and friends are not. Further, “wealthy people generally count other wealthy people as their family and friends, while low-asset individuals can only hope to get wealth transfers from strangers and lotteries.” For this reason, the Tax Court’s jurisprudence “favors the more fortunate both because wealthy individuals have access to more gifts and because they have access to the ‘right’ gifts.” Deeming family gifts “the right gifts” is a value judgment and perhaps one better suited to democratically elected and accountable legislators.

Similarly, tax professor Dorothy Brown examines the race- and class-based judgments embedded in the judicial Code that prevent taxpayers from deducting commuting costs as business expenses. As she writes:

Commuting expenses are not deductible because they are regarded as a personal preference reflecting where to live, but . . . the decision where we live may not be controlled in its entirety by personal preference. What if I want to live near work to minimize my commute but can’t afford to? What if I want to live near work, but I either can’t buy or can’t rent because of my race?

The commuting rule has its roots in case law from the mid-twentieth century that pre-dates many large-scale historical shifts in employment and transportation such as the rise of national and multinational businesses with multiple primary locations, of telecommuting, of air travel and most recently of remote work arrangements. How section 162 should account for these developments raises complex value judgments but not ones that Congress is considering. Instead, section 162 is firmly in the judicial Code. Again, there is no actual Constitutional problem here; the Tax Court is clearly adjudicating disputes in section 162 cases rather than writing any laws. Yet, if the tax judiciary is the only branch regularly revisiting certain foundational Code sections, it may start in practice to usurp the elements of the formal legislative role, and vice versa. For this reason, the formalist perspective does suggest some reason for concern about the tax separation of powers.

Moving onto the functionalist approach, its multiple versions make it somewhat difficult to evaluate the tax separation of powers from its perspective. Nourse’s formulation of functionalism suggests that the relevant criterion may be some kind of basic workability. Political philosopher Jeremy Waldron takes a different tack, writing that the paradigmatic functional separation of powers involves something like a principle of legislative supremacy, where the legislature has an “initiating place on the assembly line of law-making/law enforcement.” Taking the Waldron approach suggests a similar problem to the one identified under the formalist criteria. As this Article documents, with regard to judicial Code sections, Congress is not taking an initiating place; in fact, Congress is not looking at judicial Code sections at all. With regard to legislative Code sections, the Waldron principle is technically satisfied. Congress is, after all, amending the sections. But as a practical matter, there is reason to doubt that altering the depreciation schedule for certain fruits and trees occurred on the actual initiative of a member of Congress.

Turning back to Professor Nourse’s functionalist formulation, the tax separation-of-powers system is “workable” in that tax law gets written and adjudicated but it is not clear how to evaluate whether it is “working” well. For instance, it is hard to see what is particularly “workable” about having the legislature rather than the judiciary look at the depreciation schedule for fruits and trees. Understanding how businesses use different kinds of assets and what rules should govern them should perhaps require consistency, technical expertise and familiarity with relevant facts. Practically speaking, an argument can be made that judges offer these things at least to the same extent that legislators do. On the other hand, deciding what kind of disasters should give rise to deductible losses is perhaps a decision that requires immediate responsiveness to current events rather than prolonged judicial procedure. Given that a workability can cut multiple ways, the that the split between judicial and legislative Codes is probably not irrevocably unworkable, but the “overall balance” that a functionalist would want may not be perfectly calibrated.

Then, as the most recent separation-of-powers scholarship suggests, the test for a valid interbranch power allocation is about anti-domination and equality. Under that view, the tax separation of powers is also imperfect. For instance, the political economy analysis points to a failure of the political equality Bowie and Renan find important. Well-financed interest groups have the potential to wield outsize power in deciding which Code section goes where. Most taxpayers, of course, have no way of raising their concerns with Congress. For instance, the Lovelands and other taxpayers facing the collection-due-process framework might have views about what that should look like. Those views might, from a policy perspective, be right or wrong, but Congress will not be considering them either way; it does not regularly consider the collection-due-process rules. For another example, scholars concerned with racial wealth inequality might be eager for a legislative hearing as to how they propose to fix the tax law on gifts, but Congress has not been holding any hearings on section 61.

Contemplating the fact that separation of powers as construed generally might lead to problems with domination and political equality, Levinson writes:

The problem, from this perspective, is not just that concentrated wealth becomes concentrated power, but also that the two are linked together in a mutually reinforcing dynamic. The political power purchased through wealth may allow economic elites to enact self-interested policies that increase their wealth, and hence their political power. This dynamic creates a feedback loop of increasing inequality in both domains: the rich get richer; the powerful get more powerful.

That dynamic seems to be reflected at least to some extent in the divide between judicial and legislative Codes.

It is not as clear that the tax separation of powers allows one branch to “tyrannize” another, as would raise a concern under Macey and Richardson’s approach. However, the textual differences between the two Codes point to one potential area for, if not tyranny, somewhat dominant Congressional behavior. The legislative Code written during the studied period consisted largely of highly technical statutes, possibly drafted with industry help, that leave little interpretive room for the judiciary. An industry can attempt to forestall the Tax Court from ever issuing an adverse ruling by having Congress write rules that, while they look complex, are actually easy to follow if a taxpayer knows how to do so. As Cauble has written, provisions like the many judicial Code sections “that ground tax law in rules rather than standards can . . . make tax consequences more predictable”—however only for sophisticated taxpayers who “contemplate tax law prior to acting.” For this reason, enacting narrow, technical and rule-like statutes that Tax Court could rarely reach may not allow the ideal amount of judicial check on Congress.

On the other hand, while this Article has the executive branch has generally been outside the scope of this Article, perhaps its agencies are also partially controlling the legislative agenda and doing legislative drafting, perhaps even in the public interest. Maybe, for instance, the Department of Energy or energy experts at Treasury are working with Congress to draft energy tax credits. Agencies deciding what is in the legislative Code do not necessarily create the problematic feedback loop that Levinson posits. Additionally, at least some of the legislative Code involves income support credits that have nothing to do with powerful interest groups. Existing scholarship does not explain how the earned income and child credits got onto Congress’s agenda so frequently during this period, but these credits got there, which implies that placing a topic into the legislative Code is not just a matter of special-interest maneuvering.

Perhaps taking these political economy factors into account and applying a non-domination lens to the separation of tax powers suggests a bigger role for the judiciary. Energy companies could have pursued their tax agenda through an adversarial process like litigation, and courts could have interpreted the credits more aggressively. However, “there are reasons to be wary of relying on the judiciary to reduce the checks and prods imposed by wealth.” Litigation, like legislation and regulation, is susceptible to wealthy interest-group pressures. How well one navigates the litigation process depends on resources; “[l]itigation is expensive[, and ][w]ell-financed groups are able to litigate more frequently, to influence more effectively the information tribunals receive, and to strategically settle cases that may produce unfavorable precedents.”

Moreover, as observed above, “under the adversarial system, courts generally consider only the arguments of the actual litigants.” While individual interest groups may have few opportunities to badger Tax Court judges, so too does anyone with a public-spirited mindset, as “[i]ndividuals or organizations interested in the future implications of a case, but not in the judgment itself, generally lack standing and receive inadequate consideration.” Insofar as policy considerations are important to a tax case outcome, the Tax Court is “often not presented with the full array of arguments.” The case-by-case nature of litigation also means that “focus on the particular courts tend to underweigh, or be underinformed about, the systemic context and consequences of their decisions.”

Generally, while the political economy dimensions of the split in Codes may not be ideal from a non-domination perspective, is also not clear that re-sorting Code sections by branch would resolve concerns about domination, much of which seem to be larger systemic problems of which the split between judicial and legislative Codes is only a symptom.

In other words, whether one takes a formalist, functionalist, or other approach to evaluating the divide between judicial and legislative Codes, some normative concerns arise. Yet, because those concerns are varied, they do not all point in one direction in terms of potential solutions. As a result, while a different interbranch allocation of topics or more collaboration between the two branches might address some of the normative problems raised here, no across-the-board policy fix suggests itself.

V. Conclusion

This Article set out to answer the question: who does what in tax law? Data from the Article’s novel comparative study revealed a divide in substance between different branches of the federal government. In tax, the judiciary and the legislature mostly work on separate topics. While the two branches appear to split by function, more or less along the lines that traditional doctrines about the separation of powers would predict, these branches also, more surprisingly, divide up topics. The judiciary takes up broad, foundational tax questions like what is income, what is a business expense, what is a charitable gift, and what is a loss. In contrast, the legislature devotes itself primarily to tax credits, sometimes for income support but more often for narrowly defined interests.

The substantive divide between judicial and legislative Codes relates both to differences in how the texts of the Code sections are written, as well as to political-economic factors, namely, who in society these Codes affect. These two potential explanations open up more questions for future comparative research, particularly the normative questions that this Article has highlighted. At the present time, the judicial Code contains sections that might fall under the heading of common law statutes, which would imply that, from a statutory-interpretation perspective, judges have broader policymaking authority than they otherwise might. Should judges do more to drive policy in the judicial Code? If so, in what ways? Looking back to an earlier time period than the one examined in this study, are there sections of the legislative Code that contained more sweeping language that may have provided an opening into which judges could move with relative ease? Or should Congress have stopped writing statutes that could at some point become part of the judicial Code? If this is a move for Congress to take, why? What is more, the substantive split between Codes suggests an outside role for special interests in choosing which tax provisions land, respectively, in the legislative vs. judicial branches. Could either the judiciary or the legislature find ways of leveling that playing field? Finally, for reasons of length and scope, this Article did not examine the third branch of government. What does the tax separation of powers look like in regard to the executive branch? These questions and related ones provide a robust research agenda for articles to come.