The Immodest Textualism Of Justices' Retirement Tax Opinion

March 15, 2019Article
Law360 Expert Analysis

On March 4, in BNSF Railway v. Loos,[1] the U.S. Supreme Court resolved a circuit split about whether time lost from work because of a personal injury qualifies as “compensation” under the Railroad Retirement Tax Act. By answering yes, the court enables railroad employers to deduct railroad retirement taxes from the lost wages portion of personal injury awards and settlements with their employees. The opinion is unsatisfying, not because it is bad or wrong so much as it is analytically immodest.

Textualism, the now-familiar mode of statutory interpretation championed and popularized by former Justice Antonin Scalia,[2] requires judges to determine only what the statute says. Done properly, textualism disempowers judges by restricting the ever-present urge to bend a statute’s meaning to a judge’s desired outcome in the case. In this way, textualism is the interpretive means to achieve judicial restraint. It has become the norm, and this is a good development. (Naturally squabbles continue about whether an avowed textualist followed the theory properly in a particular case — we’re still lawyers, after all.)

But textualism too has its excesses. While it is good that we fear judges with red pens on their ears, eager to “correct” the legislature’s work, we must also guard against the judge as a cloistered academic — jamming the world into a theory, however the fit. Textualism’s cannons are for lawyers what supply-and- Michael Arndt demand graphs are for economists: Invaluable foundational tools that explain much. Over-reliance on these tools, however, zaps the life from issues, turning live questions into academic hypotheticals. When this happens, the debate never leaves the classroom, even if there is a history that might question a premise or orient the discussion. Minimum wage debates remain theoretical despite an abundance of real-world cases that could corroborate, refine or modify the theoretical answer. Likewise, in the Loos opinion, the eight decades of litigation involving similar injuries appeared rather beside the point.

Justice Ginsburg wrote for the seven justice majority and began with the statutory text. Compensation under the Railroad Retirement Tax Act is “any form of money remuneration paid to an individual for services rendered as an employee to one or more employers.”[3] The U.S. Court of Appeals for the Eighth Circuit — and the two dissenting justices at the Supreme Court — found that damages for personal injuries were not paid “for services rendered” but rather for services that were not, and could not be, rendered. The majority found this reading unpersuasive by way of the statutory context.

The court likened the Railroad Retirement Tax Act’s “compensation” to the Federal Insurance Contributions Act’s “wages.” Both statutes were passed during the Great Depression and aimed to provide sustainable retirement programs for workers. Both statutes require contributions from the employer and employee. Although different terms are used for the taxing triggers in these statutes — compensation and wages, respectively — the terms are defined similarly and expansively. For these reasons, this statutory context informs the arguments.[4]

Having made the comparison, the majority then relies on two cases that interpreted “wages” under the Federal Insurance Contributions Act.[5] In those cases, the court considered whether back pay for wrongful termination and severance pay qualified as “wages.” In both cases, the court answered yes, even though “wages” are defined as remuneration for service “performed by an employee.” The Eighth Circuit's and the dissent’s reasoning clearly did not apply in those cases because back pay and severance were not payments for services performed by the employee. Rather, as the majority explains, both cases recognized that “wages” moved beyond the periods of active service and captured periods of inactive service. The court found this broader conception also applied to Railroad Retirement Tax Act’s “compensation.”

This broader reading was confirmed by other statutory context. The court addressed but did not spend much time on the companion statute, the Railroad Retirement Act. That act specifically includes pay for time lost — unlike the current version of the Railroad Retirement Tax Act. The court’s holding puts the funding statute on par with the benefits statute, but the court — perhaps due to practical considerations discussed below — did not dwell on this point.

Instead, the majority discussed the previous versions of the Railroad Retirement Tax Act and the consistent regulations. Both confirmed the court’s reading: The Railroad Retirement Tax Act includes — and has always included — absences from active service within the employment relationship. Justice Ruth Bader Ginsburg cited a 1938 regulation that stated as much and explained that “[t]his understanding has governed for more than eight decades.”[6]

The opinion is certainly an acceptable textualist holding. The interpretative mode did not, however, mandate the outcome.[7] As pointed out by SCOTUSblog just nine months ago, the Supreme Court has distinguished between the Railroad Retirement Tax Act and the Federal Insurance Contributions Act.[8] The court could have used the same approach in this case by giving meaning to the differences in language in the Federal Insurance Contributions Act and the Railroad Retirement Act, on the one hand, and the Railroad Retirement Tax Act, on the other.

Even if the Federal Insurance Contributions Act cases govern, there are factual distinctions: Both of those cases had defined periods of absence from work while that is not necessarily true under this ruling. As both the railroads and the Internal Revenue Service acknowledged, a separate statute treats general verdicts as consisting entirely of lost wages — even if that was at odds with the pleadings in the case and the laws regarding verdicts in the forum. This default assumption — treating a general verdict as all lost wages in the absence of a specific allocation — undercuts the railroad’s argument as to why pain and suffering was different.

At oral argument, counsel for BNSF explained that pain and suffering was not “payment for services rendered in the same way because the amount of lost wages is directly tied to salary for services rendered.”[9] That’s not true in a case where a plaintiff receives a general verdict that is more than his claim for lost wages. The entire verdict would be lost wages and would not be “directly tied to salary.”

These types of quibbles are perhaps unavoidable. If congressional texts were abundantly clear and everyone agreed on the proper textualist analysis in each case, then there would not be circuit splits in the first place. Even if we agree on the proper approach, there may be disagreement. This is not a revelation.

The more striking feature is the textualist’s conceit. The majority found that the understanding that compensation included absences from service “governed” for more than eight decades. Governed whom?

Loos was one of approximately four thousand cases under the Federal Employers Liability Act just last year. There have been thousands and thousands of cases since the Railroad Retirement Tax Act was enacted in 1937, so one would have thought this governing understanding would have applied to all of those cases. As the dissent pointed out, however, none of the parties “identified a single case where the IRS has sought to collect RRTA taxes on a FELA judgment in the 80 years the two statutes have coexisted.”[10] This announced the suspicion — articulated in many of the cases considering this issue — that the taxability of awards was a newfangled litigation strategy at odds with decades of contrary practice.

At oral argument, the IRS acknowledged that this may have happened but responded with two points: (1) The IRS would not be in a good place to follow thousands of FELA cases each year, and (2) even if railroads did not submit the taxes, that would not change the regulation.[11] BNSF, for its part, suggested that the railroad companies were keenly interested in the solvency of its retirement system, raising the specter that employees would claim retirement benefits despite not contributing.[12]

The majority found this rationale “plausible” and more reliable than the dissent’s “conjecture” that the railroad companies’ sought to gain negotiating leverage.[13] Each side could muster more arguments on this front, but notice that these arguments would apply equally to a brand new statute. There is an academic constraint to the conversation: The decades of practice are not worth looking up because they could not matter.

To be sure, the IRS is correct when it says that violating a law does not render the law meaningless. Widespread speeding does not mean there are no speed limits. But, when the law is less clear than a speed limit, the history may give us pause or a little less certainty in our textualist conclusion. If, for example, the railroad retirement system was imperiled by workers claiming benefits after refusing to pay their share of taxes, there would almost certainly be evidence of that — either from the Railroad Retirement Board or from railroad companies who have paid after employees refused to do so.

This evidence was unnecessary as a “plausible” suggestion was sufficient to proceed to the textualist analysis. Likewise, how the issue arose — many decades after the legislation and after thousands of cases without a reported opinion — might modify a de novo textualist review. History, of course, will not change the text’s meaning, but it might bolster a contested textualist reading. Perhaps the short shrift given to these considerations is simply a matter of the record in the case, and the dissatisfaction is best blamed elsewhere. If so, we would have expected a discussion of this absence or perhaps even remand for a better record. Instead, we got a textualist opinion with little regard for the statute’s applications in thousands of cases across many decades.

While textualism constrains the judge’s ability to ignore the legislature, it must maintain a certain worldliness or context to resist devolving into pure theory. We rightly fear a judge re-writing statutes but may be no better off with the judge buried in a book with the curtains drawn.

[1] Burlington N. Santa Fe Ry. v Loos , 2019 U.S. LEXIS 1734, ___US___ [2019].


[3] 26 U.S.C.S. Section 3231(e) .

[4] We also pointed this out in our oral argument preview,

[5] United States v. Quality Stores Inc. , 134 S. Ct. 1395, 1400, 188 L. Ed. 2d 413
(2014) (considering severance pay); Social Sec. Bd. v. Nierotko , 327 U.S. 358, 365-66 (1946) (considering back pay).

[6] Slip, p. 3.

[7] For list of how lower courts addressed the issue, see


[9] Oral Argument Tran., 14-15.

[10] Dissent, p. 2.

[11] Oral Arg. Transcript, 22.

[12] Oral Arg. Transcript, 13.

[13] Slip, p.5, n.2.