Unions win or lose elections according to a majority of valid ballots cast. That’s been NLRB practice since the 1930s, but arguably, this is not the approach that makes the most sense. Why do card checks require majority support from a whole bargaining unit, but representation elections only require a majority of votes that are cast? Employers have pushed back on the “votes-cast rule” over the decades, as we’ll discuss below, but Constellation Energy Generation is now asking the Board to outright overturn the doctrine.
Currently, Constellation has a matter pending in front of the Board. This involves an International Brotherhood of Electrical Workers (IBEW) election that took place in July. IBEW Local 15 won, 87 to 73, with one challenged ballot, but if the union had been required to win a true majority from the 181-worker bargaining unit itself, it would have needed 91 votes, as Constellation argued in a brief.
Where did the votes-cast rule come from?
This rule emerged from the Board’s 1936 R.C.A. Manufacturing decision, which revolved around a union vote where a rival union boycotted the election, and only around a third of the eligible bargaining unit cast votes. The Board then certified the winning union while reasoning that a turnout requirement would allow either a minority union or an employer to manipulate election results by keeping voters away through intimidation.
The current issue is whether a case that revolves around such extreme facts should have governed nearly a century of union elections. Most elections look nothing like what took place in 1936 Camden, NJ, where a rival union organized a boycott and where ballots did not include a “no union” option. In Constellation’s case, around 89 percent of the bargaining unit voted, and such a high turnout makes a boycott unlikely.
A majority… of whom?
To consider that question, we need to look at the text of the NLRA. Under Section 9(a), bargaining representatives must be “designated or selected for the purposes of collective bargaining by the majority of the employees in a unit appropriate for such purposes.” In its brief, Constellation argues that “the majority of the employees” should be interpreted to include all members of the bargaining unit, not simply those who showed up to vote.
To support this argument, Constellation points out that Congress specified the words “valid votes cast” in Section 9(c)(3) as applied to runoff elections and “a majority of the employees eligible to vote” in Section 8(a)(3) concerning deauthorizations. This language, as Constellation argues, shows that when Congress wants to specifically define the measure of voters in a union election, it’s explicitly done so through statutory text.
To stress that point more, Constellation is looking for consistency in how the NLRA is interpreted regarding the definition of “the majority of the employees” when it comes to fairness in union elections. The company wants the same standard applied to elections as for card check. For the latter, the Board requires a majority of the whole unit to support unionization. It’s only fair to have consistency in standards.
Why now?
This isn’t the first or even the second time that employers have made an argument similar to what Constellation is offering now. In 1940 and 1944, respectively, employers in NLRB v. Whittier Mills Co. and NLRB v. Central Dispensary & Emergency Hospital argued that Section 9(a) requires a majority of the entire unit, rather than simply a majority of votes cast. In both cases, the courts rejected that interpretation.
Fast forward to 2026, and Constellation is arguing that “it is high time for the Board to revisit R.C.A.” because of the Supreme Court’s 2024 ruling in Loper Bright. That case ended Chevron deference, so courts no longer generally defer to agencies (and no longer accept what the Board decides is reasonable labor policy). Accordingly, Constellation’s brief argues that the Board must stop supplying its own meaning to the NLRA and instead follow the statute’s plain text.
R.C.A., the company argues, was a policy judgment made regarding an extreme situation, and after Loper Bright, agencies must follow a statute’s “single, best meaning” regardless of policy preferences.
Constellation also cites Hospital Menonita de Guayama, Inc. v. NLRB, the July 2026 D.C. Circuit decision that applied Loper Bright while striking down the successor bar. In doing so, the D.C. Circuit pointed out that Section 9 requires majority support for a bargaining representative, which is in line with the argument that Constellation is making. Hospital Menonita is one of the first appellate court decisions that have struck down an NLRB doctrine in a post-Loper Bright world. Also, if Constellation refuses to bargain with IBEW, and the NLRB orders it to do so, the company can take its case to the D.C. Circuit.
What’s next?
For the moment, employers have a wait on their hands. The NLRB could decide Constellation’s election challenge solely on a separate issue regarding whether certain positions held by voting workers should be classified as supervisors, which would take them outside the NLRA. If the Board does this, it could choose to entirely bypass the company’s Section 9 arguments. However, if the Board does reconsider the Section 9 issue, there’s a question of how far back in time any change could go. Would any relief simply apply to election challenges currently in process or in the future, rather than taking a retroactive view?
That’s a can of worms that the NLRB would probably be hesitant to open, considering the 90-year run for the R.C.A. presumption. The Board isn’t likely to offer retroactive relief from the votes-cast rule, but it’s not a stretch to imagine how challenges to pending elections would pile up pretty darn fast. Before any of that would happen, the Board must decide whether it wants to touch the Section 9 argument at all.
