The National Labor Relations Board used flawed data to support a rule change that would reduce unions’ power to defend against anti-labor campaigns, a Bloomberg Law analysis found. Deficiencies in the data could weaken the board’s defense in a legal challenge, two administrative law professors said.
The Republican members of the NLRB want to revamp the so-called blocking charge policy that has been in place for eight decades, which pauses elections to approve or de-certify a union when unfair labor practices are alleged. At the heart of their argument is time: They contend, in a proposed rule released in August, that unions can file frivolous actions as a stalling tactic when unions believe results will go against them, and that resulting delays in elections deprive employees of “free choice.”
The NLRB determined that 155 blocking actions were filed from 2016 through 2018, and that the median number of days those cases were delayed ranged from 122-145 days. But a Bloomberg Law review of data supporting the rulemaking found dozens of cases in which the board overstated the length of delays attributable to blocking charges over the last three years—overshooting the mark in one instance by more than 12 years, and in another by five years.
The board’s data over-counted delays in more than one-third of cases—55 in all—in which they said blocking charges were filed. In those cases, the board listed a statistical impossibility: a delay that exceeded the total number of days between when a petition to approve or de-certify a union was filed and when an election was actually held, after the NLRB investigated a blocking charge.
The errors open the NLRB to a potential legal challenge under the Administrative Procedure Act, which instructs courts to invalidate regulations and other actions by federal agencies if they’re proven to be “arbitrary, capricious” or an abuse of discretion, according to the two administrative law professors who reviewed Bloomberg Law’s findings. Litigants have won that kind of claim against the Trump administration at a higher rate than under any other president—a trend that lawyers and administrative law professors say is due to a rush to create new rules amid cutbacks in administrative staff.
“They could handle a minor error easily, just by issuing an addendum, but this sounds like a big mistake,” said Richard Pierce, a law professor at George Washington University who’s authored over 20 books on administrative and regulatory law. “If they proceed without correcting it, the resulting action would be hard to defend.”
Management-side attorney Michael Lotito said he doesn’t believe any errors in the Notice of Proposed Rulemaking will be particularly problematic, so long as the agency accounts for any discrepancies and relies on accurate data in issuing the final rule.
Pierce suggested the agency consider restarting the rulemaking effort from scratch, given that the significance of the errors could bolster an argument that the agency acted arbitrarily even if it corrects or amends its notice.
NLRB spokesman Edwin Egee repeatedly declined to comment specifically on the discrepancies or to discuss the board’s methodology. The NLRB can’t “make public statements about this data during the rulemaking process,” he said.
Egee said the board included “a significant amount of data pulled from agency records, including a list of over 4,600 cases,” in an effort to encourage the public to comment on the proposal. He added that the board “looks forward to reviewing and considering all comments submitted, including any questions regarding the data, with the issuance of our final rule.”
The proposed rule would allow for an election to take place without delay even if a blocking charge is filed. Votes would be impounded if such a charge were made. If the NLRB substantiates an allegation of unfair labor practices, the election could be held a second time. Unions and worker advocates say that change would encourage employer interference in unionization drives and illegal assistance in efforts to de-certify unions.
‘Not a Data-Driven Argument’
The data discrepancies appear to be the result of multiple factors. One of the most likely is that the board incorrectly counted cases in which more than one allegation of improper conduct was made.
- For example, one case involving Yale University (01-RC-183014) was erroneously listed as nine separate cases. That created a large discrepancy in the board’s assessment of the period of delay. The proposal listed these “separate” cases as being blocked for 677 days, but said it took only 178 days to hold the election—meaning 499 extra days were counted in each case. As a result, the board inflated the blocking period count by 4,491 days, or more than 12 years.
- Another example involves an ongoing case. The board said that Piedmont Gardens, Grand Lake Gardens, 32-RC-087995, a petition filed in 2012 for a unionization drive among workers at retirement communities in California, was blocked for 4,634 days, or more than a dozen years. In reality, the case was filed just seven years ago, meaning the board overstated the delay by five years.
There were other problems with the board’s use of agency data.
Three cases in its list of petitions that involved blocking charges were duplicates. That means those cases—Waste Management of Pennsylvania, Inc., 06-RD-216734; Hospital Damas, Inc., 12-RC-217728; and Hospital Damas, Inc. 12-RC-217749—were counted twice.
The board’s lone Democrat, Lauren McFerran, who dissented from the rulemaking, faulted the GOP majority for overstating the number of petitions that involved blocking charges. So too did the AFL-CIO after reviewing the board’s data.
“My colleagues not only err by artificially inflating the length of time periods that their cited cases were blocked, they also err by artificially inflating the number of ‘blocked petitions pending,’” McFerran wrote in her dissent.
The board’s Republican members did not respond substantively in the rulemaking notice to McFerran’s accusations. “We need go no further in discussing the details of the dissent, other than to note that we already have her predetermined opinion about the proposals, regardless of what comments or further analysis may ensue,” the majority said.
When Bloomberg Law asked Egee in a subsequent conversation about McFerran’s criticism, he said that “from the board’s perspective, this is not a data-driven argument; this is a substantive argument.”
Call to Update Dockets
The data issues highlight problems the agency has experienced with incomplete records in online public dockets. The NLRB inspector general recently zeroed in on this problem.
Bloomberg Law reviewed the online dockets for all 155 cases in which the board said a blocking charge was involved. In 74 of those cases the online docket didn’t mention that any blocking charges had been filed or provide other indications that an election was blocked for any period of time.
Egee declined to respond to questions about whether the lack of blocking charge information in these 74 cases is due to an agency failure to post information and documents on its case information website.
The NLRB inspector general recommended after a September 2019 audit that the agency “develop and implement a system of controls to address” issues regarding the “data accuracy and reliability” of its electronic case-processing system, NxGen.
After Bloomberg Law questioned the board about the apparent public docket problems, the agency announced Oct. 21 that it would begin enforcing a requirement that most motions and case documents be filed electronically. That decision in effect will implement a rule enacted in February 2017.
Deadlines and Attrition
The two administrative law professors—Pierce and Case Western Reserve University School of Law professor Jonathan Adler—said the discrepancies in the board’s data could be evidence of the NLRB’s rush to take deregulatory action and the effect attrition can have on the federal workforce.
Those two factors, they said, have contributed to a record number of successful legal challenges against agency rules and policies under the Trump administration.
The administration had lost 55 of 59 such legal challenges as of Nov. 6, according to the Institute for Policy Integrity at the New York University School of Law. A majority of those challenges were filed under the Administrative Procedure Act and argued that proposed or enacted regulations were “arbitrary and capricious.”
“The NYU data reflect incompetence at an unprecedented level,” Pierce said. “Typically, agencies win about 70% of the cases in which their actions are challenged.”
Staffing also has become an issue at the NLRB. The agency’s field staff declined from 1,136 workers to an estimated 857 from fiscal years 2011 to 2018, an overall 25 percent decrease, according to the union that represents NLRB staffers. Union representatives said a large majority of that overall staffing loss happened during the Trump administration and under the board’s current leadership.
Administrative pressures surrounding the regulatory process may be especially pronounced at the NLRB, where the current Republican leadership has undertaken a rulemaking agenda that is more aggressive than anything the agency has experienced in decades, Adler said.
Pierce said the errors are substantial enough to warrant an addendum or correction.
Adler said that if data used in the rulemaking “is demonstrably false, that’s usually a problem. But if there are different possible assumptions the agency can make, provided the agency is transparent and explicit about why it’s making those assumptions, courts will generally accept that.”
Lotito, the management-side attorney, who is co-chair of law firm Littler Mendelson’s employer advocacy group, was more optimistic about the rulemaking and the board’s chances in litigation.
“The purpose of the NPRM (Notice of Proposed Rulemaking) is to get stakeholder input, and part of that deals with info about data,” Lotito told Bloomberg Law. “I’m sure if there needs to be clarification with respect to anything in the NPRM, that’ll get fleshed out by commenters, and the board will have the benefit of that analysis to incorporate into its final rule.”
The comment period for the proposal ends Dec. 10, with replies due Dec. 24.