Notes on Labor Data

Improving Success of Individual Certification Petitions at the NLRB

August 24, 2021

0k2k4k6k8k10k12k↑ Certification petitions19502000 →
0k2k4k6k8k10k12k↑ Elections19502000 →
0k2k4k6k8k10k12k↑ Union certifications19502000 →
Solid lines are RC petitions from NLRB case records, by fiscal year filed (petition counts for 1965-1983 are from the Annual Reports). Dashed lines are the NLRB Annual Reports' own counts: RC petitions by fiscal year filed, and elections and certifications by fiscal year a case closed, which include RM cases before FY1964.
0%20%40%60%80%100%↑ Petition to elections19502000 →
0%20%40%60%80%100%↑ Election to certifications19502000 →
0%20%40%60%80%100%↑ Petition to certifications19502000 →
For petition to elections and petition to certifications, the dashed lines are a year's elections or certifications over that year's petitions, not a cohort of petitions.
0k100k200k300k400k↑ Single-union eligible voters19502000 →
0k100k200k300k400k↑ Single-union workers in certified units19502000 →

Increasing Unit Size

The relatively favorable number of workers is not only due to rising success rates, but also to the fact that the average size of single-union bargaining units that win a certification election is also increasing.

406080100↑ Average unit size for elections19502000 →
406080100↑ Average unit size of certified units19502000 →

Data

Data from the current NLRB case management system, the CATS system (1999-2011), and the CHIPS system (1984-2000), data on representation elections from 1965-1998 compiled by the AFL-CIO c/o JP Ferguson. The year 1999 is excluded because of data artifacts due to the transition between data systems.

About 5,000 cases in the 1961-1998 data, mostly closed in 1982, have no filing date. Since the NLRB assigns docket numbers in order within each region and case type, I interpolate their filing dates from the nearest dated cases by docket number.

The number of RC cases filed from 1965-1983, and the series from 1950-1995 labeled as the annual reports’, are compiled from NLRB Annual reports. The case-record series use NLRB fiscal years of filing: July to June through 1976, and October to September after. Cases filed in the July to September 1976 transition quarter are left out. The annual reports count elections by the fiscal year a case closed, include employer-filed RM cases before 1964, and do not follow petitions to their outcome, so their petition-to-election and petition-to-certification ratios divide a year’s elections or wins by that year’s petitions.

Single-union elections are elections with only one union on the ballot. Petitions consolidated into one election count as one election, though each still counts as a petition. When an election’s eligible-voter count is more than ten times the unit size on the petition, and at least 1,000, I treat it as a keying error and use the petition’s count.

The 1965-1983 elections come from the AFL-CIO file, which only includes elections that ended in a certification of representative or certification of results; the later data include every election held. Checked against the annual reports, the AFL-CIO file has about 95% of single-union elections in every year, so the earlier years are somewhat understated relative to the later ones.

Subscribe