The NLRB Will Eliminate Faster Union Elections. How Much Does it Matter?
In the four years since the union election rules changed, the election time period has been two weeks shorter, and the union win rate increased by more than 3%.
In 2014, Obama’s NLRB adopted a number of rule changes that were intended to streamline how union elections were conducted. These changes followed a lengthy NLRB rule-making process which received tens of thousands of comments and went into effect in April, 2015. Perhaps the most important result was that the time period between the filing of a union petition and the election would be shorter. That’s why this rule change was opposed by employer groups as promoting what they called “ambush elections.” Their complaint was that employers had less time to “educate” workers about the drawbacks of organizing a union. In other words, they would have less time for their union busting campaigns.
Now with Trump’s NLRB, nothing good can be left alone. The board published a Request for Information (RFI) about this rule in 2017. It has recently announced a new rule that would implement more than a dozen changes to the process and would involve “extending deadlines and adding steps” that are expected to increase the election time period. I wanted to see what the impact of the 2015 rule has been and what we might lose with this change.
The Time to Election
Let’s verify that the time period is indeed shorter by looking at the median days until the election. It’s been more than four full fiscal years since the rule change, which occurred in the middle of federal FY2015. To see what the impact has been, the chart below shows the median time between the filing of the petition and the date of the election for standard RC union elections for four years before and after the rule change. The year of the 2015 rule change is shown in red. We can see a clear difference — the time period from 2011–2014 had an average median time of 38 days, falling to 23 days for the 2016–2019 time period, a decrease of about 2 weeks. In both time periods, the median is remarkably stable. The 33 days for 2015 makes sense since the rule change happened in the middle of that fiscal year. This data is from the NLRB website here.
The Election Win Rate
Here is the union win rate for these elections for the same time period. We again have a difference, but it’s fairly small — the time period from 2011–2014 had an overall win rate of 66.9%, rising to 70.4% for the 2016–2019 time period, an increase of almost 3.5%. This data is from the NLRB annual election reports here.
The post-2015 declining win trend may be interesting, but I’m not sure if that’s meaningful yet. Possibly it means that unions benefited right away from shorter elections, but employers have gotten better at dealing with it over time, leading the win rate to drop slightly.
But it would make sense that win rates would go up if the election time is shorter since there are two fewer weeks for the employer to engage in union busting. The No Holds Barred report from 2009 documents the employer’s standard legal and illegal union busting activities. They can range from required attendance at anti-union meetings, anti-union leaflets, one-on-one meetings with supervisors, threats to close the workplace, surveillance of workers, and retaliation against pro-union workers. These campaigns are often very effective at intimidating and spreading fear among workers and eroding their support for having a union. A recent report, Unlawful, found that employers spend hundreds of millions of dollars every year on union busting consultants and are charged with unfair labor practices (ULP) in 42% of elections.
Why Only a 3.5% Win Increase?
So this raises the question of why the union win rate has only increased by 3.5%. Let’s discuss what happens in an election. Unions need to file for an election with at least 30% of the workers signed on cards, but they should really file with well over 50%. After the petition is filed, if that’s the first the employer hears about it, they will ramp up their anti-union campaign. The employer wants to reduce the pro-union vote to less than 50%, and with this rule change there are two fewer weeks to do that. Let’s say it takes the employer a week to get their campaign running. That means in a 23 day median time period, they have about two weeks for union-busting, instead of four weeks under the old time-frame. This is half the time, and that can make a difference.
But since 23 days is the median time, that means half the cases have a longer time. Moreover, often the employer learns about the union campaign well before the petition is filed, so the union busting begins earlier. In these cases, the two fewer weeks of the election period, though helpful to unions, are not enough to make a huge difference in win rates. There’s still plenty of time for the union busting to wreak havoc. Overall, this 3.5% likely represents the small set of elections where the two weeks less time made a big enough difference — what would have been a slight loss changed to a slight win.
It’s also worth asking, is 3.5% a meaningful increase? Or is this just the random fluctuation of win rates? Moreover, it’s possible there are some other factors involved aside from the election time period. For example, despite all the ULPs, unions have actually gotten better over time at winning elections, and the win rate could have gone up anyway. The answer may be coming — if the election time period gets longer again, we’ll see if this trend reverses.
But with these results, it’s tempting to say that every week you shorten the election period, the union win rate rises by about 1.7% percent, but that would probably need further analysis. We should look at the election win rate for various election period time brackets. Possibly we would see a roughly linear relationship, with the win rate rising as the time period falls. It would also be interesting to see how the size of the bargaining unit matters here. The annual NLRB election reports unfortunately don’t list the election filing date so each case would have to be looked up individually to get this data. But I think this trend is probably right.
I’ll also note something else that’s potentially interesting, which is that another set of data on the NLRB website, which lists the number of petitions, elections and union wins, shows a higher win rate after 2015 of 71.5% and a larger win rate difference before and after 2015 of almost 6%. The difference between this data and the election report data I have been discussing is that this is for all cases while the election reports list only closed cases that year. This would need further analysis, but the difference suggests that unions lose some of their apparent wins in post-election litigation, and this litigation may have increased after the rule change. That would make some sense, since if employers have less time before the election to fight the union, they may fight even harder afterwards.
Withdrawn Election Petitions
It’s also interesting to look at the percentage of election petitions withdrawn by the union. A union would withdraw a petition after filing if it determined later on that a loss looks likely. So if an employer is engaging in union busting and decreasing the support for the union over time, at some point support for a union might drop below 50%, and the union is more likely to withdraw the petition. So we would expect the number of withdrawn petitions to decrease when the election time is shorter.
That’s in fact what we see here, though the difference is again fairly small. The time period from 2011–2014 had an overall withdrawal rate of over 30.3%, decreasing to 27.5% for the 2016–2019 time period, a nearly 3% decrease. The NLRB data is here. I’ll emphasize that aside from many election losses, these withdrawn petitions are another casualty of union busting campaigns. And again, we’ll see if this trend reverses with longer election periods.
The Impact of the Shorter Election Period
Assuming all these rate changes are meaningful, we can calculate a rough estimate of the impact of the shorter election time in terms of the increase in the number of workers who were able to win a union. After the rule change, a higher percentage of elections were actually held, since fewer petitions were withdrawn, and of those held, a higher percentage were won.
So what would have happened if the election time period was never changed? In the 2016–2019 time period, a total of 7,153 election petitions were filed, 4,594 elections were held, and 3,233 elections were won. Taking out 2.8% more petitions that would have been withdrawn, and reducing the number of subsequent election wins by 3.5% gives us about 295 fewer elections won. With an average bargaining unit size of 63.3 workers in these years (calculated from the NLRB annual election reports for 2016–2019), this gives us 18,673 fewer workers who would have won a union. So we can maybe say that the election change helped about 18,673 more workers organize a union, or over 4,600 workers per year.
If this seems like a small number of workers, let’s remember that relatively few NLRB elections are held these days. In fact there were only 1,047 elections in FY2019, far lower than the more than 8,000 elections/year in the 1970s, and as low as the election numbers seen in the first years of the NLRA in the 1930s. This is a major problem. Though union elections are just one avenue for worker organizing, they do remain an important one. But I have written previously that the NLRB election strategy is simply not organizing enough workers to reverse the decline of the labor movement. Only about 200,000 workers won a union through these elections in the 2016–2019 time period, or about 50,000 per year. That number would have to be ten times larger to cause a real increase in union density.
And Now with this New Rule Change?
With the election rules going back closer to the old system, and if nothing else changes, we could see the union win rate slide back down a bit to the mid 60% range and tens of thousands fewer workers over time will succeed in their efforts to organize. Possibly the rules will be changed again for the better in the future. But it’s clear that the NLRB union election process will remain tough terrain for the labor movement.