I was skeptical of unions. Then I joined one.

A union isn’t just right for Vox Media, but for everyone.

On November 17, 2017, right after the Vox Media editorial staff started a push to unionize, I sent out my worst tweets of all time. “I am against #VoxUnion,” I wrote in one of the tweets. I wrote in another, “Vox Media is a generous company (unusually so for digital media), and some people want to take advantage of that.”

“I am generally fine with and even supportive of unions,” I concluded. “Just not this one.”

I wasn’t convinced, based on my experience, that we needed a union to ensure the company treated us well. And I was worried that “lazy” or bad workers could take advantage of union protections to stay on the job — something I feel that police unions, for instance, have helped do with even the worst cops.

Almost immediately, I was barraged by much of lefty Twitter with a huge ratio. A few people tried to genuinely debate me, explaining that unions could be good for even workers who feel they are well off. But it was mostly insult after insult, and after a while, I stopped looking at my notifications.

A year and a half later, sleepless but amped up on coffee and solidarity from a 29-hour marathon final bargaining session, I celebrated with the rest of the bargaining committee, made up of select union members who negotiated directly with the company, as the union and Vox Media management reached a contract agreement.

I had done a complete 180 on unions.

Umair Irfan/Vox

Good organizing and outreach from colleagues helped change my mind about the need for a union at Vox Media specifically. But so did approaching the research on unions the same way I would any topic in my reporting: by looking at the data and talking to experts.

Research suggests that unions have their biggest effects from density. When more people are part of a union, unions don’t just boost their workers’ wages and benefits; they also lift up those they don’t represent.

Scott Eisen/Getty Images

Unions accomplish this in two ways: The first is through bargaining on wages and benefits, which, because unions tend to represent lower- and middle-class workers, helps people who generally haven’t gained as much from the US economy in recent decades. Second, politically active unions push for progressive policies that lift up the entire working and middle classes, not just their members. Indeed, unions were crucial to some of the biggest gains in this area in the past century, from the New Deal to the Affordable Care Act.

In doing this, unions also help address income and wealth inequality, which have fueled social and political discord in the US in recent decades. Based on reviews of the research, the decline in unions — of about 66 percent since the 1940s and ’50s — can explain about 10 to 30 percent of the rise in inequality we’ve seen in the past several decades.

Toward the end of my journey in unionization, between arguments over the finer details of the contract, the bargaining committee members reminisced on all the work we had put into this up to that point — a year-and-a-half effort that would soon, finally, come to a close. One of them remarked to me: “My favorite part of this process was how much you’ve changed on unions.”

When I wrote those 2017 tweets, I thought unions could be good for some workplaces but others were good enough without unions and so they should be avoided.

I was wrong. We need more unions everywhere.

Unions balance out the workplace

When I first sent out my tweets, I believed that unions could do some good in some places — mainly in low-skilled jobs, like those in fast food and the auto industry. Particularly at Vox Media, I didn’t see the need for a union. The company had long done right by me, offering what seemed like generous benefits with pretty good health insurance, “unlimited” vacation time, 401(k) matching, parental leave for both mothers and fathers, and more.

I saw unions as a balancing act to corporate interests, offering protections to lower-skilled workers who, without collective action, didn’t have much power over their bosses. They would make sense at McDonald’s and Amazon warehouses (both of which are not unionized) and at GM car factories, I thought. But not high-skilled industries like digital media, where workers could, on their own, use their skill sets as leverage over their bosses.

It was, admittedly, a pretty selfish — and, in retrospect, naive — way of looking at a union.

The first thing I learned is not everyone had the same experience; even in a company that genuinely does try to be the best in digital media, things can slip through the cracks, and a bad manager can make a world of difference. I had always gotten along very well with my bosses at Vox, but that could change in one corporate reshuffling. I also started to worry about the future: What if, in a very volatile journalism industry, I’m laid off, or Vox is sold off to another company? Who’s to say the next owners would be as good as the current ones?

I began to see myself as one company reshuffle or sale or economic downturn away from losing all I worked for.

Scott Olson/Getty Images

Research, meanwhile, has consistently shown that unions are good for most people in them. A 2017 review of the evidence by John Ahlquist, a political economist focused on labor and inequality at the University of California San Diego, found that men in the private sector at unionized workplaces make about 15 to 25 percent more than those at non-unionized ones. Another review by Jake Rosenfeld, a sociologist focused on unions and economic inequality at Washington University in St. Louis, reached similar conclusions, noting that unions consistently produce a premium for workers in them.

A recent study by Henry Farber, Daniel Herbst, Ilyana Kuziemko, and Suresh Naidu, using surveys and other data going back to the 1930s, found that this union premium has been remarkably consistent over the decades. And while less educated workers seem to get a bigger premium, higher-skilled workers still get one too.

Part of this is the result of collective bargaining, as unions negotiate higher pay for their members. But Rosenfeld told me that unions also give a “cultural voice” to workers — one that checks executive excess. It’s this concept, first described to me by my coworkers, that really attracted me to a union.

Scott Olson/Getty Images

Companies’ stated goals and values can be undermined by bad managers or overwhelmed by market forces. And even at a good company, these problems can pop up despite efforts to the contrary, and it’s always a risk to speak out about the problems alone.

The cultural voice gives a way to address those problems that would otherwise go unheard. During my year or so in the bargaining committee, I saw this firsthand: People would come to us with problems, the committee would raise the problems to management, and the company, which sometimes was genuinely not even aware of the problems, would try to fix them.

That’s exactly what I hoped a union could accomplish.

Unions fight income inequality

When I started out at Vox in 2014, I wasn’t making much money: $30,000 as a writing fellow in Washington, DC. I have risen up the ranks since then to make far more than that. In writing my anti-union tweets, I was focused on one side of the equation: that I had the opportunities to rise up quickly through the ranks.

Jane Tyska/MediaNews Group/The Mercury News via Getty Images

What I gave less thought to was that I had been making $30,000 to begin with. Vox, as it grew out of startup mode, had since publicly increased salaries for the lowest-paid writing position. But I realized I wanted to make sure the higher pay for the lower end was written into a contract, to guarantee no one went through what I did again. In fact, I was willing to give up some of my money, or raises, for the benefit of the lower-paid. I wanted to make Vox Media just a bit more equal when it came to pay.

It turns out that this kind of wage compression is one thing unions do very well: Ahlquist’s research review found that unions tend to bring wages up for a workforce’s lower end while keeping wages for the higher end a little flatter or even leading them to decrease. One study he cited found that unionization led to a 12 percent reduction in pay for the top management, such as the CEO or other executives. Part of that is likely a balancing act: As more of a company’s revenue or surplus goes to its lower-level workers, there’s just going to be less for higher earners.

This is the societal case for unions: In general, they help reduce inequality. And it’s the decline of unions that helps explain, in part, why inequality has gotten worse and worse over the past several decades.

Since the 1970s, the top 1 percent in the US has reaped the bulk of the benefits of the growing economy, through explosive income growth, while everyone else has seen much smaller gains or even stagnation. Meanwhile, the share of American workers in unions has declined from around a third of workers in the 1940s and ’50s to about 10 percent today — a trend that shows little sign of reversing.

Christina Animashaun/Vox

“From the research perspective, if you’re interested in more progressive domestic policy, unions have been the answer to unchecked corporate power,” Rosenfeld, the sociologist, said. “More unions mean the kind of counterweight to policies that benefit the economic elite. There’s pretty good decades of research demonstrating this.”

By using data going back to the 1930s, Naidu’s study found about 10 percent of growing economic inequality could be explained by declining union membership. “Not nothing,” Naidu told me, adding that it’s likely a conservative estimate because the study couldn’t capture all the downstream effects of declining unionization (such as whether a lack of unions lobbying for workers made it easier to move some jobs overseas).

Other research has put the estimate higher. According to Ahlquist’s review, the decline in unions explains about 30 percent of the growth in wage inequality for men in the private sector from 1973 to 2007, and about 20 percent for women.

Michael Strain, an economist at the right-leaning American Enterprise Institute, told me it’s likely true that “the decline of unions accelerated the growth of inequality.” But he argued that there have been bigger forces at play — particularly technological advances, from the internet to automation, that led to more jobs and higher pay for more skilled workers and the opposite for those at the lower end. We should focus on addressing the fallout from that, Strain said, instead of boosting unions that, he believes, may stifle economic growth.

The evidence indicates that technological change has played a big role in the rise of inequality, but unions have too. The most convincing study I found, published by researchers Tali Kristal and Yinon Cohen, looked at computerization in different workplaces to gauge what drove inequality. It concluded that “declining unions and the fall in the real value of the minimum wage explain about half of rising inequality, while computerization explains about one-quarter.” Depending on the time and industry, union decline could even be the biggest factor.

Christina Animashaun/Vox

This study does not come with 100 percent certainty. All of the research in this field involves imperfect data and methodological compromises.

But it’s the best data we have. And there seems to be widespread agreement that unions have some effect on reducing inequality, particularly when there are more of them. That’s a case for more unions — everywhere.

Unions also balance out politics

Unions’ equalizing effects apply to politics too: While the increasingly wealthy and big corporations have fueled a rightward shift in American politics and policy since the era of Ronald Reagan, unions could help — as they especially had before they were weakened in the 1980s — swing the pendulum back in the other direction.

The most illuminating study on union’s effects on politics comes from researchers James Feigenbaum, Alexander Hertel-Fernandez, and Vanessa Williamson. Their working paper compared bordering counties across state lines: one county in a state without anti-union right-to-work laws (which allow workers to refuse to pay union dues and fees even after a workplace votes one in, depriving unions of a revenue source), and the other in a state with them.

Scott Olson/Getty Images

They found that right-to-work laws reduced the presidential vote share for Democrats by 3.5 percentage points, with similar effects in other races. Overall voter turnout also dropped by 2 percentage points overall. (For reference, Donald Trump in 2016 beat Hillary Clinton by 0.7 points in Wisconsin, which adopted a right-to-work law under Republican leadership in 2015.)

At the state level, the study found that right-to-work laws also shifted policy to the right on labor issues, and led to different policy outcomes regarding, for example, prevailing wage and minimum wage laws — the kinds of policies that can help close income inequality.

Hertel-Fernandez told me that this is an underappreciated aspect of unions, but an important one: Just as unions balance out the workplace to help workers, they also help balance politics by creating a powerful set of organizations that can counter economic elites and corporate interests that often have a big say in Washington, DC, or state capitals.

Hertel-Fernandez compared what unions do for the left to what gun clubs and evangelical churches do on the right — mobilizing voters, educating them about issues, and even creating pathways to running for office. “If you’re in a union, you have experience showing up to meetings where you’re speaking in public, running for elected office for your union,” he said. “You gain these skills that you otherwise might not have if you weren’t in a union.”

Bastiaan Slabbers/NurPhoto via Getty Images

I relate to this. Before I posted my tweets, I honestly hadn’t given unions or labor issues much thought. Sure, I read about them on the news, going back especially to Wisconsin’s fight over an anti-union law in 2011. But my policy interests by and large lie in other areas — including criminal justice issues, guns, and drug policy.

Since joining a union, I have spent a lot more time reading about unions and labor issues. As a member of the union’s bargaining committee, I also got a lot of experience — some more than I, well, bargained for — over how to negotiate, organize, prioritize different constituents’ wants and needs, communicate with a broader public, manage expectations, set up events, talk and argue in public, and more. I had little need for these skills before. And unions appear to do this for a lot of people, based on Hertel-Fernandez’s research.

Every union isn’t the same. Some unions, like those for police or coal companies, may not advocate for progressive policies. And some workers may not like it when a union gets too political; Republican members of a union, for example, may not be too happy if their union endorses Elizabeth Warren over Donald Trump.

But unions’ political activism, on net, seems to shift America toward supporting progressive policies that can help workers, offering another explanation for studies finding more unions mean higher wages and less inequality.

Those on the right know this. As Republican strategist Grover Norquist wrote in 2017, “If Act 10 [an anti-union law for public-sector employees] is enacted in a dozen more states, the modern Democratic Party will cease to be a competitive power in American politics. It’s that big a deal.”

Unions don’t have to come with big downsides

I had another concern with Vox Media’s union: Given the volatile nature of the journalism industry, could a union lead to layoffs? This is a common argument from critics of unions: Once wages go up, employers cut back on overall hiring and work hours to make up for the higher costs.

There’s some research suggesting this is true. One line of studies, by economist Brigham Frandsen and others, looked at what happens to workers after close union representation elections, based on the idea that places where workers approve a union by a small margin aren’t too different from places that narrowly reject a union.

In one of the better versions of these studies, looking at unionization efforts in nursing homes, Frandsen and his colleagues found unionization did appear to reduce employment for some workers, particularly those with higher incomes, even as they led to higher wages, particularly for lower-paid workers.

Joanne Rathe/The Boston Globe via Getty Images

This line of research doesn’t provide total assurance. One of Frandsen’s studies indicated that places where unions marginally win and the places where unions barely lose might actually be more different than economists initially thought, undermining the basis for such studies. Still, they’re the best evidence we have about unions’ effects on employment.

Naidu, who was part of the study measuring unions’ effects going back to the 1930s, argued that the research on unions’ effects on employment “is not great” but that, ultimately, “the net effects are likely small.”

Altogether, the research and statistics don’t completely dispel concerns, raised by Strain at the American Enterprise Institute, that unions may lead to slower economic growth and less employment. But the effect seems small, so hard to detect in studies that it may not exist at all. There’s a bit of give and take, too: Even if unions do depress economic growth to a small extent, that could be worth the cost if it guarantees that the remaining growth benefits the lower and middle classes, instead of the super wealthy, to an extent that is not true today.

Besides, some countries with strong unions still have really strong economies. Experts pointed out that Norway has much higher union membership than the US, with more than half of workers in unions, but still a lower unemployment rate (3.4 percent) than America (3.7 percent). Norway is also a far more economically egalitarian country than the US. Yes, there are big differences between the economies of Norway and the US, but the numbers at least suggest that higher union density isn’t incompatible with a healthy economy.

Separately, I also worried about unions protecting their worst members — whether that’s people who were “lazy” and created more work for everyone else, or those who were sexual harassers or worse. Police unions are particularly notorious in this regard; many defend their worst members after shootings, blame the victims of such shootings, and even claim that “blue racism” against police officers is as bad as actual racism.

Scott J. Ferrell/Congressional Quarterly/Getty Images

But how unions deal with their worst members really falls on individual unions and their members to decide. Unions can be democratic, and they can determine what they value and create agreements that reflect that. We didn’t want to be a police union at Vox, so while our contract provides more general protections for members, it’s still very much possible to fire people for, say, plagiarism or sexual harassment.

Marissa Brookes, a political scientist focused on unions at the University of California Riverside, put it this way: The union “is not good or bad. It has the potential to create a great deal of good if we steer it toward diversity and equality. But like any organization, it can be captured by bias, captured by special interests.”

If the United States were to see more unions, some of them would undoubtedly be bad, just as some governments, politicians, and corporations are corrupt. But the downside of bad unions, or unions’ potentially negative effects on employment or economic growth, would have to be weighed with the proven benefits unions have on net for workers’ wages and overall inequality.

Policy reforms can encourage more unions

Historically, economist Richard Freeman found in 1997, unions have grown in spurts. He argued that there needed to be a push from the bottom up to a certain level of density or membership, at which point unions would rapidly expand as they gain more members and resources to mobilize.

For this to happen again, there needs to be more interest in unions. There need to be more workers like those of Vox Media and other digital media outlets, but in all sectors, willing to organize and with the ability to actually do it.

A survey released last year found there’s already solid public support for unionizing. Researchers Thomas Kochan, Duanyi Yang, William Kimball, and Erin Kelly wrote, “To put these findings in perspective, if all of the non-union workers who have a desire to join a union had the opportunity do so, union membership could increase by approximately 58 million workers, essentially quadrupling the number currently represented by a union, which would raise union density to 54%.”

Helen H. Richardson/The Denver Post via Getty Images

That’s not quite how unionizing works. Workers can’t just sign up to be in a union; each workplace has to go through the arduous organizing process and get at least half of their qualifying coworkers to support it. But the finding indicates that there’s a lot of support for unions — which is backed by Gallup’s surveys too.

So why haven’t more workplaces organized? By and large, for the same reason that unions collapsed in the first place, experts told me: Union-unfriendly laws and regulations made it easy for employers to shut down unions. So with the dawn of globalization and other shifts in the economy (from more competition abroad to a more fractured business environment), many companies didn’t hesitate — especially with the support of politicians like Reagan — to take advantage of the weak laws to tear down unions.

Kate Andrias, a legal scholar focused on labor law at the University of Michigan, described current federal labor laws as heavily tilted against unions.

One example: Penalties for violations of labor laws are extremely weak, to the point “that it’s actually economically rational for employers to engage in them,” Andrias explained.

Imagine that an employee manages to win an unfair labor practice claim in front of the National Labor Relations Board, which is already a difficult challenge. For that victory, she’d only be entitled to back pay minus any wages earned in the interim, Andrias said. Especially for a bigger company, paying part of a worker’s wage just isn’t a significant penalty.

A lot of anti-union conduct is also legal, such as replacing striking workers. And the laws and rules themselves are largely outdated, because they were built for direct employer-employee relations instead of the tangle of subcontracting that’s become more common in recent decades.

“The weaknesses were always in the law,” Andrias said. “But there was a shift — that historians write about — in which employers began much more aggressively exploiting those weaknesses in the 1980s.”

Among the union-friendly experts I spoke to, they generally echoed the same goal for labor policy: create a system of sectoral bargaining. This kind of system, used in some countries in Europe, essentially creates unions for entire industries. So fast-food workers, say, would all be represented by a single union that would negotiate a baseline contract with the whole industry, from McDonald’s to Taco Bell. This would help address concerns about how unions can work in a more fractured economy and lead to universal or near-universal union density.

But that’s probably not going to happen in the US anytime soon, given that even milder reforms foundered while Congress and the White House were held by Democrats. In the meantime, Andrias said, there are some smaller ways that unions could get a policy boost at the federal level: increase penalties for violating the law, bar employers from permanently replacing workers who strike, eliminate prohibitions on secondary boycotts, and shorten time periods for representation elections (to prevent employers from stalling the process), among other fixes.

Short of that, people could start today by trying to organize their own workplaces. As noted above, it’s not an easy process. It is definitely risky.

But unions are worth it. It’s still weird to write that, nearly two years after I tweeted about lazy workers taking advantage of Vox Media’s union.

But as I dug deeper and deeper into the research, and as I engaged in the actual organizing and bargaining processes, I was repeatedly proven wrong, in large part because I initially focused way too much on the bad examples of unions instead of the good ones. When you stack up all the research and look at the broader picture, though, the net effect of unions — bad examples included — is good for the typical worker.

I hope more Americans go through the transformation that I did. We’d all be better for it.