Victory Lab

A New Job for Labor

Citizens United didn’t just permit corporations to spend freely on elections; it opened the door to unions, too. Can the AFL-CIO get Obama re-elected?

National AFL-CIO President Richard Trumka speaks during a rally in opposition of the Wisconsin Gov. Scott Walker's bill that threatens collective bargaining rights February 18, 2011 in Madison, Wisconsin.

AFL-CIO President Richard Trumka speaks during a 2011 rally opposing Wisconsin Gov. Scott Walker’s bill threatening collective bargaining rights

Photo by Mark Hirsch/Getty Images.

Last Friday, at the hour when most Wisconsinites would be leaving work, a union canvasser named Chad Pichler was knocking on doors in a neighborhood on Milwaukee’s western edge.* Pickler was two hours into his daily shift. He had already successfully contacted 15 voters, asking each how they planned to vote in the following week’s recall election and recording the results on an iPad that synced the data with the AFL-CIO’s computers in Washington, D.C.. Pickler, 38, wore a black hoodie, along with stone earrings and a lip ring bisecting the right side of his smile. He had a laid-back way of introducing himself to voters that initially suggested amateurism, but demonstrated a deft way of returning to questions that a voter deflected on the first pass—and he displayed what passes for institutional knowledge in the world of the professional door-knocker, like rattling a suspicious gate to draw out any dogs that could be hidden on the other side. “I’ve been doing this since ’92,” Pickler said.

He mounted the steps of one house, rang the bell, and asked for Alex. Alex was a tall man in a blue janitor’s jumper with a gleaming shaved head that revealed a scar along one side. He was surprisingly fit for what Pickler’s iPad told him was a 74-year-old. Pickler said he was taking a poll: What issue mattered most to Alex? The man began to respond, but quickly stalled; his face showed puzzlement, then frustration, then despondence. “I just had brain surgery and I can’t get the words out,” he said.

Pickler struck a reassuring tone and slowed the pace of his speech. He asked Alex about whether he supported Scott Walker or Tom Barrett in the recall election, and there the answer came quickly: Like nearly everyone Pickler had talked to that day, Alex supported Barrett, Milwaukee’s mayor and the Democratic candidate for governor. “I’m with the AFL, but I’m retired,” Alex said proudly.

“We’re with the AFL,” Pickler said, leaving a photocopied flyer about Barrett he suggested Alex could share with his children. Pickler descended the steps and looked at his iPad for the next address on his route, but continued talking about Alex. “Milwaukee’s got that old union history—socialist history—so a lot of the older guys I talk to are more political than most,” he said.

For decades, however, the AFL was prevented from speaking with men like Alex about politics. The Taft-Hartley Act that restricted corporations from trying to influence elections had been written to constrain labor unions, ensuring that they could use general treasury funds only for communication with active members and their families. Unions created affiliated PACs and other vehicles to reach the broader electorate, but the bulk of their campaign activity was concentrated on member households—in recent decades, a perpetually shrinking pool of targets.

But the Supreme Court’s 2010 Citizens United decision removed those restrictions, permitting unions to spend freely on campaign activity. Some Republicans claimed that, as a result, the decision could ultimately tip the balance of power between labor and industry in the Democrats’ direction. “Unlike corporations, unions are far better positioned to take advantage of the ruling because they have virtually no other restraints on their capacity to engage in political action,” wrote Steven J. Law, a former Chamber of Commerce official now serving as president of the American Crossroads super PAC directed by Karl Rove, in the Wall Street Journal after the ruling. Union leaders scoffed at that suggestion, arguing that they could never keep up with unbridled spending by business interests. “Look, we don’t think of it as a good decision at all,” says Mike Podhorzer, the AFL’s political director.

But for all the dystopic visions conjured by those on the left about the post-Citizens United landscape, Podhorzer does allow that in legal terms the decision marked a return to terrain where organized labor had once been at its most politically fearsome. “For unions up until 1947, the kind of organizing we’re trying to recreate was norm—working folk talking to one another in their neighborhood about politics and issues. Taft-Hartley made it illegal for us to do that,” says Podhorzer. “But until the late ’40s that’s what made Republicans and conservatives very concerned about the influence of unions.”

Tuesday’s result in Wisconsin, where Walker won by a seven-point margin, does not augur well for labor’s ability to defend its state-level policy interests or exact electoral retribution against its enemies. Labor strategists have attributed their loss solely to the fact that Walker’s campaign outspent Barrett as much as tenfold, the ratio shrinking only slightly when spending from outside groups including unions is included. That conclusion seems a bit fanciful: Much of the money was spent in an effort to persuade an electorate that, exit polls confirm, was already highly polarized and had made up its minds about the recall.

But labor is right to conclude that if it is going to be an operational counterweight to conservative committees funded by large checks from corporate interests or wealthy individuals—especially in races with a more fluid, persuadable electorate—it will not come through parity in media spending. Labor will have to find a way to win in the field.

So as AFL strategists begin to analyze the results in Washington their goal will not be to catch up with the opposition in terms of spending as much as to widen the gulf between how the two sides practice politics. How much can organized labor recreate its culture from the 1940s, with new tactics for an era of both mass media and high-tech niche communication? Any such quest for a 21st-century comparative advantage will probably come down to a deeper test of the value of street-level expertise in campaigns. Is there anything in politics money can’t buy?

* * *

Last fall, as Ohio prepared to vote on Senate Bill 5, the ban on collective bargaining pushed by Governor John Kasich, the AFL-CIO’s political department decided to run an experiment. In 2003, the AFL had launched Working America, which was in legal terms a union, even though it had no collective-bargaining ability. Working America would target swing-state neighborhoods rich in the type of working-class, culturally conservative white voters who had fallen not only out of the union’s ranks but were slipping from the Democrats’ grasp, as well. Some may have been former union members, or grew up in union families, or worked in private-sector industries that had never been unionized. “There are people who have good associations with unions from growing up, but over time—with talk radio—they’ve moved away,” says Peter Drummond, Working America’s Wisconsin state director.

Canvassers would knock on doors and ask residents if they wanted to join Working America. Anyone who signed a piece of paper became effectively an AFL member, eligible to be included in the union’s political programs. With operations in a dozen cities, Working America grew to 3 million members. It became the sphere through which the AFL could do traditional union-style organizing—heavy on direct voter contact like phone calls and door-to-door canvasses—in a cleverly expanded universe of sympathetic voters who were not dues-paying members.

With the Citizens United decision in 2010, the AFL could use its methods to speak to anyone—though their opponents could, too. But thanks to initiatives like Working America, the AFL was ahead of the competition. The union already knew a lot about how to talk to its members outside of mass media: Research had repeatedly found that they were more responsive to mail from a local than a national or an umbrella federation. The AFL had also deployed some of the earliest experiment-informed programs to understand which messages and themes most successfully changed members’ minds. In Ohio, a field experiment demonstrated that labor knew how to get its targets to the polls: Those who received at least one contact from the AFL turned out at a rate 18 points higher than those who received none. Three contacts boosted turnout by just under 30 points.

At the same time, the AFL was using Ohio to test a substantive departure from standard political practice. The difference between media and field operations—the so-called “air war” and “ground game”—is typically not one just of venue, but of content and purpose. Since the rise of television and radio advertising, campaigns have largely relied on the air to make their arguments, a setting where they fully control the message and presentation. The ground has been for vote-counting: individual interactions with voters at doors or over phone lines to identify their preferences and then, in the days before the election, to remind and mobilize those who had declared themselves supporters (or otherwise looked like they might be). In these contexts, persuading the undecided was usually a secondary concern.

In Ohio, the AFL’s political department concluded, Working America would depart from that two-track model. Instead it would deliver its message through face-to-face encounters, developing a basic canvassing script and training staff to make the case against SB5.

Working America identified a pool of Ohio households that the AFL’s statistical models profiled as likely to be friendly to the argument, and then randomly assigned canvassers across it so that there would be a control group that went unvisited for comparison. The AFL polled across the groups before and after to measure the effect those visits had on residents’ opinions. The targets who answered the door were 14.7 percentage points more likely to oppose SB5. The campaign had an equal impact among the general public as it did moving union households and Working America members. Its greatest sway, in fact, was over the most conservative voters.

On election day, SB5 was overwhelmingly rejected by Ohio voters, and AFL officials emerged from the experience confident that they could use face-to-face canvassing to change minds even as millions were being spent by both sides on television ads. They anticipated further experiments in low-profile contests, like the following June’s mayoral race in San Diego, to develop a strategy for competing in the first presidential election after Citizens United. But in January, union organizers in Wisconsin succeeded in another show of force by gathering a staggering number of signatures to force a vote to recall Walker. The AFL’s testing ground would not be a sleepy local vote but an election observers swiftly designated the year’s second-most important.

* * *

The imposition of what Podhorzer labels “the Taft-Hartley handcuffs” happened to coincide with the rise of television advertising, which within decades came to dominate political communication. Organized labor could not turn to the airwaves as reflexively as candidates or parties did, and so it became good at the mechanics of political fieldwork because legal requirements forced it to be disciplined about targeting and contact. Even before the rise of databases and statistical modeling, unions had to master voter lists to pick out their members and know which doors they were allowed to knock on and which numbers to call. Geography eased the physical demands: Union members were often consolidated into urban precincts that could be manageably walked.

Walker’s financial edge over Barrett manifested itself primarily on television. There was plenty of grassroots activity on the right, thriving on enthusiasm for Walker, but it rarely took the form of visiting voters at their homes. The 22 field offices administered in partnership between the Republican National Committee and the Republican Party of Wisconsin were run almost exclusively as phone banks to identify supporters and then remind them to turn out. Only in a few exceptional cases, in conservative suburbs with a friendly walking density, did the party deploy volunteers to do the same work in person. “With the sheer volume of supporters we have, you can make more phone calls in an hour than you can do in knocking on doors,” says state-party spokesman Ben Sparks.

One could see a different approach at the Madison Labor Temple, the recall supporters’ primary staging area in the capital. When Barrett started drawing attention to a Walker plan to privatize state lands, an AFL data analyst matched a publicly available list of Wisconsinites with hunting and fishing licenses to the union rolls, and had phone bank volunteers deliver a special persuasion message to them—even though many might have been too conservative to fall into labor’s usual pool of targets. Outside the temple, the new, post-Citizens United reach of these tactics were visible: Canvassers who were dispatched across liberal Dane County no longer had to skip houses without union members, an adjustment that would improve the efficiency of door-knocking operations and likely open up new neighborhoods for focus. “The doors are better just because you’re more likely to actually speak to someone. People just don’t answer the phone these days—damn caller ID,” said Katheryn Burns, a K-1 teacher in Madison who had been working on the recall on nights and weekends. “Once you get people at the doors they’re pretty friendly.”

The newest component of the AFL’s program, however, was one that had no face-to-face component at all. This spring, the AFL launched its own super PAC, Workers’ Voice, the center of which was an online platform called Friends and Neighbors. Workers’ Voice bought online ads inviting liberals to enlist as volunteers, at which point they were asked to link their Facebook accounts to Friends and Neighbors. The AFL’s proprietary code trawled through a volunteer’s social network, looking to match up names to the roster of Wisconsin voters whom the AFL’s microtargeting models had identified as targets for its campaign contact. Then volunteers were given the chance to call their acquaintances directly from the site. “The hope is that it is going to be a resource for the economically progressive community, and to be a way to reach people who are young and don’t have much exposure to unions,” says Podhorzer.

In terms of software, none of this is novel: Millions of volunteer calls were placed through Barack Obama’s website in 2008, and the RNC just launched a home-calling tool as part of its Social Victory Center. Others have developed the ability to match Facebook accounts to a voter file and classify friend networks based on their state or district, party, or other demographic categories. But Workers Voice represented a radical new role for the institution behind it: The ideal interaction, from the AFL perspective, was a liberal activist with no labor ties calling friends who also were not union members.

Even if it resembled a high-tech upgrade to Podhorzer’s Edenic pre-Taft-Hartley America, this vision of the union as a facilitator of interactions between individuals with whom it had no economic relationship represented a major departure from the culture of control and self-interest that often exists around union activity. Still, Workers’ Voice was far from a free-for-all: The AFL’s microtargeting selected which friends a volunteer would be asked to call, depending on its campaign strategy. All the data collected about one’s social network and generated through the conversations, including preferred phone numbers, were fed back in to the AFL’s databases to refine its future projections and direct door-knocking operations. Soon the data generated in Wisconsin will be available for analysis in Washington, allowing labor leaders to more keenly attribute at least some blame for Barrett’s dismal performance.

A lot of things could have gone wrong. Labor officials were cheered to see exit polls showing that one-third of voters came from current union households, a big jump from the 2010 rate, but Barrett won only 62 percent of their votes. Labor’s aggressive persuasion efforts could have had a backlash effect, or the unions’ targeting could have simply missed its mark, helping to mobilize the wrong voters. (What if those hunters and fishers cared less about Walker’s view on public lands than his deficit-cutting plans and the AFL delivered an inadvertent reminder to vote for him?) In a highly polarized, emotional environment, voters may have been wary of sharing an unpopular opinion in face-to-face encounters. Indeed, union representatives went into election day with what now seem to be potentially inflated assessments of their support: Mark Hoffman, the business manager for a Madison-area local of the International Brotherhood of Electrical Workers, said just hours before the polls opened that he had spoken to all 857 of his members—and that all but two told him they were voting for Barrett.

The AFL left Wisconsin with reams of new data on the effectiveness of those various programs, although Podhorzer says that the institution’s changed role will force him to refine new metrics for measuring success in political communication. The experimental methods used in Ohio were based on clinical trials that isolated voters to discern the effects of specific top-down communication. An effort based on pulsing ideas through existing social relationships would require the tools of epidemiology. “The premise of what we’re doing is that it’s built the way the world works,” Podhorzer says. “The way the world works is that people live in networks and not in isolation.”

The AFL may adjust its tactics after Wisconsin, but also conclude that it was a massive strategic failure that led to the embarrassing loss. As many Democrats said quietly for months, labor may have picked the wrong fight. It was unable to rally around a single candidate and wasted its energies in a contested primary. By the time Barrett became the nominee a month ago, the majority public opinion may have so stiffened in its skepticism of the recall—and a fundamental sympathy to Walker’s predicament—that no amount of canvassing or friend-to-friend phone calls could have softened it.

Correction, June 9, 2012: This article originally misspelled AFL canvasser Chad Pichler’s last name. (Return to the corrected sentence.)