When U.S. senator Ted Cruz mounted the stage on election night in 2018, after narrowly fending off a surprisingly robust challenge by Democratic congressman Beto O’Rourke of El Paso, the metaphors in his victory speech quickly turned martial. Cruz deemed the election, which had invigorated Democrats in a way not seen in decades, a “battle of ideas” and said that he’d turned back “an assault.” He triumphantly declared that “the people of Texas decided this race.”

Strictly speaking, that wasn’t true. If the campaign was indeed a war, it was one defined by its conscientious objectors. Although Texas’s turnout that year was 18 percentage points higher than in 2014, in the previous midterm election, it was still the eleventh-lowest level of turnout in the country. Cruz won 4.26 million votes; O’Rourke, 4.05 million. Both vote totals were dwarfed by the 7.42 million registered Texans who, despite the $115 million the candidates poured into the election, didn’t cast ballots. And that doesn’t count the millions of Texans who were eligible to vote but didn’t bother to register.

In the aftermath of that race, the Texas Democratic party saw opportunity in apathy. Manny Garcia, then the state party’s executive director, told Reuters that when it came to Texas Democrats and Republicans, “there are more of us than there are of them.” Garcia was echoing a persistent Democratic aphorism that maintains that Texas isn’t a red state, but a nonvoting one. It’s an age-old belief that explains the recent strategy of the Texas Democratic party—and of the Beto O’Rourke campaign, which is waging a long-shot bid this year to upset Greg Abbott in his bid for a third term as governor. 

Ahead of the 2020 election, Garcia and other state Democratic leaders believed that if the party could turn out the nonvoters, it could flip half a dozen U.S. House seats, defeat U.S. senator John Cornyn, and gain the nine Texas House seats necessary to control that legislative body. So the party went all in on voter outreach, which had traditionally been left to individual campaigns to coordinate. The party had at least $25 million in its coffers and invested heavily in improving its voter model, which assigned every eligible Texan both a partisanship and a likelihood-to-vote score. Then, Garcia recently told me, leaders decided to ignore the latter rating and deliver mail, texts, and digital ads to every Texan modeled as a Democratic-leaning. The party also focused on voter registration, with officials confidently declaring that of the five million voting-age but unregistered Texans, the model had identified 70 to 75 percent as likely Democrats. 

Curiously, the Republican Party of Texas also saw salvation in the nonvoters. In July of 2019, the GOP, chastened by Cruz’s narrow victory and by the loss of several legislative seats in the suburbs of big cities, assembled a team of election whizzes who’d helped the party in the past before leaving for work in states with more-contested races. Karl Rove, a key architect of Texas’s transformation from a Democratic to a Republican state in the 1980s and nineties, had moved on to the national arena but returned to advise the state party on voter registration efforts. Also returning to his roots was former Texas GOP chair Steve Munisteri, who’d overseen Republican sweeps in 2010, 2012, and 2014 before leaving for the greener pastures of the Rand Paul presidential campaign and then the Trump White House, where he served as principal deputy director of the Office of Public Liaison. Munisteri split his time between assisting Rove’s effort and aiding the campaign of Cornyn, whose favorable ratings were poor. 

Munisteri felt certain that Democrats were focusing too heavily on voter registration at the expense of efforts to get out the already-registered nonvoters—whom political scientists agree are much easier to convince to vote. He also thought Democrats’ confidence in the partisan allegiance of the unregistered was wildly inaccurate and based on incorrect math. (After accounting for noncitizens and felons who are not eligible to cast ballots, and weeding out duplicate entries on the Secretary of State’s list of the unregistered, he felt there were half as many eligible unregistered Texans as the Democrats believed there were—and thus way fewer Democratic-leaning ones.) “They just don’t understand the numbers or haven’t done the research,” Munisteri recalled thinking when he read the Democratic party’s press releases. (Responding to Munisteri’s criticism, Garcia told me, “In a state suffering under one-party rule for so long, it does no one any good for the [Democratic] party to play the role of a cautious pundit.”)

The GOP poured resources into getting registered voters to the polls, outspending the Democrats more than tenfold on turnout efforts. The Cornyn campaign alone targeted 3 million of the 7.5 million voters on the rolls who hadn’t turned out in 2018, and whom the GOP had modeled as likely Republicans. It contacted each an average of twelve times until it had confirmation they had voted. When it came to the unregistered voters, Republicans, especially those leading the Cornyn campaign, set out with a lethal combination: much bigger pocketbooks than the Democrats and a much more modest goal to register about 100,000 GOP-leaning Texans. The party set up registration efforts outside gun shows, Trump rallies, and other places where large majorities of attendees would lean Republican. Collectively, those efforts exceeded expectations and led to the registration of 212,942 Texans. (In Texas, voters don’t register by party, so the partisan allegiances of those the GOP registered are unknown. But given where it did its registration outreach, the vast majority were likely Republicans.)

The 2020 election set turnout records. More than 11 million Texans voted, the highest raw total ever, and the 67 percent turnout was the highest seen since 1992, when George H. W. Bush and Bill Clinton topped the ballot. Munisteri was impressed by the Democrats’ turnout, especially given that the party had not done any door-to-door canvassing during the COVID-19 pandemic: Joe Biden garnered 1.38 million more votes in Texas than Hillary Clinton had in 2016. But the GOP had gotten out the vote even more successfully, with Cornyn winning more votes than Donald Trump and Republicans holding onto all of their U.S. House seats and control of the Texas House and Senate. 

While Democrats had registered 150,000 more Texans than the Republicans—and spent three times less per registration—the GOP got 26,000 more new registrants to actually vote for its candidates. According to Democrats’ models of voter likelihood, the GOP was more successful at getting both regular voters and low-propensity voters to the polls. Republicans also did a better job than Democrats did of turning out the Black and Hispanic voters who were likely to favor their party. 

In the face of these results, and after losing every statewide race for the past 28 years, the Texas Democratic party responded with business as usual. This year, O’Rourke is once again proclaiming that “Texas is not a red state” and that nonvoters will turn it blue. 

Ahead of the election, I wanted to investigate that theory and find out who the state’s nonvoters are and what they believe. Political polling doesn’t much care for the views of those who eschew the democratic process, so I traveled Texas to meet them, starting in El Paso, the lowest-turnout major city, and heading east to hit four of the state’s ten lowest-turnout counties. 

At the entrances of dollar stores and grocery markets, I greeted dozens of folks who said they wouldn’t be voting this year. I met dozens more by violating no-talking rules in public libraries and by risking political conversations in dive bars. The majority of those who told me they weren’t planning to cast ballots were registered to vote; a larger majority said they had no idea an election was scheduled for November 8. From conversations, it was clear that Democrats are partially correct: Texas is a nonvoting state. But there are a lot of Republicans who don’t cast ballots, too. And when it comes to more-liberal nonvoters, the Democratic party is not yet well positioned to reach them. 

Every election has a prevailing issue, or two, that will drive the bulk of voters to the polls. But the reasons for staying away from them are often quite eclectic. Take the cashier in Pecos, in the heart of the West Texas oil patch, who told me she wasn’t going to cast a ballot because she’d heard the saying that “one vote can make a difference” and worried she’d make the wrong decision. Or take the retail clerk in Marfa, the artsy liberal enclave and seat of Presidio County, where turnout was only 46 percent in 2020, who told me he didn’t care about the election and just wanted customers to stop complaining about the rising prices of goods on his store’s shelves, whatever the result. 

Despite the quirks among nonvoters, they share certain characteristics. Nationally, they are disproportionately between the ages of 18 and 29, and they are less white, less educated, and less highly paid than voters. Fifty-one percent of nonvoters in 2016 had not attended college, versus 30 percent of voters; 56 percent of nonvoters earned less than $30,000, versus 28 percent of those who had cast ballots. While non-white Americans made up only 25 percent of national voters in 2016, they made up 48 percent of nonvoters. (There isn’t great data from Texas specifically, but political scientists told me the state mirrors national trends. It’s easy to see how: in 2020, all but one of the 25 lowest-turnout counties in Texas had populations with Hispanic majorities.) 

Those demographics might seem to scream Democrat, as many in the party were apt to tell me. And GOP elected officials certainly act as if they agree. Despite evidence that Republicans can win high-turnout elections, the Republicans who dominate the Legislature have made what they call “election integrity” (and what civil rights groups call voter suppression) a priority in the two sessions since O’Rourke came uncomfortably close to unseating Cruz. The Legislature passed extreme voter restrictions last year even after winning nearly everything in 2020. The state is one of twelve where residents cannot register to vote online, and independent analyses have ranked it among the five hardest states in which to cast ballots. “The proof is in the pudding,” Gilberto Hinojosa, the chair of the Texas Democratic party, told me. “If Republicans really, sincerely believed that nonvoters would vote for Republicans, they’d do everything possible to ensure that every single one of them would go vote.”

Still, there are good reasons to doubt the demographic assumptions leaders in both parties make. As Daron Shaw, a University of Texas political scientist, told me, while nonvoters here have a demographic profile that might seem Democratic, “the dominant characteristic of nonvoters is their psychology: they’re just not that interested; they’re not that engaged; they’re not that involved in politics.” 

And with less involvement in politics come less-defined partisan allegiances. By way of illustration, Shaw pointed to North Carolina in 2008—perhaps the prime example of a successful turnout effort by Democrats—where the Obama campaign found and engaged enough low-propensity Black voters to flip the state blue for the first time since Jimmy Carter in 1976. Shaw noted that even then, while Black regular voters were modeled as 85 percent likely Democrats, that number dropped to the seventies for low-propensity Black voters. And four years later, in a higher-turnout election, Obama lost North Carolina. “When nonvoters do become interested, when they do get engaged, they tend to be disproportionately responsive to whatever the short-term forces are politically,” Shaw said. 

Traveling the state, I met many Texans who helped underscore the thesis that nonvoters, even if they seem likely to be Democrats, might not reliably be. There were those who favored liberal policies in the abstract—a $15 minimum wage, say—who nonetheless thought the Democratic party delivered them nothing worth voting for. One proud liberal, a barista in Plainview, told me he was not going to vote because his candidates don’t win and he didn’t want to waste his time waiting in line. Many others—whether a Black grocery-store employee in a red enclave in the South Plains or a young Latino in a blue city—told me that even the politicians they liked, and who won, never did anything to improve their lives. This was an especially common opinion among once-Democratic voters, who made clear they are no longer voting for a reason. 

Outside a Dollar General in El Paso, where turnout in 2020 was 54.6 percent and as recently as 2010 was below 25 percent, I met two Hispanic government employees who told me they won’t head to the polls in November. Mario, a 27-year-old, said he hadn’t voted since casting a ballot for Obama in 2012. He could list some of Obama’s legislative achievements that he supported, including the passage of Dodd-Frank, which imposed new regulations on financial firms following the 2007 Wall Street crash. But he said nothing is being done for his wages and well-being anymore, and he confidently declared that he wouldn’t cast another vote. “I just believe that none of the candidates offer anything and actually go through with it.”

A few minutes later, I met Rick, a 41-year-old veteran who used to work for the city, who said his wages had stagnated despite local elected Democrats promising prosperity. “There’s no real change that I’ve seen.” He said that even on other policy matters, government wasn’t addressing his needs, pointing to disaster-control efforts after deadly flooding in June 2021. “They say, ‘Oh, there’s [a solution] for all the flooding.’ [But] it still floods. I don’t know where all the money goes to.” He told me the only reason he’d go to the polls anymore is to vote against any candidate threatening to cut Veterans Affairs funding.

There were also plenty of nonvoters I met who helped explain why Rove and Munisteri see such opportunity for the GOP in that cohort. Victor, a 27-year-old Hispanic former voter in Plainview, the seat of Hale County, where 75 percent of voters went for Trump, wasn’t planning on casting a ballot and told me he would never vote for Democrats because of their stance on abortion. Tim, a compressor operator in Pecos, told me that Democrats “have had a lot of good stances that are good for the general public: corporate taxes, unionization, and in a kind of a backwards way, race issues.” But he said O’Rourke had lost him. “The whole take-your-guns thing, it’s too intense.”

What little polling there is on nonvoters indicates that a slim majority nationally are Democratic-leaning, but not nearly by the margins many expect. To get a sense of what the numbers look like in Texas—which is more conservative than the country at large—I chatted with dozens of local election officials, pollsters, political scientists, and political operatives of both parties. There was some bipartisan consensus on what percentage of nonvoters are Democrats. In a 2020 postmortem, the Texas Democratic party projected, based on its modeling, that 51 percent of the voting-eligible population was Democratic—which, given the actual 2020 margins, means the party’s candidates would have expected to take around 56 percent of nonvoters that year. That aligns pretty squarely with GOP operatives’ assumptions based on the RNC voter file, which ascribes scores to every voter based on thousands of consumer data points, a method pioneered by Rove. (Is this voter a member of the Sierra Club? Does she subscribe to Garden & Gun? Does he shop at Whole Foods or Cabela’s?) 

Despite these GOP innovations, neither party has demonstrated a consistent ability to get its slice of the nonvoters to the polls. In 2019, Shaw and John Petrocik, a political scientist at the University of Missouri, analyzed presidential, congressional, and gubernatorial elections in all fifty states from 1966 to 2018. In their book The Turnout Myth, they found that “there is no axiomatic connection between the turnout rate and the partisan outcome of the election.” In crude terms, Democrats have won plenty of low-turnout elections (think Jimmy Carter in 1976, or Bill Clinton in 1996, the lowest-turnout presidential election of the last ninety years). And the GOP has won plenty of high-turnout ones (such as George W. Bush’s victory in 2004, or Greg Abbott’s 2018 triumph in the highest-turnout gubernatorial election in Texas in at least half a century). 

Shaw acknowledged that the Supreme Court decision overturning Roe v. Wade might be a short-term factor that motivates typical nonvoters to go to the polls for Democrats this year (although the early data indicate that it has not had an outsized effect on voter registration in Texas.) But he doubted that relying on nonvoters, who are notoriously fickle, election after election is a long-term solution for the party. 

After their losses in 2020, Democrats might have determined they couldn’t again bank on nonvoters. A working definition of insanity is doing the same thing over and over again and expecting a different result. But the party evidently favors a different mantra: “If at first you don’t succeed, try, try again.” The Texas Democratic party has modestly adjusted its estimate of unregistered likely Democrats down from 3.5 million to 2 million and has announced they are key to the party’s path to victory this year. “Texas Democrats’ takeaways from the 2020 election are clear,” a party press release from last summer reads: “to take back our state from Texas Republicans, Democrats need to register more voters.”

Rove and Munisteri have stayed in Texas to run their sophisticated registration effort for the GOP. Meanwhile, the Democratic party has continued to raise money for its modeling—though some donors were scared away by the wholesale defeat in 2020—and leaders are excited by resuming door knocking after putting it on hold during the pandemic. Staffers for O’Rourke, once again the party’s standard-bearer, told me they are also using their own model—one more similar to the GOP’s, based on electoral history, individual data, and extensive surveys of nonvoters in the state. “The raw math suggests that there are enough of these low-investment voters to win Texas,” Jason Lee, the Beto for Texas deputy campaign manager, told me. “We like to say we can never run out of those targets, we can never get to them all, because there’s so many of them.” 

Probably the only silver lining to decades of losses by the Democrats is that there are many experienced hands to pitch in to assist campaigns in trying to reach all of them. Indeed, losing an election as a Democrat in Texas has become a sort of internship for launching a registration or turnout operation. Cristina Tzintzún Ramirez, a labor organizer who lost in the Democratic primary for U.S. Senate in 2020, now leads NextGen America, a progressive nonprofit trying to mobilize the youth vote. Between running for president and governor, O’Rourke launched Powered by People to do the same. Two-time losing U.S. House candidates Julie Oliver and Mike Siegel teamed up last year to head Ground Game Texas, an organization testing whether progressive ballot initiatives such as minimum-wage hikes or marijuana decriminalization—which tend to perform better on the ballot than progressive candidates do—can drive turnout for Democrats. (The fact that these policies outpoll the Democrats who support them might also be evidence that the party’s brand is underwater, especially in Texas.)

Party leaders say they are running a “coordinated campaign” this year, with candidates sharing data and others outside the apparatus assisting with voter registration. But coordination requires many willing participants, and even with the host of new organizing efforts, the Democratic turnout infrastructure across the state is weak. In the bluest parts of the state, extreme gerrymanders—which Republicans were again able to pass given Democrats’ failure to take the Texas House in 2020—have reduced candidates’ need to find new votes to win. Many elected representatives don’t even try to drive turnout, as numerous unsuccessful Democratic candidates, including former state senator Wendy Davis, the party’s nominee for governor in 2014, told me. Indeed, eleven of the twelve lowest-turnout U.S. House districts in Texas in 2020 were all Democratic strongholds. (Democrats only won thirteen U.S. House seats in the state that year.)

And in traditionally red areas of the state, once O’Rourke’s tour bus leaves, there often isn’t anyone to follow up with voters his campaign might target. In 55 counties, the Texas Democratic party does not have a local chair to run a door-knocking campaign. Texas Monthly reached out to the chairs of the other 199 counties in Texas, and heard from fewer than forty who are leading door-to-door canvassing efforts this election cycle. 

In the South Plains, I met with Stuart Williams, the former Democratic chair of Lubbock County, who described his frustration at the state party’s lack of investment in outreach campaigns in West Texas. Last year, Williams, a gay, Black grocery-store employee, cofounded the 134 PAC, dedicated to building the party in rural West Texas counties where it doesn’t exist. He praised what O’Rourke is doing, but pointed to 2018 as an example of how visiting every county isn’t enough. That year, O’Rourke won Texas’s thirty most populous counties by 5 percentage points. In the other 224, all of which he visited, he lost by 47. Williams said that candidates merely showing up in election years won’t be enough to convince folks to vote for Democrats after years of the party becoming tainted as one full of “baby killers” who oppose the oil and gas industry. “It’s like 1964 in Mississippi, during Freedom Summer,” he said, referring to publicly being a Democrat in rural West Texas. “You could lose business, lose friends. You could be ostracized from your community, ostracized from your church.”

O’Rourke’s campaign managers told me they weren’t concerned about outreach, despite the challenges, because they have more than 100,000 volunteers. But I wanted to see what voter outreach really looked like in some rural counties, so on the Monday before Labor Day, when campaign season typically goes into overdrive, I traveled to Pecos. The town of under 13,000 is the seat of Reeves, the seventh-lowest-turnout county in Texas. (The county has a Democratic chair, but she did not respond to requests for an interview.) Before Trump carried the county handily in 2020 in an election in which local turnout only hit 49 percent, Reeves had not voted for a Republican presidential candidate since 1984. In fact, the county had been so blue, Kellie Nagy, the elections administrator, told me, that in most elections there was no Republican on the ballot, but this year there’s one for county clerk. 

While a historically blue county that went red in a low-turnout election might seem the type of place the Democratic party would target, many nonvoters I talked with in Pecos didn’t even know there was an active gubernatorial race underway. At the county library, stacks of primary-election voter guides remained on a credenza beside other community brochures five months after the primary. The receptionist, a student named Mariana, told me she was on the “Democratic side of things” and that her biggest concern was gun control. But she wasn’t planning to vote. She didn’t know much about the candidates, she said, and no one had reached out to inform her.