Thomas Hofeller preached secrecy as he remapped American politics from the shadows. The Republican Party operative, known as the master of the modern gerrymander, trained other G.O.P. operatives and legislators nationwide to secure their computer networks, guard access to their maps, and never send e-mails that they didn’t want to see published by the news media. In training sessions for state legislators and junior line drawers, he used a PowerPoint presentation that urged them to “avoid recklessness” and “always be discreet,” and warned that “emails are the tool of the devil.”
Hofeller did not follow his own advice. Before his death, in August, 2018, he saved at least seventy thousand files and several years of e-mails. A review of those records and e-mails—which were recently obtained first by The New Yorker—raises new questions about whether Hofeller unconstitutionally used race data to draw North Carolina’s congressional districts, in 2016. They also suggest that Hofeller was deeply involved in G.O.P. mapmaking nationwide, and include new trails for more potential lawsuits challenging Hofeller’s work, similar to the one on Wednesday which led to the overturning of his state legislative maps in North Carolina.
Hofeller’s files include dozens of intensely detailed studies of North Carolina college students, broken down by race and cross-referenced against the state driver’s-license files to determine whether these students likely possessed the proper I.D. to vote. The studies are dated 2014 and 2015, the years before Hofeller helped Republicans in the state redraw its congressional districts in ways that voting-rights groups said discriminated on the basis of race. North Carolina Republicans said that the maps discriminated based on partisanship but not race. Hofeller’s hard drive also retained a map of North Carolina’s 2017 state judicial gerrymander, with an overlay of the black voting-age population by district, suggesting that these maps—which are currently at the center of a protracted legal battle—might also be a racial gerrymander.
Other files provide new details about Hofeller’s work for Republicans across the country. Hofeller collected data on the citizen voting-age population in North Carolina, Texas, and Arizona, among other states, as far back as 2011. Hofeller was part of a Republican effort to add a citizenship question to the census, which would have allowed political parties to obtain more precise citizenship data ahead of the 2020 redistricting cycle. State legislative lines could then have been drawn based on the number of citizen voters, which Hofeller believed would make it easier to pack Democrats and minorities into fewer districts, giving an advantage to Republicans.
Other documents show that Hofeller was hired by a Massachusetts Republican who sought to use the Voting Rights Act provision for majority/minority seats to draw a single district containing all of Boston, so that Republicans could make inroads into an otherwise entirely Democratic congressional delegation. Hofeller drew several sets of maps, but the effort went nowhere. Additional files document his work in Mississippi, Alabama, and Virginia, among other states.
E-mails also connect Hofeller to redistricting efforts in Florida. Top Republican officials in the state have denied that they played any official role in drawing the state’s legislative and congressional districts in 2011. A 2010 state constitutional amendment barred partisan gerrymandering in Florida. E-mails show that Hofeller communicated with and visited top G.O.P. political operatives in Florida in 2011. The operatives helped organize or draw state legislative and congressional maps that matched the districts that were later enacted. The operatives insisted, at a trial, that drawing the maps was only a hobby. A Florida judge found that argument unconvincing, concluding that the G.O.P. conducted a stealth redistricting operation that snuck partisan maps into the public process and made a “mockery” of the state’s constitutional amendments.
The files mostly pertain to Hofeller’s work in North Carolina, where he drew—and defended in court—the state’s legislative and congressional maps multiple times, after judges ruled them to be either unconstitutionally partisan or racial gerrymanders. The congressional lines that he helped draw in 2016 were struck down by a federal court as a partisan gerrymander; that decision was vacated this past June, in a 5–4 ruling by the U.S. Supreme Court.
Perhaps one of the clearest and ugliest gerrymanders in North Carolina—or in the entire nation—is the congressional-district line that cuts in half the nation’s largest historically black college, North Carolina A&T State University, in Greensboro. The district line divided this majority minority campus—and the city—so precisely that it all but guarantees it will be represented in Congress by two Republicans for years to come. North Carolina Republicans have long denied that this line, between the state’s Sixth and Thirteenth Congressional Districts, was intentionally drawn to dilute black voting power, which would be a violation of the constitutional prohibition against racial gerrymandering.
Hofeller’s files, though, show that he created giant databases that detailed the racial makeup, voting patterns, and residence halls of more than a thousand North Carolina A&T students. He also collected similar data that tracked the race, voting patterns, and addresses of tens of thousands of other North Carolina college students. Some spreadsheets have more than fifty different fields with precise racial, gender, and geographic details on thousands of college voters.
A spreadsheet named “NC College Voters for ZIP ID” contains voter data for more than 23,100 North Carolina university students, including thousands in Greensboro. The detail for the North Carolina A&T students is precise: students are sorted by residence hall. That means that Hofeller knew which A&T students lived in Aggie Village, on the north side of campus, and which resided in Morrow or Vanstory Halls, on the south side—along with a detailed racial breakdown and information about their voting status. As Hofeller sought to create two reliably Republican congressional districts, his computer contained information on the precise voting tendencies of one of the largest concentrations of black voters in the area.
And if Hofeller cross-referenced that spreadsheet against another included on his hard drive, this one saved as “80 pct College Voters on Non-Match List”—which identified 5,429 North Carolina college students who appeared to lack the necessary I.D. required to cast a ballot at the time he drew the congressional maps—he could have crafted this line along Laurel Street, with even greater specificity about who would and would not be likely to vote.
North Carolina Republicans have admitted that the maps were intentionally drawn to lock in a partisan advantage, but not to harm minority groups. It is difficult to determine from a review of the files exactly how Hofeller used the data on race, but the documents serve as proof that he possessed the data and incorporated it into his decision-making.
His hard drive contained maps of Greensboro with the titles “Greensboro Master Race,” “Greensboro – Pct Blk – City Only VAP,” “Greensboro 45+ BVAP Compactness” and “Greensboro 50+ BVAP Compactness.” (VAP is short for “voting-age population,” and BVAP is an abbreviation for “black voting-age population.”) The “Greensboro Master Race” file contains a color-coded map that shows the city’s black population, age eighteen and older, by precinct.
Two files on Hofeller’s computer —“NC Juducial [sic] 2017 — stats” and “NC Judicial 2017 — Map”—break down the state’s eleven judiciary districts with race data. A PDF of a spreadsheet file shows that mapmakers constructed a file with partisan and racial information, including the percentage of Republican voters in each district, but also a column labelled “% 18+_AP_Black.” Those percentages are then overlaid on a color-coded map by district.
Hofeller also appears to have intensively researched the impact of the state’s voter-I.D. laws. In 2013, Republicans passed one of the strictest voter-I.D. laws in the country, which rejected forms of identification typically used by students, government employees, and racial minorities. It also cut the number of early-voting days by a week and put an end to same-day voter registration during that period. Hofeller created spreadsheets listing the voters, college students, and racial minorities who lacked driver’s licenses.
He appears to have gathered much of this information after the state N.A.A.C.P. challenged North Carolina’s stringent voter I.D. law. Hofeller worked with the Republican legal team that was defending the law. In February, 2014, Hofeller wrote to a Republican attorney about an effort to match North Carolina’s master voter-registration file against the driver’s-license base to see which voters might be affected by the I.D. law. The team also looked to compare that information against that of the closest Department of Motor Vehicles offices.
“After the addresses are standardized and geo-coded for all three files, the next step will be a determination of the distance from each voter address to both the nearest DMV office and nearest in-county early voter center,” Hofeller wrote, on February 4, 2014, according to the letter from his files. “After geocoding and distance computations are completed, the extracted, augmented voter information can be merged back into the full voter database. Then we will be able to perform the required analyses.”
This letter from 2014 also suggests widening the scope of the project to test for citizenship. Hofeller asks, “Do we wish to link up with the DMV to match the list of voters, with state IDs, against the federal citizenship file to check for non-citizens?”
In the letter, Hofeller estimated that the data costs would be around seven thousand five hundred dollars, and he capped his fee at sixteen thousand dollars, or two hundred dollars an hour, for a maximum of eighty hours. “This will include project management and the production of tables and maps,” he wrote, but, noting that the matter could end in court, he wrote that “everyone needs to understand” that giving depositions or attending trial would be billed at three hundred and twenty-five dollars an hour.
Hofeller invoiced the Republican attorney and others, including national Republican organizations, for tens of thousands of dollars between 2011 and 2017. He often applied what he called a “professional discount.” In an August, 2011, e-mail to another Republican in North Carolina, Hofeller wrote, “I would propose billing this as a flat rate of $36,000 for the four-month period—unless you want the billing to show actual hours. If the issue comes up in deposition or court, we will not then be faced with details, but rather with a flat rate of $9,000 per month.”
More than two dozen PowerPoint presentations in Hofeller’s files show how he travelled the nation throughout 2009, 2010, and 2011, instilling the importance of redistricting in conversations with state legislators. He wrote memos to Party leaders warning them of potential issues that could interfere with the G.O.P. redistricting strategy, including one that cautioned against repeating mistakes made in Colorado. “We had to end very promising litigation in Colorado in the middle of the past decade because we ran out of money,” he wrote. “Once again, the GOP is being hampered by lack of litigation funding. One to three congressional districts could be at stake in addition to our ability to control the legislature.”
Hofeller also encouraged Republicans to send public comments to the Census Bureau to urge it to continue to allow prison gerrymandering. The practice, which generally aids Republicans, counts inmates as residents of the community where their prison is located, not where their home address states they reside. But prisoners who are convicted felons are often barred from voting. “The Bureau should not engage in a adjustment of the census counts when the actual effect on reapportionment and redistricting is not clearly known for individual states,” he wrote. “This change is being encouraged by Democratic or Liberal organizations.”
E-mails also show that Hofeller worried that his redistricting handiwork could be undone by an anti-Trump wave in 2016 that handed state legislatures to Democrats. On September 6, 2016, Ryder e-mailed Hofeller a story from The Hill about Democratic efforts down-ballot. Hofeller’s reply: “Yes, maybe our redistricting ace card will be ‘trumped’ to use a bridge analogy.”
In an August, 2016, e-mail to a consultant to California Senate Republicans, Hofeller expressed frustration about Trump’s hold over the Party, and confidence that his maps would survive any blue wave. “Meanwhile the GOP continues to bury its collective heads in the sand or in other higher places. Trump is only a product to this stupidity,” he wrote. “Do not worry about us in North Carolina in terms of redistricting. Even in the coming political bloodbath we should still maintain majority control of the General Assembly. The Governor cannot veto a redistricting map, so the Democrats hope is that the Obamista judiciary will come to their rescue.”
E-mails suggest that Hofeller’s commitment to the Republican cause never wavered. The day after receiving a grim prognosis for lung cancer and a kidney tumor, Hofeller wrote a friend that he didn’t plan to slow down. “I still have time to bedevil the Democrats with more redistricting plans before I exit,” he wrote, on May 21st. “Look my name up on the Internet and you can follow the damage.”