Ga. State Conference of the NAACP v. Georgia

MARTIN, Circuit Judge *1359In this action, the Georgia State Conference of the NAACP and certain black voters1 ("plaintiffs") challenge the 2015 redistricting of Georgia House of Representatives Districts 105 and 111. Plaintiffs say these Districts resulted from unconstitutional gerrymandering2 based on the race of voters. Doc. 1 ¶¶ 1-4, 20-25. This Order addresses plaintiffs' request that we enjoin the State from holding elections in Districts 105 and 111, as defined in Georgia Act No. 251, 2015 Ga. Laws 1413 ("H.B. 566"). Doc. 103. After careful review, and with the benefit of oral argument, we deny plaintiffs' motion for a preliminary injunction. I. BACKGROUND The Georgia Constitution says that the electoral districts for members of the Georgia House of Representatives "shall be changed by the General Assembly as necessary after each United States decennial census." Ga. Const. art. III, § 2, ¶ 2. In keeping with our Constitution, the Georgia General Assembly redrew state House of Representative districts in 2011. Compl. ¶ 39 (Act No. 1EX). Then in 2012, the General Assembly modified the House district map again. Id. ¶ 40 (Act No. 277). The 2011 and 2012 redistricting plans were precleared by the United States Department of Justice, as required by the Voting Rights Act of 1965. Id. ¶ 41. The 2015 plan at issue in this case was adopted after the Supreme Court's decision in Shelby County v. Holder, 570 U.S. 529, 133 S.Ct. 2612, 186 L.Ed.2d 651 (2013), and was not precleared. Compl. ¶ 42. This lawsuit challenges the third redistricting done after the 2010 census, which was enacted into law in 2015, by way of H.B. 566. The motion for preliminary injunction before us seeks to enjoin only two House districts redrawn by H.B. 566: District 105 and District 111. District 105 is in Gwinnett County. District 111 is in Henry County. These Districts were redrawn in 2015, when the racial makeup of the area was changing. State employees charged with reapportionment brought to light the "changing demographics" in Henry and Gwinnett Counties. Deposition of Gina H. Wright ("Wright Dep."), Doc. 112 at 23:25-25:6; Deposition of Dan O'Connor ("O'Connor *1360Dep."), Doc. 111 at 220:2-221:24; Deposition of Randall O. Nix ("Nix Dep."), Doc. 124 at 178:12-181:24. One person interested in the "changing demographics" was Dan O'Connor, who works in the Georgia Legislative and Congressional Reapportionment Office. In August 2014, Mr. O'Connor sent an email to Representative Chuck Efstration of House District 104 noting that "white registration in [Gwinnett County] between Jan 2000 and this month dropped by about 3,000 while black registration in Gwinnett has quadrupled from 22,443 in 2000[ ] to 96,553 in the latest count." Doc. 103-3 at 1. In a February 2015 email to Speaker Pro Tem Jan Jones, Mr. O'Connor noted that 2014 population data showed both District 105 and 111 to be "at least 35% black in voter registration." Doc. 103-4 at 1. Mr. O'Connor continued, "Generally, once a district gets in the 30-35% black range, it becomes more of a target for Democrats." Id. This is because voting in Georgia is highly racially polarized. Expert Report of Jowei Chen, Ph.D. ("Chen Report"), Doc. 63-1 at 4. Districts with large black populations are likely to vote Democratic. See Doc. 103-17. Georgia House Districts 105 and 111 were political battlegrounds before the 2015 redistricting we consider here. In 2014, Representative Joyce Chandler, a white Republican, won the District 105 election with only 52.8% of the vote. Doc. 103-87 at 3. Representative Brian Strickland, also white and Republican, won District 111 with only 53.1%. Doc. 103-37 at 6. This made Districts 105 and 111 two of only three House Districts in Georgia's 2014 election where Republicans won by a margin of ten points or less. Id. at 1. And with the "changing demographics" in the counties where these districts lie, there was good reason to think the 2016 elections would be even closer. Both Representative Chandler and Representative Strickland went to Gina Wright, who served as the Executive Director of the Reapportionment Office, to enlist her help. Wright Dep. at 23:5-24; Deposition of R. Brian Strickland ("Strickland Dep."), Doc. 127 at 126:20-127:11. Ms. Wright's office is located in the same office building that houses Georgia's legislators, but Representative Chandler visited Ms. Wright both at her home and her office. Wright Dep. at 23:12-15; Nix Dep. at 58:25-60:1. Both Representatives Chandler and Strickland also approached Representative Randall Nix, chair of the House Reapportionment Committee, to express their interest in redrawing the lines of the districts where they had been elected, to increase their likelihood of being reelected. Nix Dep. at 139:16-141:18, 163:11-164:21. Reapportionment Office staff worked with the Representatives to redraw their districts. The Reapportionment Office uses computer software known as Maptitude to develop redistricting plans. Declaration of Gina H. Wright ("Wright Decl."), Doc. 137-1 ¶ 7. Maptitude shows the detailed effects of any given redistricting option. For example, Maptitude can display information about the performance of the Republican or Democratic parties, based on the election results in any given precinct. Id. ¶¶ 7-8. Indeed Maptitude shows partisan data on a street-by-street basis within a precinct, even though it only has partisan data for the precinct as a whole. Thus, Maptitude's street-by-street political data is nothing more than the precinct-wide party-affiliation percentages assigned to streets based on the number of people who live there. Id. ¶ 9; Wright Dep. at 106:7-13. Maptitude can also display data about the *1361race of the people living in a given area. But since the race data comes from the census (as opposed to election results), it is more detailed, and Maptitude can display precise information about people's race down to the street or block-level in any precinct. Chen Report at 32-33. Any Maptitude user can choose what data to display, using "the pending changes box" as she is working on a map. Wright Dep. at 105:10-16. This record, taken as a whole, shows the redrawing of Districts 105 and 111 was a group effort. Representative Nix described one meeting in which Ms. Wright and all the potentially affected legislators sat together and used Maptitude to review options for redrawing the districts. Nix Dep. at 143:3-157:8. Representative Nix testified that as Ms. Wright was "clicking around" different options for redrawing the map, data about the race and political affiliation of the voters in those districts were displayed on the screen. Id. at 148:8-12, 150:8-12, 151:2-9. He also said the meeting he attended was one of several held by Reapportionment Office staff and interested members of the House of Representatives to work through the redrawing of Districts 105 and 111. Id. at 155:18-156:13. And while Mr. O'Connor testified at his deposition that he had no involvement in the 2015 redrawing of Districts 105 and 111, O'Connor Dep. at 134:13-20, 140:4-5, Representative Nix said he talked to Mr. O'Connor "probably daily" as the 2015 maps were coming together.3 Nix Dep. at 92:4-12, 93:6-20. The record supports a conclusion that any changes made to Districts 105 and 111 were vetted and approved by interested legislators and Mr. O'Connor. Even so, Ms. Wright says in her declaration that she "alone worked on that portion of the HB 566 (2015) redistricting plan that touched any part of Gwinnett and Henry County, including HD 105 and HD 111." Wright Decl. ¶ 5. In her deposition, Ms. Wright discussed her general process for mapping in reference to the 2011 redistricting process: When I work on a plan, I use the pending changes box, which is a feature in Maptitude that shows you how your numbers change as you select that geography you're selecting.... And you can set and move which fields you want to have in that change box as you go. So for me I usually keep a combination. I like to have the political data there as well as other data, racial data, population data, all of those other things. Wright Dep. at 105:10-16. In her declaration, though, Ms. Wright says she did not keep race data open in her pending changes box while redrawing Districts 105 and 111. Wright Decl. ¶ 10. Instead, Ms. Wright says she considered only partisan and population data. Id. Ms. Wright says she looked at race data only when she finished redrawing the districts "to make sure that [she] did not do significant harm in that respect as well." Wright Dep. 30:6-8; see also Wright Decl. ¶ 22, 40. In the end, the new versions of Districts 105 and 111 were drawn in the legislative *1362office building by some combination of Reapportionment Office staff and legislators. While Representative Nix made much of the fact that no changes were made to districts unless all affected legislators agreed, Nix Dep. at 79:19-80:5, nothing in this record suggests the affected communities had any input. Neither does the legislative process suggest transparency or public engagement. H.B. 566 was introduced to the House on March 5, 2015. Doc. 103-53 at 1. The bill itself contains no maps-simply a list of precincts and numbers.4 See H.B. 566. Representative Nix described the bill best when he said, "[i]f you look at the bill, there's not a whole lot of detail that anybody can look at and get out of that" without assistance from the Reapportionment Office. Nix Dep. at 182:24-183:3. The House Reapportionment Committee voted to approve the bill at the one and only hearing it ever held about it on March 9. Doc. 103-63 at 1; Nix Dep. at 186:9-11. This committee meeting was also the first time legislators, other than those who had been involved in the redrawing process, ever saw the maps. Nix Dep. at 183:12-20. Two days later, the full House unanimously voted in favor of redistricting Districts 105 and 111, when it passed H.B. 566. Doc. 103-53 at 1. Mr. O'Connor prepared a summary of H.B. 566 in which he offered an explanation for the redrawing: "District line changes can be made for a variety of reasons-as some examples, eliminating a split precinct (a precinct divided into multiple districts), reuniting a neighborhood or community of interest, or addressing technical concerns." Doc. 103-64 at 2; see also O'Connor Dep. at 133:7-134:12. But in his deposition, Mr. O'Connor said he did not recall if any of those justifications for redistricting were actually served by the redrawing of Districts 105 and 111. Id. at 134:6-136:4. Similarly, Ms. Wright testified she worked to adhere to traditional redistricting criteria like these in redrawing Districts 105 and 111, including compliance with state and federal law, improving contiguity and compactness, and keeping counties, precincts, and communities of interest together. Wright Dep. at 20:24-21:12. "Further down" on the list of priorities, she said, "would be incumbency." Id. at 21:9-10. But in redrawing Districts 105 and 111, Ms. Wright acknowledged her "objective was to make these districts, if at all possibly anyway, better for these incumbents to get reelected." Id. at 30:1-3. It is true, as Ms. Wright testified, that there are ways in which the new maps of Districts 105 and 111 maintained the traditional principles of redistricting. But more often, the new maps had a negative impact on these principles. For example, the new maps created districts that were less compact; deviated more from the ideal district size; split more municipalities across district lines; and split more districts across county lines. Chen Report at 3, 26-31. The new maps kept the same number of split precincts in District 105 but more than doubled the number of split precincts in District 111, which went from two to five. Id. at 29-30. And while Representative Nix characterized the changes proposed in 2015 as "very minor," Nix Dep. at 62:8, the *1363new versions of Districts 105 and 111 look quite different from the old. An expert for the plaintiffs calculated that nearly 15,000 people were moved in or out of District 105, and over 31,000 people were moved in or out of District 111. Expert Report of Dr. Gerald R. Webster ("Webster Report"), Doc. 62-1 at 3. These numbers are significant where the target population of a House district is 53,820 people. Id. No one disputes the new maps gave Districts 105 and 111 more white voters and fewer black voters. In total, the black share of the voting age population in both districts decreased by just over 2% as a result of the redistricting. Chen Report at 25. This may not seem like much, but speaking in terms of percentages distracts from the fact that the redistricting likely changed the outcome of the 2016 election in both Districts 105 and 111. Id. at 21-22. If the old maps had been used, both districts would have become significantly more diverse, and significantly more Democratic. Id. at 10. Plaintiffs' expert estimated that the percentage of black voter turnout in the 2016 District 105 election was more than four points lower than it would have been had the old map been used. Id. at 14. In District 111, the percentage of black voter turnout was estimated to be almost three points lower than it would have been under the old map. Id. at 15. But in any event, the redistricters were not trying to change the demographic makeup of Districts 105 and 111 dramatically. Their express purpose was to change Districts 105 and 111 just enough to protect the incumbents there, without endangering the incumbent Republican House members in the neighboring districts. And that's exactly what they did. Under the new map, Representatives Chandler and Strickland were both narrowly reelected. Representative Chandler won in District 105 with 50.5% of the vote to 49.5% for her Democratic challenger. Doc. 103-87 at 5. Representative Strickland won District 111 with 51.7% of the vote to 48.3% for his Democratic challenger. Id. at 6. An expert for the plaintiffs estimated that if the 2016 elections had been held using the old maps, both Representatives Chandler and Strickland would have lost. Chen Report at 21-22. Even Mr. O'Connor agreed that both Representatives likely would have lost their seats had their districts not been redrawn. O'Connor Dep. at 90:11-19. II. ANALYSIS A preliminary injunction is "an extraordinary remedy that may only be awarded upon a clear showing that the plaintiff is entitled to such relief." Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 22, 129 S.Ct. 365, 376, 172 L.Ed.2d 249 (2008). To warrant a preliminary injunction, a moving party must establish: "(1) a substantial likelihood of success on the merits of the underlying case, (2) the movant will suffer irreparable harm in the absence of an injunction, (3) the harm suffered by the movant in the absence of an injunction would exceed the harm suffered by the opposing party if the injunction issued, and (4) an injunction would not disserve the public interest." N. Am. Med. Corp. v. Axiom Worldwide, Inc., 522 F.3d 1211, 1217 (11th Cir. 2008) (quotation omitted). No party to this lawsuit disputes that redrawing Districts 105 and 111 made them more white and less black. But to state a claim for racial gerrymandering, the plaintiffs must show more than that. They must show "that race was the predominant factor motivating the legislature's decision to place a significant number *1364of voters within or without a particular district." Bethune-Hill v. Va. State Bd. of Elections, 580 U.S. ----, 137 S.Ct. 788, 794, 197 L.Ed.2d 85 (2017) (quotation omitted). A plaintiff can make that showing "either through circumstantial evidence of a district's shape and demographics or more direct evidence going to legislative purpose." Miller v. Johnson, 515 U.S. 900, 916, 115 S.Ct. 2475, 2488, 132 L.Ed.2d 762 (1995). This is particularly hard to do when the State offers a defense rooted in partisan gerrymandering, as it did here. We did not move these voters because they are black, the State tells us. We moved them because they were Democrats. And under current Supreme Court precedent, the State tells us this motive is perfectly acceptable. But if the State has "placed a significant number of voters within or without a district predominantly because of their race," they have engaged in unconstitutional racial gerrymandering, even if the ultimate objective of those moves was partisan advantage. Cooper v. Harris, 581 U.S. ----, 137 S.Ct. 1455, 1473 n.7, 197 L.Ed.2d 837 (2017) (quotation omitted and alteration adopted). This record documents that Ms. Wright had racial data available to her for redrawing Districts 105 and 111. She testified that race data was available on Maptitude down to the census-block level. Wright Dep. at 25:4-26:8, 105:24-25. In contrast, data about political affiliation was accurate only on a precinct-wide basis. As we've described, this meant Maptitude had much more granular data about race than about political affiliation. Chen Report at 32-33. Also, Ms. Wright testified her general practice was to keep open a box in Maptitude that gave the racial data for any potential district. Wright Dep. at 105:10-23. Representative Nix corroborated this with his testimony that racial data was visible on the Maptitude display at one of the group sessions where Ms. Wright tried out various options for new district boundaries. Nix Dep. at 150:8-12. And of course, Ms. Wright admitted she "eventually" looked at race data "to make sure that [she] did not do significant harm in that respect as well." Wright Dep. at 30:6-7. Also, although Ms. Wright ultimately denied having racial data visible while drawing the new lines for Districts 105 and 111, her testimony indicates that she knew where black residents were located on the maps. After all, she saw the location of black residents both when reviewing potential redistricting options with the affected House members before she undertook to finalize the lines, and after she had finalized them. Wright Dep. at 29:19-30:9, 105:10-23; Nix Dep. at 150:8-12. This being the case, can anyone say the process for redrawing House Districts 105 and 111 was blind to the race of possible voters? This record leaves no doubt that Ms. Wright, Mr. O'Connor, and all the other stakeholders involved, knew plenty about the racial demographics of Districts 105 and 111. Mr. O'Connor certainly knew about the racial effects of H.B. 566. He wrote an email the day after the House Reapportionment Committee approved the bill, explaining that as a result of the new map, "the black percentage in Rep. Chandler's district [District 105] drops by two points ... while Rep. Efstration's district [adjacent to District 105] increases from 22% to 24% black." Doc. 103-65. This email supports an inference that those involved in the redistricting were keeping a close eye on race, and considered a decrease in "the black percentage" in District 105 to mean they'd accomplished their mission. *1365The plaintiffs' evidence that race predominated this redistricting process is compelling. They have documented that the whole idea of redistricting House Districts 105 and 111 arose amidst talk about the changing demographics in Gwinnett and Henry Counties. Plaintiffs also point to the large and increased number of split precincts in the new map, especially in District 111. Splitting precincts increases the unpredictability of political data. Mr. O'Connor explained the problem with relying on political data from less than a whole precinct in an email sharing results from the 2016 presidential election. He wrote: Keep in mind that a lot of districts are estimated because of split precincts (that is, precincts which have two or maybe even 3 districts within them-unfortunately, precinct lines and district lines do not coincide all the time). Split precincts make data accumulation on State House districts very difficult. Doc. 103-47 at 3. Again, this difficulty arises because partisan data is only available on a precinct-wide basis. Any attempt to estimate partisan data at the sub-precinct level can therefore only be that-an estimate. The testimony of Ms. Wright and Robert Strangia, another Reapportionment Office employee, makes clear that Maptitude only made a rough estimate of partisan data when providing that data for less than a whole precinct. Wright Dep. at 110:10-111:8; Deposition of Robert M. Strangia ("Strangia Dep."), Doc. 110 at 25:12-26:22. To estimate partisan data for less than the whole precinct, Maptitude assumes the proportion of Democratic and Republican voters is the same throughout the precinct. Strangia Dep. at 25:12-26:22. There is no street-by-street data, like with race. As a result, "it is impossible for a map-drawer to gain detailed knowledge of whether one split portion of a precinct is more heavily Democratic or Republican-leaning than another split portion of the same precinct." Chen Report at 32-33. Or at least, it would be impossible for the map-drawer to gain such knowledge if she only relied on the partisan data available in Maptitude. In fact, it is very unlikely that the split precincts in Districts 105 and 111 had the same proportions of Democrats and Republicans throughout. But in order to know this, we must rely on the fact that these districts are racially polarized, then look at the street-by-street racial data available on Maptitude. Plaintiffs' expert points out that in six of the eight split precincts at issue in this case, the new maps pulled in portions of precincts with lower percentages of black voters, and left out portions with higher percentages of black voters. Chen Report at 34. Sometimes dramatically so. For example, the portion of the McDonough Central precinct inside District 111 had a voting age population that was 26.9% black, while the portion left outside was 50.1% black. Id. at 38. And the extreme racial polarization of voting in this district meant that the portion of McDonough Central inside of District 111 was more likely to vote Republican than the portion left outside. See id. at 2. Plaintiffs say all of these favorably split precincts show that Ms. Wright and her colleagues impermissibly used race as an effective proxy for party in deciding where to draw district lines. Certainly, if the redistricters just guessed, based on estimates of party affiliation, that splitting precincts the way they did would result in favorable outcomes for their candidates, they got pretty lucky. This is not direct evidence of racial gerrymandering, but it is circumstantial evidence that we find persuasive. *1366The State counters this argument by noting that, in many cases, it would not have been possible to include the other halves of the split precincts, with relatively larger proportions of potential black voters, and maintain contiguous districts. Doc. 137 at 30-31. But analyzing what Ms. Wright did not do in terms of splitting these precincts diverts us from our more important task of examining why she split these precincts in the first place. Republicans knew they were working with narrow margins for these districts. If they did not move enough Republicans into Districts 105 and 111, Representatives Chandler and Strickland would lose their seats. If they moved too many Republicans into their districts, the neighboring Republican seats would be at risk. Splitting precincts was risky. It made it harder to know the exact political consequences of redrawing a district in a particular way because there was no available data to say whether one portion of a precinct was more or less Republican than another. And Ms. Wright redrew District 111 to have five split precincts, out of a total of thirteen. Chen Report at 9, 28. This meant Ms. Wright did not have accurate partisan data for more than a third of the precincts in District 111. Yet she was startlingly effective in aid of the white Republican incumbents she was trying to help. The precise and effective results she got make it hard to accept that Ms. Wright and her colleagues did not have precise data directing the choices they made. Plaintiffs ask us to infer that Ms. Wright and her colleagues looked at the available block-by-block racial data on Maptitude in order to estimate political outcomes and give them some comfort in how to split precincts. This inference is certainly supported by circumstantial evidence. Beyond that, this record also supports the idea that Mr. O'Connor or the representatives themselves were familiar enough with the racial makeup of the relevant neighborhoods to just know it. The expert for the State seemed to have this sense as well. Referring to Mr. O'Connor, that expert testified: If he had knowledge of this particular area, then he could probably tell you something about the politics of-I mean, I know-I don't know him personally, but I've worked before with local political experts who can tell you almost block by block what a neighborhood is like. This neighborhood was established by Polish workers who came to work in a boiler factory. They're now over 65. The neighborhood is now half Hispanic. But people have intense local knowledge that's really much richer than what you'd get from tract level census data or precinct level demographics. Deposition of John R. Alford, Doc. 145 at 35:13-36:1. Regardless of how the redistricters knew the information about the racial distribution of voters within a precinct, if they used that information to decide where to split particular precincts, or whether to split precincts at all, they engaged in impermissible racial gerrymandering. Plaintiffs' expert testimony on split precincts is compelling. Yet we recognize that it does not rise to the level of the evidence found sufficient in Bush v. Vera, 517 U.S. 952, 116 S.Ct. 1941, 135 L.Ed.2d 248 (1996). There, the Supreme Court pointed to the surgical precision with which redistricters drew lines around black neighborhoods and the "extreme and bizarre" district lines that resulted. Id. at 971, 116 S.Ct. at 1957. The redrawn shapes of Districts 105 and 111 are much less *1367egregious, and Ms. Wright offers a race-neutral reason for each line-drawing decision she made. For example, Ms. Wright testifies she relied on Maptitude's political data estimates in deciding where to draw each district line. Wright Decl. ¶¶ 7-8. And for District 105, she points out that Maptitude's political estimates calculated the new district to be 5% more Republican than the old. Id. ¶ 21. This case therefore turns on a credibility determination, where one side has taken an oath that race was not a factor in how the redistricting lines were drawn, and the other side is not in a position to swear that it was. On this record, the "extraordinary remedy" of a preliminary injunction is not warranted.5 See Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 24, 129 S.Ct. 365, 376, 172 L.Ed.2d 249 (2008). The concurring opinion agrees that this result is "indisputably correct," yet it characterizes our analysis as "impugn[ing] the veracity of state employee witnesses who testified in depositions under oath." Conc. Op. at 1370. But of course, we have denied plaintiffs the preliminary injunction they seek precisely because we found that plaintiffs cannot refute Ms. Wright's testimony about how and why she drew the maps. At the same time, when our review of the record revealed contradictions in the testimony presented to us, we have viewed it as our obligation to take those complexities into account. That being the case, we recognize that the statements of Ms. Wright and Mr. O'Connor are sometimes at odds with other evidence in the record. For that reason, we reject the concurring opinion's "veracity characterizations," id. at 1370, based solely on the videotaped depositions of Ms. Wright and Mr. O'Connor, without recognition of other contradictory statements and documents that are plainly a part of this record. The concurring opinion also faults this opinion for "fail[ing] to provide critical facts necessary to explain the ultimate conclusion reached on plaintiffs' preliminary injunction motion." Id. at 1370. In support, the concurring opinion points to two pieces of evidence we omit from our recounting of the facts. We address each in turn. First, the concurrence highlights Representative Nix's testimony that he decided to proceed with redistricting for non-political reasons, including members' desires "to move a boundary to encompass family members or recently purchased land or reunite previously-split precincts." Id. at 1371. But Representative Nix's testimony in this regard was about districts redrawn as part of H.B. 566 other than Districts 105 and 111, which are the subject of this motion. Representative Nix's complete testimony listed a number of districts and representatives who asked for redistricting based on non-political reasons, like accommodating family members and purchases of property. Notably, Representatives Chandler and Strickland were nowhere on this list. Nix Dep. at 62:19-66:5. And as we've already described, Representative Nix also testified that both Representatives Chandler and Strickland approached him about redrawing their districts so they could improve their political performance. Id. at 141:14-18, 163:11-164:1. *1368Second, the concurrence cites from Ms. Wright's deposition, where she described her "understanding about the changing demographics in Gwinnett County" in terms of the changing political affiliation of voters in the county. Conc. Op. at 1371. The concurrence says we misrepresent the record by implying that Ms. Wright and others used "changing demographics" to refer to changes in the racial makeup of districts rather than any host of other statistical characteristics. Id. at 1371 n.2, 1371-72. Yet the concurrence's charge against us does not change our understanding of Ms. Wright's deposition testimony. We still read her testimony to indicate that she viewed "demographics" as a shorthand for race, politics, or some combination of the two. See, e.g., Wright Dep. 90:5-10 ("Q: Did you also know that there was a large growth in the black population? A: I don't know that I knew the numbers of it, but as we mentioned before, yes, we knew there was changing demographics and growth in different areas, so."). And based on the racially polarized voting in Georgia, Ms. Wright's and others' conflation of race and politics is hardly surprising. We set out to write an accurate opinion with a thorough account of the record. Our careful review of the points made in the concurring opinion does not convince us that we failed in this regard. We do feel it necessary to correct the record regarding the concurring opinion's characterization of our conclusion. The concurrence says we have "conjure[d] up a group sitting in a room clicking on Maptitude to move black voters from one district to another with the intent to depress black voter strength." Conc. Op. at 1371. As for the group sitting in a room clicking on Maptitude, our opinion conjured this up only to the extent that we rely on the testimony of Representative Nix that it happened. He plainly testified that interested parties did indeed sit together in a room and click through Maptitude in an attempt to draw safer Republican districts. Nix Dep. at 143:3-157:8. But nowhere in this opinion have we found that this group gathered "with the intent to depress black voter strength." If the record supported such a conclusion, a preliminary injunction would properly be entered. As we've said, we view the plaintiffs' evidence to be compelling. However, precisely because that evidence fell short of documenting an "intent to depress black voter strength," we have concluded no preliminary injunction is warranted. III. CONCLUSION Both Representative Strickland and Representative Chandler downplayed the impact of redistricting on their reelections. "I think I outworked" my challenger, Representative Strickland said. Strickland Dep. at 119:16. "I just have to work hard, period," Representative Chandler said. Deposition of Joyce H. Chandler ("Chandler Dep."), Doc. 126 at 75:25-76:1. But what both Representative Chandler and Representative Strickland did as well was to ask that more Republicans be put into their districts and that Democrats be taken out. This movement of voters helped these Representatives get reelected. This would be a more obvious case if it were a challenge to partisan gerrymandering. The state openly acknowledges it redrew Districts 105 and 111 with political ends in mind. Doc. 137 at 25. Of course these plaintiffs did bring a partisan gerrymandering claim, but this panel dismissed that claim because plaintiffs did not present us with any judicially manageable method for measuring discriminatory effect. *1369Doc. 28 at 32-35. And the plaintiffs did not replead that claim. We are thus left with a racial gerrymandering claim in which all involved state employees and officials have taken an oath that race played no role in their decision-making. Ms. Wright and her colleagues openly undertook to help Republican incumbents. In doing so, the 2015 redistricting moved many black voters from districts where their votes would have made an impact into districts where they did not. Do voters know the people they elect can and do shed their own voters to improve their ability to be reelected? It is said that "[t]he object of districting is to establish fair and effective representation for all citizens." Vieth v. Jubelirer, 541 U.S. 267, 307, 124 S.Ct. 1769, 1793, 158 L.Ed.2d 546 (2004) (Kennedy, J., concurring in judgment) (quotation omitted). But fair and effective representation is decidedly not what the voters removed from House Districts 105 and 111 got. Even so, our application of Supreme Court precedent to the record before us leaves us to conclude that plaintiffs are not entitled to the preliminary injunction they seek. Plaintiffs' Motion for Preliminary Injunction, Doc. 103, is DENIED . Footnotes [1] The individual plaintiffs are Lavelle Lemon, Marlon Reid, Lauretha Celeste Sims, Patricia Smith, Coley Tyson, Austin Thompson, Darryl Payton, Audra Cunningham, Sabrina McKenzie, Jamida Orange, Andrea Snow, Sammy Arrey-Mbi, Lynne Anderson, and Coretta Jackson. See Complaint ("Compl."), Doc. 1 ¶¶ 21-25; First Am. Compl. ("Am. Compl."), Doc. 84 ¶¶ 21-30. Mr. Reid, Ms. Smith, Mr. Tyson, and Mr. Thompson live in District 105. Doc. 1 ¶¶ 22, 24, 25; Doc. 84 ¶ 21. Ms. Lemon, Ms. Sims, Mr. Swanson, and Mr. Payton live in District 111. Doc. 1 ¶¶ 21, 23; Doc. 84 ¶¶ 22, 23. [2] Gerrymandering is a term used to describe the drawing of legislative district lines by a dominant political party to entrench itself in power at the expense of rival parties. See Ariz. State Legislature v. Ariz. Indep. Redistricting Comm'n, 576 U.S. ----, 135 S.Ct. 2652, 2658, 192 L.Ed.2d 704 (2015). [3] In her declaration, Ms. Wright also said she did not discuss the redrawing of Districts 105 and 111 with Mr. O'Connor. Wright Decl. ¶ 20. This is somewhat at odds with her deposition testimony that she could not recall whether she had such discussions. Wright Dep. at 192:17-193:14. In her deposition, Ms. Wright also said she spoke with Mr. O'Connor "pretty much daily." Id. at 192:21-193:1. And there is evidence that at least one meeting was scheduled between Ms. Wright, Mr. O'Connor, and Representative Strickland in December 2014 to discuss District 111. Doc. 103-37 at 1. [4] For example, the portion of H.B. 566 defining District 105 begins: District 105 Gwinnett County VTD: 135001-HARBINS A 050608: 2005 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2036 2037 2038 2039 2040 2041 2056 2057 2058 2059 [5] Having decided that plaintiffs do not show a substantial likelihood of success on the merits, we do not reach the remaining preliminary injunction factors. See Schiavo ex rel. Schindler v. Schiavo, 403 F.3d 1223, 1226 (11th Cir. 2005) ("Controlling precedent is clear that injunctive relief may not be granted unless the plaintiff establishes the substantial likelihood of success criterion.").