Opinion
Civil Action No. 2:03-CV-354.
June 9, 2005.
Richard Scott Gladden, Law Office of Richard Gladden, Denton, TX, Zeb Davidson Zbranek, Zbranek Firm PC, Liberty, TX, Otis W. Carroll, Jr., Ireland Carroll Kelley, PC, Tyler, TX, G. Michael Fjetland, International Legal Group, Sugar Land, TX, for Frenchie Henderson, et al.
William Andrew Taylor, Andy Taylor Associates, Houston, TX, Cassandra B. Robertson, David C. Mattax, Don R. Willett, R. Ted Cruz, Attorney General's Office, Robert M. Long, Houdyshell Long, LLP, Austin, TX, for Rick Perry, et al.
Before HIGGINBOTHAM, Circuit Judge, and WARD and ROSENTHAL, District Judges.
We are asked to examine again, in light of the Supreme Court's decision in Vieth v. Jubelirer, the claims that the redistricting plan for the election of the thirty-two members of Congress from Texas, adopted by the Texas legislature in 2003, is unconstitutionally tainted by excessive partisan purpose. Ultimately, we will adhere to our earlier judgment that there is no basis for us to declare the plan invalid.
541 U.S. 267 (2004).
We conclude that claims of excessive partisanship before us suffer from a lack of any measure of substantive fairness. The claims accept that some partisan motivation is inevitably present in the political enterprise of redistricting, but urge that at some point it can become unconstitutional, presumably a denial of equal protection. No party before us states with clarity the precise constitutional deficit. Although the lead plaintiffs invoke the structure of equal protection analysis, they identify no suspect criterion or impinged fundamental interest in insisting that if the state acts with the "sole" purpose of partisan advantage in drawing legislative districts, regardless of its effects, the state must offer a "compelling explanation" for its effort. The conduct that plaintiffs condemn is offered only in unstructured form; their condemnation of practices such as targeting incumbent members and ignoring "communities of interest" and other "traditional" principles of redistricting comes untethered to constitutional texts.
The most frequently invoked image of the evil resulting from excessive partisanship in drawing congressional districts is the non-competitive district, a product of a member choosing his constituents. We are asked to recognize this as anti-democratic and implored to find a means to curb it. The vision of the House of Representatives controlled by members who do not face serious opposition to reelection is urged as a stain upon its historical image as an institution embracing the teaching of Cincinnatus. The argument ignores a historical fact; the Texas delegation has enjoyed non-competitive districts for at least the past four and one-half decades, long before there were two political parties with any strength in the state. The emergence of Texas as a two-party state has not altered this reality, although it has given rise to forces that have caused the Texas delegation now to approximate the relative statewide voting strength of the two parties. As we will explain, there is little to suggest that taking up the tools plaintiffs offer in attacking the 2003 Texas plan will in fact remedy this awkward reality.
After addressing the claims of excessive partisanship, we will turn to a narrower and seemingly more plausible contention that does not attempt to measure how much partisanship in redistricting is constitutionally excessive, but instead uses the requirement of one-person, one-vote as a tool to limit how often redistricting can occur. This contention aims only at "voluntary" mid-decade redistricting that occurs when a legislature replaces a valid existing plan put into place after the last census; it does not attempt to set a standard for direct judicial supervision over partisan influence in redistricting, but instead proposes a rule that is intended to prevent the specific type of redistricting that occurred in Texas in 2003. The argument is that a legislature seeking to displace a valid extant plan may not rely on decennial census figures to meet the stringent demands of one-person, one-vote, but must instead prove that its proposed plan distributes population in an equipopulous manner by use of actual current figures. While its relative simplicity is seductive and avoids the need to measure how much partisanship is constitutionally excessive, we are not persuaded that it is appropriate for this court to endorse this application of the one-person, one-vote requirement as a means to the end of limiting political influences on redistricting.
I
The history of this case and of the efforts of the Texas legislature to draw lines for its thirty-two congressional districts is set out in our previous opinion, and we will not repeat it here. While the appeal from our judgment upholding the plan adopted by the Texas legislature was pending before it, the Supreme Court decided Vieth v. Jubelirer. In Vieth, the Court affirmed the decision of a three-judge court rejecting claims by three registered Democrats who vote in Pennsylvania that a redistricting plan for congressional districts adopted by the Pennsylvania legislature should be set aside because it constituted an impermissible political gerrymander, in violation of Article I and the Equal Protection clause of the Fourteenth Amendment. Their complaint, in addition to other claims, alleged that the districts were "meandering and irregular" and "ignored all traditional redistricting criteria . . . solely for the sake of partisan advantage." The three-judge court granted defendants' Rule 12 motion to dismiss for failure to state a claim, and the Supreme Court affirmed.
See Session v. Perry, 298 F. Supp. 2d 451 (E.D. Tex. 2004).
Id. at 272-73 (internal quotation marks and brackets omitted).
Shortly thereafter, in Cox v. Larios, the Court summarily affirmed the judgment of a three-judge court that had rejected a redistricting plan of the Georgia legislature as failing to conform to the principle of one-person, one-vote. The district court held that because the legislature sought to give advantage to certain regions of the state and to certain incumbents in an effort to help Democrats and hurt Republicans, Georgia was not entitled to the 10% deviation toleration normally permitted when a state is drawing lines for its legislature.
124 S. Ct. 2806 (2004).
Larios v. Cox, 300 F. Supp. 2d 1320 (N.D. Ga.) (three-judge panel), summarily aff'd, 124 S. Ct. 2806 (2004).
Then, after the summer recess, the Court remanded the present case "for further consideration in light of Vieth," making no reference to its decision in Larios. Responsive to the remand order, we received briefs and heard oral argument from all parties and amici.
Jackson v. Perry, 125 S. Ct. 351 (mem.) (Oct. 18, 2004).
Although in our prior opinion we turned back many attacks upon the legislative plan for electing members of the Texas congressional delegation, we read the remand order to be a directive to reexamine only our rejection of the claim that the Texas plan is an illegal political gerrymander. This mandate does not include consideration of other attacks, with the possible exception of the claim that the Texas plan failed to abide the command of one-person, one-vote. Variations of this one- person, one-vote claim have been in the case from its inception, but came to the fore only in the arguments following the remand to this court. The Court made no mention of one-person, one-vote in its remand order, nor was it at issue in Vieth. However, this issue was present in the Court's examination of Georgia's plan in Larios in a way arguably related to the present case. Any examination of compliance with one-person, one-vote thus faces the threshold hurdle of whether the claim is within the mandate of the remand. We will treat this question in due course.
II
The light offered by Vieth is dim, and the search for a core holding is elusive. This observation is not a criticism, but a recognition that Vieth reflects the long and twisting historical narrative of political gerrymanders in the United States.
The most recent chapter in this history of partisan influence upon the drawing of legislative districts involves the federal judiciary's effort to play the role it claimed for itself in Davis v. Bandemer. Judicial reluctance to surrender this role is understandable. The move to the one-person, one-vote principle in Reynolds v. Sims both answered some of the critics of Baker v. Carr and fulfilled the predictions of others who warned against entering the political thicket. While hardly analogous to the quest for standards for reining in partisan gerrymanders, the relatively quick two-year process culminating in Reynolds encourages those reluctant to concede the futility of finding an effective standard for Bandemer. And as we will suggest, we have yet to calculate the full costs of achieving the clear and easily administered standard of Reynolds.
478 U.S. 109 (1986) (finding political gerrymandering claims justiciable).
377 U.S. 533 (1964) (applying one-person, one-vote principle to malapportionment claims).
369 U.S. 186 (1962) (finding malapportionment claims justiciable).
See, e.g., Colegrove v. Green, 328 U.S. 549, 556 (1946) (opinion of Frankfurter, J.).
In addition, there is hesitation to concede that any solutions must come from legislatures and other political players whose critics say lack the ability to restrain themselves. This fear is fueled by the advent of computer-driven redistricting, which has taken this hoary practice to a new level. There is wariness of lines that fall precisely where the draftsmen intend — an absence of randomness or sufficient extraneous forces that the draftsmen must accept. At bottom it is a concern that the power to draw lines is inadequately checked, an implicit accusation that the political process is inadequate to the task. There is also the reality that many members of the House of Representatives enjoy a more secure tenure than members of the Senate for the simple reason that Senators run statewide, while their colleagues in the House may run in districts crafted to their advantage. Answers to such questions do not come easily.
See Vieth, 541 U.S. at 345-46 (Souter, J., dissenting), citing Samuel Issacharoff, Gerrymandering and Political Cartels, 116 HARV. L.REV. 593, 624 (2002) ("[I]ncumbent entrenchment has gotten worse as the computer technology for more exquisite gerrymandering has improved and political parties have ever more brazenly pursued incumbent protection."); Pamela S. Karlan, The Fire Next Time: Reapportionment After the 2000 Census, 50 SIAN. L.REV. 731, 736 (1998) ("Finer-grained census data, better predictive methods, and more powerful computers allow for increasingly sophisticated equipopulous gerrymanders."); Richard H. Pildes, Principled Limitations on Racial and Partisan Restricting, 106 YALE L.J. 2505, 2553-54 (1997) ("Recent cases now document in microscopic detail the astonishing precision with which redistricters can carve up individual precincts and distribute them between districts with confidence concerning the racial and partisan consequences.").
Even before the computer enhanced the ability to draw precise lines, the politicians were hardly without their own devices. At the same time, these perceived legislative line-drawing failures are distant from the state legislature's six-decade gridlock that the Court faced in Baker v. Carr. See 369 U.S. at 187-95 (Tennessee General Assembly had not been reapportioned since 1901, despite the Tennessee Constitution's decennial requirement); see also Reynolds, 377 U.S. at 583 (same, in Alabama).
One commentator has noted:
For the most part, redistricting appears to be done by barons dividing up fiefdoms, not by democratically accountable representatives. . . . The gerrymandered House contrasts with elections the same day for nongerrymandered Senate seats and governorships. About half of all gubernatorial and U.S. Senate elections were competitive in 2002, compared with fewer than 10% of House elections.
Richard H. Pildes, The Supreme Court 2003 Term — Foreward: The Constitutionalization of Democratic Politics, 118 HARV. L.REV. 28, 63-64 (2004) (footnotes omitted).
The Founders were no strangers to the self-interest afflicting legislators charged with drawing the lines for their own seats. Nor were they blind to the need to locate the power to curb potential abuses. With Article I, Section 4, they gave to the legislatures of each state authority to prescribe the times, places and manner of holding elections for Senators and Representatives. The Founders also insisted upon a superintending of the exercise of this power granted to the states. They gave that assignment to Congress by granting it the power to "make or alter" such regulations. Congress has exercised this power from time to time, as Justice Scalia recounted in his opinion in Vieth. While not directly speaking to the difficulties of gerrymandered state legislatures, this explicit placement in the Congress of the power to supervise the authority granted to states, coupled with the difficulty faced by judges of divining rules or standards adequate to distinguish a judicial decision resolving issues of partisanship in redistricting from a legislative act, has to date left the courts in the indefensible position of undertaking a task they cannot perform.
U.S. CONSI. art. I, § 4 ("The Times, Places and Manner of holding Elections for Senators and Representatives, shall be prescribed in each State by the Legislature thereof; but the Congress may at any time by Law make or alter such Regulations, except as to the Places of chusing Senators.").
Id.
541 U.S. at 275-77 (plurality opinion).
In upholding the Texas plan for congressional districts, we followed an unbroken line of cases declining to strike down a redistricting plan as an illegal partisan gerrymander. We left any change in direction to the Supreme Court, making only brief observations about our own years of work in this case and in drafting plans with the feared computers. We observed that the Court could make an honest case of Bandemer by either setting a standard or concluding that the issue was not justiciable. We expressed deep reservations over any approach that would dilute the critical Voting Rights Act by deploying it in name only or by borrowed concepts. We were and remain wary of employing metrics to determine how much is too much partisan motive or effect in redistricting, convinced that such an approach could not move the Court from its stasis under Bandemer. We need not further recount these observations.
See, e.g., O'Lear v. Miller, 222 F. Supp. 2d 850, 859 (E.D. Mich.) (three-judge panel), summarily aff'd, 537 U.S. 997 (2002); Marylanders for Fair Representation, Inc. v. Schaefer, 849 F. Supp. 1022, 1043 (D. Md. 1994) (three-judge panel); Terrazas v. Slagle, 821 F. Supp. 1162, 1172-75 (W.D. Tex. 1993) (three-judge panel); Pope v. Blue, 809 F. Supp. 392, 397 (W.D.N.C.) (three-judge panel), summarily aff'd, 506 U.S. 801 (1992); Ill. Legislative Redistricting Comm'n v. La Paille, 782 F. Supp. 1272, 1275-76 (N.D. Ill. 1992); Badham v. March Fong Eu, 694 F. Supp. 664, 671 (N.D. Cal. 1988) (three-judge panel), summarily aff'd, 488 U.S. 1024 (1989); see also Vieth, 541 U.S. at 280 n. 6 (collecting cases); Samuel Issacharoff Pamela S. Karlan, Where to Draw the Line?: Judicial Review of Political Gerrymanders, 153 U. PA. L.REV. 541, 543 (2004) ("In the ensuing eighteen years [since Bandemer], not a single challenge to a congressional or state legislative reapportionment managed to satisfy this standard.").
Session, 298 F. Supp. 2d at 457, 474-75.
In Vieth, the four Justices in the plurality voted to end the search for a workable standard, concluding that the legality of partisan gerrymanders is not justiciable and should be left to the political arena. Four dissenting Justices offered various possibilities for a standard that might serve a judicial role. Justice Kennedy cast the pivotal fifth vote to affirm the dismissal of the partisan gerrymandering claim. Although Justice Kennedy found each of the standards offered in the dissents deficient, he declined to abandon the search for a standard and, presumably, provided the fifth vote necessary to remand the present case.
541 U.S. at 305-06 (plurality opinion).
Id. at 317-41 (Stevens, J., dissenting); id. at 343-55 (Souter, J., joined by Ginsburg, J., dissenting); id. at 355-68 (Breyer, J., dissenting).
Id. at 305-17 (Kennedy, J., concurring).
Upon our reading of Vieth, then, our mandate requires us to look at this record again, with the message that the Court is unpersuaded by contentions that it can never properly locate a standard adequate to a judicial role in policing partisan gerrymanders, but that the standards the plurality rejected in Vieth were inadequate to that task. In their arguments on remand, the plaintiffs have offered various approaches for adjudicating claims of partisan gerrymandering. While the State's contention that most, if not all, of these arguments have been rejected by a majority of the Court is strong, we decline to stop there, given the unusual fracture of the Court in Vieth. We can only fairly read the remand to suggest that the Justice providing the fifth vote sees the possibility of a workable standard emerging from this case, the rejected allegations of the complaint in Vieth aside. We turn then to the various solutions offered on remand.
III
We first address the argument of the Jackson Plaintiffs that the current redistricting map is unconstitutional because it was driven solely by a partisan agenda. Before reaching the merits of this claim, however, we find it illuminating briefly to recount some of the events leading up to the passage of the redistricting plan now under attack.
Our reference to "Jackson Plaintiffs" includes the following: all plaintiffs included in the Amended Complaint filed on November 7, 2003, on behalf of the existing Jackson, Mayfield, and Manley plaintiffs; additional plaintiffs included there for the first time; and the Democratic Congressional Intervenors.
A
The history of electoral politics in Texas during the latter half of the twentieth century can be described as the story of the dominance, decline, and eventual eclipse of the Democratic Party as the state's majority party. From Reconstruction until approximately the beginning of the 1960's, the Democratic Party dominated the political landscape in Texas. In 1961, John Tower became the first Republican Senator elected from Texas since 1875. Throughout the 1960's, and for much of the 1970's, Republican voting strength on a statewide basis hovered near 35%. During this time, Republicans never held more than four congressional seats at one time.In 1978, William Clements, Jr., was elected Governor of Texas, becoming the first Republican to hold that office since 1874. In the 1978 election, Democrats won twenty out of twenty-four congressional seats and captured 56% of the vote in statewide races, while Republicans' statewide strength stood at 43%. Republican strength grew throughout the 1980's such that by 1990, the Republican Party had nearly achieved parity with the Democratic Party, garnering 47% of the statewide vote compared to the 51% for the Democrats. Nonetheless, Democrats still held the lion's share of congressional seats, with nineteen compared to the Republican's eight.
See Mike Kingston, Politics and Elections, in TEXAS ALMANAC 1982-1983 (1981), at 491.
See Appendix.
Id.
No doubt aware of the growing strength of the Republican Party, the Texas legislature, controlled by Democrats, enacted a a redistricting plan in 1991. Under this plan, Democrats won twenty-one congressional seats in the 1992 election compared to nine won by the Republicans, even though the "tipping-point" had been reached with the Democratic and Republican parties capturing an equal share of the vote in statewide races.
See infra note 48.
See Appendix.
Throughout the 1990's, Republican strength continued to grow, while the Texas congressional delegation remained firmly in the hands of Democrats. By the end of the decade, Republicans were consistently winning every statewide race on the ballot, including the offices of governor, lieutenant governor, attorney general, and seats on both the Supreme Court and Court of Criminal Appeals. Yet with the 1991 Democratic Party gerrymander still in place, Democrats captured seventeen congressional seats to the Republicans' thirteen in the 2000 election, despite Republicans garnering 59% of the vote in statewide elections to the Democrats' 40%.
See Office of the Secretary of State, 1992-2005 Election History, at http://elections.sos.state.tx.us/elchist.exe (last visited Apr. 1, 2005).
Id.
Following the 2000 census, the Texas legislature was unable to pass new lines for the Texas congressional delegation, and the task eventually fell to this court. For reasons that we will discuss, the plan produced by this court perpetuated much of the 1991 Democratic Party gerrymander. In the 2002 elections, the number of congressional seats held by Democrats remained unchanged, with Republicans gaining the two seats added by the census. In 2003, the Texas legislature, now controlled by Republicans, passed the redistricting plan that we upheld in Session and now review again in light of Vieth. Under this plan, Republicans captured twenty-one congressional seats in the 2004 election compared to eleven for the Democrats. In this election, Republicans carried 58% of the vote in statewide races compared to 41% for Democrats. It is against this backdrop that we now consider the Jackson Plaintiffs' arguments on remand.
See Balderas v. Texas, No. 6:01-CV-158, slip op. (E.D. Tex. Nov 14, 2001), summarily aff'd, 536 U.S. 919 (2002).
See Appendix.
Id. By way of comparison, when the statewide voting strength was roughly reversed in 1982, Democrats took twenty-two congressional seats to the Republicans' five. Id.
B
The Jackson Plaintiffs urge that we "distill from the Vieth opinions the principle that a decision to revise a districting map, along with particular features of the map, become unconstitutional when the evidence makes clear that the legislature was driven solely by a partisan agenda." Invoking equal protection analysis, the Jackson Plaintiffs contend that sorting voters for the sole purpose of gaining partisan advantage can serve no rational or legitimate purpose. This approach focuses on voluntary legislative redistricting — "voluntary" in the sense that it sets out to replace a valid extant plan. By definition, this approach would tolerate efforts to gain partisan advantage when the legislature is compelled to redistrict because the extant plan is invalid, such as when new decennial census figures require redistricting to comply with one-person, one-vote or to accommodate changes in the numbers of legislative members. This is so because efforts to gain partisan advantage in involuntary redistricting do not constitute the sole reason for the undertaking. Rather, the Jackson Plaintiffs' approach takes aim at mid-decade (or "mid-cycle") efforts to replace a valid extant plan, drawing on the observation in Session that structural or process-based constraints may have more purchase because they avoid the difficulties attending efforts to gauge how much is too much partisan motive or gain.
In support of their argument, the Jackson Plaintiffs point to Justice Kennedy's opinion in Vieth, in which he observes that a "determination that a gerrymander violates the law rests on something more than the conclusion that political classifications were applied. It must rest instead on a conclusion that the classifications, though generally permissible, were applied in an invidious manner or in a way unrelated to any legitimate legislative objective." Justice Kennedy, the argument continues, pointed to the Court's denial in Baker v. Carr that it needed to "enter upon policy determinations for which judicially manageable standards are lacking . . . if on the particular facts . . . a discrimination reflects no policy, but simply arbitrary and capricious action." It is suggested that redistricting for purely partisan purposes is an example of such arbitrary and capricious action. It is further suggested that this "sole reason" approach by its own terms would not apply in Vieth because the Pennsylvania legislature had to redistrict for multiple reasons, including the legal requirement of redrawing the lines after the 2000 decennial census in ways that created equipopulous districts based on that census, as well as satisfying the Voting Rights Act and accounting for "traditional" districting criteria, such as incumbent protection and minimizing split precincts.
541 U.S. at 307 (Kennedy, J., concurring).
Id. at 310 (Kennedy, J., concurring) (quoting Baker, 369 U.S. at 226) (internal quotation marks omitted).
Finally, the Jackson Plaintiffs parse the opinions of the four dissenting Justices, noting that all sought to locate a principle that would identify plans lacking a rational basis. The Jackson Plaintiffs urge that condemning efforts undertaken solely to gain partisan advantage is such a principle. They add as a final implementing principle the proposition that when a legislature controlled by a single party replaces a legal redistricting plan in the middle of the decade, the effort should be presumptively unconstitutional. They conclude that this presumption could be overcome only by a "compelling explanation."
This ignores, as it must, the reality that even with an overarching objective of feathering the party nest, the various cuts and turns of a redistricting plan with its reverberating impacts are infused with myriad mixtures of local politics and accommodation, inevitably producing lines drawn for a variety of reasons and objectives, often inconsistent with the overall objectives of partisan gains.
In response, the State urges that Vieth "squarely rejected the notion that `sole,' `predominant,' or `only' partisan intent suffices to state a claim." In Vieth, the Court found insufficient allegations that the Pennsylvania districts "ignore[ed] all traditional redistricting criteria, including the preservation of local government boundaries, solely for the sake of partisan advantage." The Court noted that the plaintiffs had alleged that when the Pennsylvania legislature turned to redistricting after the 2000 census, "the Republican Party controlled a majority of both state Houses and held the Governor's office. Prominent national figures in the Republican Party pressured the General Assembly to adopt a partisan redistricting plan as a punitive measure against Democrats for having enacted pro-Democrat redistricting plans elsewhere." The Court in Vieth had before it allegations that the Pennsylvania map was drawn "solely" and "exclusively" for political ends by a single-party-controlled legislature. These allegations were insufficient to overcome the motion to dismiss the political gerrymandering claim. Justice Stevens argued for a test based on such factors. The State points out that the contention by the Jackson Plaintiffs mirrors the dissent of Justice Stevens, or is at least functionally identical to it, and that Justice Kennedy expressly rejected the "standards proposed . . . by our dissenting colleagues."
541 U.S. at 272-73 (plurality opinion).
Id. at 272 (plurality opinion).
Id. at 317-19 (Stevens, J., dissenting).
Id. at 308 (Kennedy, J., concurring).
We are persuaded that the Jackson Plaintiffs offer a standard for measuring an excessively partisan redistricting plan that is functionally equivalent to the standard offered in Justice Stevens's dissent, a view rejected by five Justices. This similarity aside, the Jackson Plaintiffs' equal protection analysis, assertedly a structural approach, fails on its merits. Specifically, they are unable to locate a substantive right or suspect criterion to trigger strict scrutiny. Rather, they claim to rely upon the most deferential standard of review under the Equal Protection Clause, the absence of rationality. Even more, they would alter rationality review to insist that justification for the plan require proof that a legislature that voluntarily engages in redistricting have purposes other than partisan advantage by offering a "compelling explanation." That is, of course, not rationality review.
See, e.g., Williamson v. Lee Optical of Okla., 348 U.S. 483 (1955); cf. Lawrence v. Texas, 539 U.S. 558, 580 (2003) (O'Connor, J., concurring) ("When a law exhibits such a desire to harm a politically unpopular group, we have applied a more searching form of rational basis review to strike down such laws under the Equal Protection Clause.")
The fact that the Texas legislature's redistricting plan replaced the court-drawn plan put into place after the 2000 census does not make the legislative plan invalid in light of Vieth because it was "solely" motivated by political motivation. As noted, the Vieth plurality rejected a "sole" motivation test as a basis for measuring when partisan influences on redistricting are impermissibly excessive. Although Vieth did not involve mid-cycle redistricting to replace an existing plan, there is no constitutional or statutory prohibition on mid-decade redistricting, as we explained in our earlier opinion rejecting plaintiffs' contention that Texas lacked the authority to draw new district lines to replace the court-drawn map put into place after the last census. In that opinion, we noted that "innumerable decisions have either assumed that a state legislature may draw new lines mid-decade or have invited a state to do so after the court has drawn a map in a remedial role." For example, in Wise v. Lipscomb the Supreme Court observed:
Session, 298 F. Supp. 2d at 460; see id. at 460 n. 14, 461.
Legislative bodies should not leave their reapportionment tasks to the federal courts; but when those with legislative responsibilities do not respond, or the imminence of a state election makes it impractical for them to do so, it becomes the "unwelcome obligation" of the federal court to devise and impose a reapportionment plan pending later legislative action.
437 U.S. 535, 540 (1978) (emphasis added) (citation omitted); see also Branch v. Smith, 538 U.S. 254, 265-66 (2003); Upham v. Seamon, 456 U.S. 37, 44 (1982); Connor v. Finch, 431 U.S. 407, 411-12 (1977); White v. Regester, 422 U.S. 935, 935-36 (1975) (per curiam).
We also pointed to the practical solutions that Congress has available to prevent or limit such mid-cycle redistricting.
Session, 298 F. Supp. 2d at 474-75.
As those on whose shoulders we stand, we suffer no illusion of commission or ability to cleanse the air of partisan politics and self-interest, or to otherwise make angels of men. Rather, we accept the common-sense understanding that any voluntary redistricting would not have been undertaken unless a majority of the legislature thought it would advance their interests. That is, the self-interest of members of the legislative body will inevitably be a "but-for" cause of voluntary redistricting, the only activity the Jackson Plaintiffs would now condemn. This condemnation is driven by the assumption that the self-interest of members is a proxy of partisan interest. Putting aside limited amendments to cure some inadvertent error made in adopting an extant legal plan, if the initiating force of partisan ambition is sufficient to strike down all that follows, the principle contended for forbids mid-cycle redistricting by judicial fiat, when neither Congress nor the State of Texas has done so. This would contradict the long-standing assumption by courts that a state may replace existing court-imposed redistricting plans with plans enacted by the state's legislature.
Further, considering self-interest as a proxy of partisan purpose is forced. It does not accommodate the reality that a representative may act out of self-interest to secure a less competitive district — conduct that may or may not be beneficial to her party. Indeed it may be executed by a trade with a member of the opposite party equally actuated by the instinct of political survival.
In addition to making tendentious use of the equal protection standard and conflating the personal ambition of party members with partisan intent, the Jackson Plaintiffs' approach is question-begging in a more fundamental way. It does not escape, because it cannot, the absence of a substantive measure of fairness. It ends the inquiry into legality with a finding that the State acted with the sole purpose of obtaining partisan advantage for the controlling party, presuming that such action is irrational and impermissible regardless of its actual effects. This approach discounts the possibility that there may be rational justifications for attempting to redistrict to improve a party's position. For example, it is not clear that acting to undo a perceived disadvantage imposed previously by an opposing party is irrational in the sense that it admits of no salutary or constitutionally acceptable result.
In concept, the Supreme Court could announce a constitutional principle that acting solely with partisan purpose has no place in the drawing of districts. In implementing this principle, the Court could then adopt a prophylactic rule forbidding voluntary mid-cycle redistricting by state districting bodies controlled by one party — a bold but candid pronouncement, the Miranda of redistricting jurisprudence. The Jackson Plaintiffs shy from this step. Rather, they urge a "process," albeit one that admits of a single conclusion: that mid-cycle redistricting is unconstitutional. The inability to formulate an enforceable principle except one that gathers its normative content from an implementing rule both raises the question of justiciability and draws into question the legitimacy of the announced principle. After all, it is a much smaller step from the two underlying building blocks of Miranda — the due process and Sixth Amendment-based right to not be convicted upon an involuntary confession and an experience-based factual judgment of the inherently coercive environment of the station house — to the implementing prophylactic of Miranda's warning requirement. The baseline in Miranda was a settled constitutional principle, not an elusive condemnation of conduct that some would say is antithetical to American ideals and others would say is politics as old as the Republic itself.
The articulation of a constitutional principle here — such as condemning as irrational the drawing of district lines with the purpose (colored "sole," "dominant," "voluntary," or otherwise) of partisan gain — must face the facts of this case. While the present plan, drawn by a Republican Party majority in 2003, has been decried as egregious, the story must begin with the earlier map drawn by a Democratic Party majority in 1991. That plan, put in place following the 1990 census, was cited in the political science literature as an extreme example of what one party can do in drawing a redistricting map to the detriment of the other. In 2000, the Democratic Party gerrymander was still in place and, although Republicans now enjoyed substantial statewide majority strength, the results of the congressional elections favored Democrats by a seventeen to thirteen margin.
See, e.g., MICHAEL BARONE, THE ALMANAC OF AMERICAN POLITICS 2004, at 1510 (2003) [hereinafter BARONE 2004] ("The plan carefully constructs democratic districts with incredibly convoluted lines and packs heavily Republican suburban areas into just a few districts."); MICHAEL BARONE, THE ALMANAC OF AMERICAN POLITICS 2002, at 1448 (2001) (describing it as the "shrewdest gerrymander of the 1990s").
Another commentator describes it as follows:
The impact of political gerrymandering on the competitiveness of elections has not been given the attention it deserves. For example, following the 1990 census, Texas experienced what is sometimes referred to as "the great partisan gerrymander of '91." The Democrat-controlled legislature carefully created conservative districts around the Republican incumbents. This "packing" strategy helped the Democrats in the 1992 election to win 21 of the other 22 districts.
Brian P. Marron, Doubting America's Sacred Duopoly: Disestablishment Theory and the Two-Party System, 6 TEX. F. ON C.L. C.R. 303, 337 (2002) (footnotes omitted).
See Appendix.
The map drawn by this court in 2001 perpetuated much of this gerrymander. It did so because this court was persuaded that it could not achieve "fairness" to political parties without some substantive measure of what is "fair." Simply undoing the work of one political party for the benefit of another would have forced this court to make decisions that could not be defended against charges of partisan decision-making — again, for the lack of a substantive standard. As the panel explained, it would follow only "neutral" redistricting standards. Once the panel had left majority-minority districts in place and followed neutral principles traditionally used in Texas — such as placing the two gained seats in the areas of growth that produced them, following county lines, avoiding the pairing of incumbents and the splitting of voting precincts, and undoing transparent offsetting movements of the same number of residents between districts — the drawing ceased, leaving the map free of further change except to conform it to one-person, one-vote. Make no mistake, this undertaking, while shorn of partisan motive, had political impact in the placement of every line. The results of this court's plan did ameliorate the gerrymander and placed the two districts gained by Texas in the census count; however, doing more necessarily would have taken the court into each judge's own notion of fairness. The practical effect of this effort was to leave the 1991 Democratic Party gerrymander largely in place as a "legal" plan.
See Balderas, No. 6:01-CV-158, slip op.
Id. at *5.
The difficulties of partisan decision-making do not go away as dominance by one party over the other is diminished. Non-competitive districts allocated among incumbents perfectly reflecting party strength as measured by their numbers of representatives are the likely outcome. This reality alone raises questions of the aptness of rules indexed by controlling parties acting in mid-cycle. Even if this foray were to enjoy some modicum of success, it leaves untouched the most common occurring occasion for partisan redistricting — the effort following the decennial census.
Indeed, the 2002 congressional election totals were identical to the 2000 results, with the exception of the two new seats. See Appendix.
Plaintiffs seize upon the targeting of certain incumbents in the present plan, Democrats who had been reelected even as their constituents voted for Republicans with increasing frequency. This was one of its most controversial features — egregiously wrong in the eyes of Democrats. In the eyes of Republicans, however, this feature was justified because the effect of the court plan was to perpetuate incumbency that was itself the product of a partisan gerrymander. Under the legislative plan, twenty-one Republicans and eleven Democrats were elected in November 2004. While this was a substantial swing, the State urges that it actually better reflected the statewide voting strength of the parties than did the court plan; that the size of the swing only reflected the distortion caused by the gerrymandered plan it replaced. That is, the displaced judicially-crafted plan of 2001, while easing the partisan outcome of the 1991 gerrymander, nonetheless left the minority party (Democrats) in control of the majority of the congressional seats.
Id.
See id.
See id.; see also BARONE 2004, supra note 48, at 1508 ("In U.S. House races, as they have since 1994, Republicans won more votes than Democrats, but fewer seats, thanks to a 1991 Democratic redistricting plan which was closely followed by a court in 2001.").
The State urges that the legislative purpose of the current plan was to remedy this unfair drawing — that its line-drawing was hardly "invidious." The State does not suggest that there is a constitutional right to proportionality, or that one constitutional wrong justifies another. Rather, the State argues that this outcome is relevant to the Jackson Plaintiffs' contention that the present legislative plan is irrational as a matter of law. As Justice White explained in Gaffney v. Cummings,
judicial interest should be at its lowest ebb when a State purports fairly to allocate political power to the parties in accordance with their voting strength. . . . [We do not] have a constitutional warrant to invalidate a state plan, otherwise within tolerable limits, because it undertakes, not to minimize or eliminate the political strength of any group or party, but to recognize it and, through districting, provide a rough sort of proportional representation in the legislative halls of the State.
See Vieth, 541 U.S. at 288 (plurality opinion) (noting that "the Constitution contains no such principle" of proportional representation).
412 U.S. 735, 754 (1973).
In other words, the suggestion is not that escaping from one impermissible partisan gerrymander is a license to replace it with another, and we do not understand the State to argue as much here. Rather, it is that even by the "irrationality" measure proposed by the Jackson Plaintiffs, the current Texas plan is not an impermissible gerrymander at all.
It is instructive to compare the results of the Texas legislature's redistricting effort with the results of the redistricting plan enacted by the Pennsylvania legislature that the Court upheld in Vieth. In the 2000 general election, Pennsylvania voters elected eleven Republicans and ten Democrats to represent them in Congress. In that same election, Republican
The following is a summary of the results of the last three election cycles in Pennsylvania:
Year Statewide Strength Pennsylvania Congressional Seats R D R D Total
2004 46% 51% 12 (63%) 7 (37%) 19
2002 47% 50% 12 (63%) 7 (37%) 19
2000 49% 48% 11 (52%) 10 (48%) 21
See Bureau of Commissions, Elections and Legislation, Pennsylvania Department of State, 2004 General Election Returns, at http://www.electionreturns.state.pa.us/(last visited Apr. 1, 2005); Bureau of Commissions, Elections and Legislation, Pennsylvania Department of State, Archived Election Results Selections, at http://web.dos.state.pa.us/cgi-bin/ElectionResults/elec_archive.cgi? which=Archive (last visited Apr. 1, 2005).
"Statewide Strength" was calculated by averaging the percentage of vote received by each party in all of the following races that occurred in Pennsylvania in a given year: Governor, Attorney General, Auditor General, and U.S. Senator. This provides a rough approximation of a party's general appeal statewide. Cf. Appendix.
Party statewide candidates captured an average of 48% of the statewide vote against 47% by Democratic Party candidates.
See table, supra note 60.
Following the 2000 census, Pennsylvania's allotment of congressional seats was reduced from twenty-one to nineteen. Pennsylvania's General Assembly then took up the task of drawing a new districting map. Under the plan that the Assembly produced, and which the Court ultimately upheld in Vieth, Pennsylvania voters elected twelve Republicans and only seven Democrats to Congress in the 2002 general election. These results were repeated in the 2004 general election, giving Republicans control of 63% of Pennsylvania's congressional seats despite the fact that Republican statewide candidates captured an average of only 46% of the statewide vote against 51% for the Democrats. In short, under the plan passed by the Pennsylvania General Assembly and upheld by the Court in Vieth, the party that garnered, on average, less than half the vote in statewide races was able to capture nearly two-thirds of Pennsylvania's congressional seats. In contrast, the plan passed by the Texas legislature resulted in the election of twenty-one Republicans and eleven Democrats to the House of Representatives in 2004, when the Republican Party carried 58% of the vote in statewide races and the Democratic Party carried 41% of the vote.
See id.
See id.
See id.
See Appendix.
The State's description of the 2003 Texas legislative plan as dismantling a prior partisan gerrymander that had entrenched a minority party, in order to allow a party with overwhelming statewide voting strength to capture two-thirds of Texas's congressional delegation, is a characterization that the record supports. Under the Jackson Plaintiffs' analysis, with its keystone declaration that partisan line drawing is per se irrational, if the Democrats in Pennsylvania were to obtain control of the state legislature, they could not then voluntarily undertake to undo the present map — even to bring Pennsylvania's congressional delegation more in line with the parties' apparent statewide strength — because they would be acting solely for partisan advantage. However, as we have explained, saying it is irrational, even saying it many times, does not make it so. And if the effects of the Pennsylvania plan did not provide a basis to find excessive partisanship in redistricting, it is hard to see how the effects of the Texas plan make it constitutionally offensive.
In short, the plaintiffs' contentions on remand are conspicuous for want of any measure of substantive fairness, one that can sort plans as "fair or unfair" by something other than a judge's vision of how the judiciary ought to work — more precisely, how the judiciary ought to run this show. We are persuaded that the inability of any plaintiff to conquer these difficulties, even when supported by able lawyers with an army of advisors, is explainable by the reality that this effort has been to give a legislative task to a court.
Of course, to paraphrase Justice Harlan, the fact that a standard for justiciabiliy is not precisely definable does not mean that it is ineffable. See Poe v. Ullman, 367 U.S. 497, 524 (1961) (Harlan, J., dissenting). However, if such a standard simply is ineffable, we should not be surprised by the inability of numerous talented individuals to define it.
C
Beyond the plaintiffs' inability to articulate a measure of substantive fairness, we are unable to locate in any of the proposals offered a clear articulation of the failures in governance to which courts are asked to direct their supervisory efforts. The plaintiffs persist in advancing claims that rest upon the perceived loss of individual rights by voters who, they allege, are victimized by impermissible classifications. The plaintiffs are unable to locate and address structural defects and conspicuously fail to connect their claims with the most visible awkwardness in today's electoral structure: the absence of competitive districts."Safe seats" protecting incumbent legislators flourish regardless of whether political parties are fighting for advantage, and appear even when "neutrals," not the legislators, are drawing the lines. Indeed, as Justice O'Connor has observed, the rough and tumble of partisan politics may work against the proliferation of safe seats. Her prescient observation bears emphasis:
That a judicially-approved plan could result in a majority of non-competitive districts is not surprising given that turning over the line-drawing process to independent commissions, as some states have done, does not necessarily avoid this result. See Steven Hill, Editorial, Schwarzenegger vs. Gerrymander, N.Y. TIMES, Feb. 19, 2005, at A29.
Indeed, there is good reason to think that political gerrymandering is a self-limiting enterprise. In order to gerrymander, the legislative majority must weaken some of its safe seats, thus exposing its own incumbents to greater risks of defeat — risks they may refuse to accept past a certain point. Similarly, an overambitious gerrymander can lead to disaster for the legislative majority: because it has created more seats in which it hopes to win relatively narrow victories, the same swing in overall voting strength will tend to cost the legislative majority more and more seats as the gerrymander becomes more ambitious.
Bandemer, 478 U.S. at 152 (O'Connor, J., joined by Burger, C.J., and Rehnquist, J., concurring).
Further, the creation and perpetuation of non-competitive districts can be facilitated by cooperation across the political aisle, as incumbents from each party negotiate to protect as much of their political turf as possible. It is much like the oft-cited observation about the marketplace that when two or more competitors come together, the conversation will inevitably turn to price.
See Gaffney, 412 U.S. at 738-39 (addressing cooperative efforts by the two parties in Connecticut to carve up the landscape).
There are other forces that disconnect Texas voters from their "representatives," including one-person, one-vote. Texas is an increasingly urbanized state, with over 60% of its population concentrated in the large metropolitan areas of Houston, Dallas-Fort Worth, San Antonio, and Austin. In drawing a map, the districts covering West Texas must reach eastward across hundreds of miles to gather sufficient population to meet the equipopulous requirement. These "reaches" must extend far beyond genuine communities of interest and come with a real risk here that West Texas farmers and ranchers will be represented in Congress by a person residing in one of the large metropolitan areas, creating a substantial disconnect between voters and their representatives. Again, this disconnect is facilitated in part by one-person, one-vote and the Voting Rights Act, both of which require map drawers to reach out for voters, departing from local communities tied by common economic pursuits and local tradition and creating opportunities for self-interested line drawing. Specifically, in reaching out, the drawers are invited to select voters favorable to their own electoral chances — a temptation driven primarily by the self-interested desire to gain safe seats, not the secondary and more remote goal of achieving "partisan purposes." In effectuating this goal of safe seats, map drawers may choose to pull certain populations into their districts for a host of reasons: some areas may produce low voter-turnout; others may be dominated by agrarian interests; and yet others may be populated by several generations of the drawer's extended family.
See U.S. Census Bureau, U.S. Dep't of Commerce, Ranking Tables for Metropolitan Areas: 1990 and 2000, at http://www.census.gov/population/cen2000/phc-t3/tab03.pdf (Apr. 2, 2001), at tbl. 3. The concentration of the Texas population can also be illustrated by comparing the census figures for the most populous and least populous counties in the state. According to the 2000 census, 45% of Texans live in five of the state's 254 counties. See U.S. Census Bureau, U.S. Dep't of Commerce, Census 2000 Summary File 1 (SF 1) 100-Percent Data, Geographic Area: Texas — County (8, 458, 627 out of total population of 20,851,820 live in Harris, Dallas, Tarrant, Bexar, and Travis Counties), available at http://www census.gov/census2000/states/tx.html. By comparison, only 2.6% of Texans live in the state's 100 least populous counties. Id.
See Michael W. McConnell, The Redistricting Cases: Original Mistakes and Current Consequences, 24 HARV. J.L. PUB. POL'Y 103, 112 (2000) ("Both as a matter of state constitutional law and as a matter of custom, legislators used to be extremely reluctant to violate city, county, and township lines. Now, under `one person, one vote,' they are required to do so. And once that constraint is lifted, they are liberated to snake lines all over the map to achieve their own purposes. The result, as Justice Harlan warned back in Reynolds, is an invitation to gerrymandering," (citing Reynolds, 377 U.S. at 622 (Harlan, J., dissenting))).
Although plaintiffs argue that the districts under the 2003 Texas plan are non-competitive as a result of partisan gerrymandering, their argument and evidence assumes, but does not show, a necessary or actual correlation between partisan line drawing and an increase in the number of non-competitive Texas congressional districts. Again, the historical record is instructive. Texas has not held a congressional election for at least the past four and one-half decades in which more than a handful of districts were "competitive." For most of this time, Texas was dominated by the Democratic Party at both the statewide and local levels. In many congressional races held during this time, the incumbent had no opponent. Self-interest of incumbents controlled. As the Republican Party came to strength during the late 1980's and 1990's, the Democratic Party continued to hold a large majority of the congressional seats. Even with the growing competitiveness between the parties, the number of non-competitive districts remained relatively constant. The Democrats resisted the statewide growth in the number of Republican voters with its redistricting plan following the 1990 census. It is important to understand that this partisan effort was greatly aided by the Voting Rights Act, in combination with the introduction in 1964 of the one-person, one-vote principle. The redistricting plan drawn following the 1990 census was the fruition of these requirements. By the time the plan now under attack was first proposed, the Voting Rights Act had effectively taken six Democratic Party seats off the table, rendering them untouchable and largely non-competitive. That is, only twenty-six of the thirty-two congressional seats allocated to Texas could be pursued by the Republican line drawers, who began their task with fifteen existing seats. To adhere to the goal of taking a share of the Texas delegation that, at a minimum, approximated its then-current state voting strength, the Republican Party had to hold its fifteen seats and pick up at least three or four of the remaining eleven seats. In sum, the 2003 Texas map drawn by the Republicans effectively achieved their goals during the 2004 election, but effected little change in the number of competitive districts. Texas has not had a significant number of competitive House races for at least the past four and one-half decades. The cases that guide our way do not provide a basis for us to hold that a redistricting plan's failure to change this long course of events is a basis to invalidate that plan.
See Appendix.
Id.
Id.
Id.
See supra note 48 and accompanying text.
D
The point is simple. It is difficult to set upon a course of treatment until the illness is diagnosed, and broad-spectrum responses have no place here — or should not. The Texas plan is not more partisan in motivation or result, including the impact on the number of competitive districts, than the Pennsylvania plan upheld in Vieth. The Jackson Plaintiffs have not identified a way to invalidate the Texas plan under the standards they urge as surviving Vieth. Yet, as the dissents and Justice Kennedy's concurrence in Vieth make clear, the disquiet with the role of partisan politics in redistricting persists, despite the difficulties in expanding the judicial role. The search for some judicial means to limit political gerrymandering takes us to the University Professors' suggestion that the principle of one-person, one-vote can serve this end. We now turn to that contention.
IV
The Amicus Curiae Brief of University Professors, which was supported in the main by all plaintiffs, seeks to apply the one-person, one-vote principle as a brake against excessive partisan redistricting by reducing the frequency of redistricting efforts. This argument, while present in various forms throughout this litigation, was brought to the fore by the Professors' brief, taking its place alongside the contentions of the Jackson Plaintiffs.
The argument is easily stated. Each new decennial census immediately places many legislative bodies in clear violation of one-person, one-vote, requiring the drawing of new districts. The process leading to a new plan may be extended for several years by court challenges and multiple legislative efforts, sometimes resulting in a court-promulgated plan. Although decennial census data is by that time no longer accurate, it is still used to draw the map. The use of this otherwise inaccurate data is sometimes described as a "legal fiction." The Professors urge that this fiction is supported by the necessity of allowing courts and legislative bodies to rely on the census figures without any demonstration of their present accuracy. This fiction, the argument continues, should not be enjoyed by a legislative body that voluntarily redistricts — that is, acts to replace a legal existing plan. In other words, dilution of the powerful command of one-person, one-vote should not be allowed when redistricting is not required by law. Pointing to the observation in our Session opinion of the practical values of a congressional prohibition of such voluntary redistricting efforts, the Professors urge that while that decision belongs to the Congress, not the courts, the enforcement of this constitutional principle is the obligation of the courts. They argue that it may be an alternative structural brake on partisan gerrymandering, conceding the present absence of a meaningful metric of substantive fairness — of how much partisan gain is too much — and declining to offer one.
See Georgia v. Ashcroft, 539 U.S. 461, 489 n. 2 (2003) ("When the decennial census numbers are released, States must redistrict to account for any changes or shifts in population. . . . After the new enumeration, no districting plan is likely to be legally enforceable if challenged, given the shifts and changes in a population over 10 years. And if the State has not redistricted in response to the new census figures, a federal court will ensure that the districts comply with the one-person, one-vote mandate before the next election."); see also Reynolds, 377 U.S. at 584 ("[I]f reapportionment were accomplished with less [than decennial] frequency, it would assuredly be constitutionally suspect.").
See, e.g., Georgia v. Ashcroft, 539 U.S. at 489 n. 2 ("[B]efore the new census, States operate under the legal fiction that even 10 years later, the plans are constitutionally apportioned."); Johnson v. Miller, 922 F. Supp. 1556, 1563 (S.D. Ga. 1995) (three-judge court) ("In the calculus of district population deviation, our only measure of the state's demographics is the decennial census. Since the population is not static, we adhered to the fiction that the census block figures are accurate to the exclusion of all others."), aff'd, Abrams v. Johnson, 521 U.S. 74 (1997); see also People ex rel Salazar v. Davidson, 79 P.3d 1221, 1233 (Colo. 2003) ("The United States Supreme Court has recognized the legal fiction that these figures remain accurate for the entire ten years between censuses.").
The argument as presented comes unadorned with supporting case citations, relying upon the exacting principle of one-person, one-vote and the observation that the limit on partisan gerrymanders the Supreme Court is seeking is lying at its feet. The simple logic of denying the benefit of the fiction that decennial census figures remain accurate throughout the decade to those who indulge in mid-cycle redistricting, together with the belief that here lies the most inviting ground for regulation, has given life to this argument.
The Court has only recently demanded exactitude in the population of districts drawn by the Georgia legislature for partisan gain. See Larios, 124 S. Ct. at 2808 (Stevens, J., concurring). It did so in Larios by denying it the 10% toleration of deviation in the drawing of lines for state legislative seats.
The State replies that injecting one-person, one-vote at this juncture is beyond the scope of the mandate; that it has a tangential tie at best to the directive to reconsider in the light of Vieth; and in any event would be attended by its own coterie of practical problems of application.
In oral argument, Professor Scot Powe was modest in the claims he made concerning the effects of applying one-person, one-vote here, observing that even if requiring the State to demonstrate compliance with current numbers would offer only a small brake, it should be required. Viewed from this perspective, the strength of the argument does not rest upon its utility in curbing partisan excesses, although it might do so. Rather, the strength lies with the power of the proposed rule itself. It asks why a state legislative body under no legal requirement to redistrict should face anything less than the full reach of the constitutional strictures of one-person, one vote — that is, why it should benefit from a fiction born of necessity.
The University Professors deny that their rule would constitute a bar on a legislature that wants to replace a court-ordered plan before the next census is conducted. At the same time, they concede that their proposed rule would, by design, make it very difficult because only the decennial census figures provide sufficiently detailed and reliable data for redistricting. They also acknowledge that "there is effectively no alternative" to decennial census data. The State agrees, noting that " Wesberry v. Sanders requires essentially perfect equality between congressional districts. In practice, that high degree of mathematical precision would be all but impossible using any data other than decennial block-level data." Judge Ward suggests in his concurrence that if a state wanted to pursue a mid-decade re-redistricting effort, it could perform a special state-wide census to provide the necessary numbers. Such a suggestion goes beyond what the parties recognized as within the realm of the practical or feasible. In their briefs and at oral argument, the litigants conceded that such data do not exist and could not practically be obtained. The proposed rule is intended to, and would, serve as a means to the end of preventing what Texas did here, to redistrict mid-decade to replace a court-imposed plan with one crafted by the legislature.
Tr. at 164.
Univ. Prof. Br. at 4; see id. at 3 ("[U]se of population enumerations from the most recent federal decennial census is essential . . . because such recent census data is the most comprehensive, precise, and objectively attainable enumeration data available."). They point out that some "[m]ore current, but incomplete information, such as population estimates . . . or minority voter registration or turn-out" exists. Id. at 5. These data, however, would be "unacceptable as a basis for actually drawing district boundaries on a systematic, statewide basis." Id. Although other parties attempted to argue that they could rely on data showing county and city population trends since the census to show that the 2003 Texas plan was not equipopulous as of 2005, the University Professors as well as the State recognized that such data would not meet the legal requirements for drawing district boundaries. See id.; State Reply Br. at 41; Tr. at 69, 133, 137; see also Valdespino v. Alamo Heights Indep. Sch. Dist., 168 F.3d 848, 853-854 (5th Cir. 1999).
State Reply Br. at 41 (citation omitted).
The practical effect of the rule would be to bar what courts have stated was within the prerogative of the state legislatures: to draw their own map to replace one imposed by a court. In none of these cases has a court suggested that the state legislature could not use the data from the last census, but instead had to conduct a special, statewide census, in order to replace a court-drawn map with a legislatively-drawn map. The departure from what courts have assumed is the proper relationship between court-imposed plans and legislatively-enacted plans signals caution in endorsing the University Professors' proposed rule.
The courts have assumed that legislative plans replacing court-imposed plans are expected to rely on the data from the last census, despite the passage of time. See Johnson, 929 F. Supp. at 1563 (upholding the Georgia State Senate redistricting plan that the legislature drew in 1995 using 1990 census data and acknowledging that "[a]t this point in the decade," the data for one-person, one-vote purposes "is largely theoretical. The real data, known only to providence, would doubtlessly lead us to another result."), aff'd, Abrams v. Johnson, 521 U.S. 74 (1997); Terrazas v. Clements, 537 F. Supp. 514, 516 (N.D. Tex. 1982) (implementing a temporary reapportionment plan based on 1980 census figures that would remain "in effect for all elections through December 31, 1983, unless valid apportionment plans are enacted sooner"); Bush v. Martin, 251 F. Supp. 484, 488 n. 3 517 (S.D. Tex. 1966) (tentatively approving a 1965 legislative redistricting plan based on 1960 census data, but noting that the Texas Legislature could adopt a plan before the next decennial census).
The University Professors assert that their proposed rule merely protects the fundamental right of people to elections conducted under maps that provide equal representation to equal numbers of persons. Yet it is difficult to discern how their proposed rule actually protects or furthers the one-person, one-vote principle, as opposed to the one-person, one-vote fiction. The Supreme Court has frequently observed that the decennial census is only briefly accurate, noting on one occasion that "the well-known restlessness of the American people means that population counts for particular localities are outdated long before they are completed." One of the University Professors has described
Karcher v. Daggett, 462 U.S. 725, 732 (1983); see also Gaffney, 412 U.S. at 745.
[o]ne of the odder features of the one-person, one-vote doctrine, when applied to the population of electoral districts, . . . that it seemingly applies only in the first election cycle out of the (usual) five in any ten-year period. That is, the practical dynamics of population growth and mobility in the United States operate to assure that mathematically identical districts (by whatever measure) in the first election are almost certainly going to be different, often dramatically so, by the third or fourth election. Indeed, as a practical matter, this differentiation might be present even by the actual occurrence of the first election, which takes place two years after the enumeration on which the reapportionment will have been based.
Sanford Levinson, One Person, One Vote: A Mantra in Need of Meaning, 80 N.C. L.REV. 1269, 1278 (2002).
The University Professors do not explain how, in an election held in 2004, 2006, or 2008, the districts drawn in 2001 using 2000 census numbers are more equipopulous if measured against current actual population data than districts drawn in 2003 using 2000 census numbers. If the University Professors' application of one-person, one-vote is adopted, the virtually certain result would be to have the next elections conducted under the court-ordered plan drawn in 2001, using 2000 census numbers. Yet all recognize that the 2001 plan is no more a reflection of today's actual population distribution than the 2003 legislative plan; both are based on the 2000 census numbers. Neither plan would produce equipopulous districts for elections held in 2006, if those districts could be measured against current data. The difficulty in discerning how the University Professors' proposed rule in fact promotes one-person, one-vote underscores its actual purpose, of limiting mid-decade "voluntary" redistricting in order to limit political gerrymandering.
The University Professors' argument would insist on current population data only for voluntary mid-cycle efforts to redistrict. It is not clear, however, whether insistence on current population figures for voluntary redistricting could be so cabined. The combined effect of such a rule and the incentives of politics are difficult to predict. Indexing liability to voluntary redistricting could create large incentives to seek a judicial invalidation of an existing plan as violative of equipopulous requirements. If, as Judge Ward posits, a state can feasibly and practically obtain mid-decade census data that is sufficiently detailed and reliable to use in drawing district lines, the proposed rule may have the perverse effect of breeding more efforts to redistrict. Such data could provide not only a means to defend a "voluntary" re-redistricting plan, but also a basis to attack valid extant plans drawn using the decennial population data. If one-person, one-vote is applied as the University Professors advocate, and the ability to obtain reliable and detailed mid-decade census information does or will exist, then would mid-decade re-redistricting be required to meet the demand of one-person, one-vote applied in this fashion? Despite the assertions of its proponents, the actual bite of a rule denying reliance on decennial data to redistricting efforts voluntarily undertaken is both important and uncertain.
Reynolds v. Sims speaks generally about the frequency of legislative efforts to reapportion and the use of decennial census data toward the end of the decade. 377 U.S. at 583-84. Chief Justice Warren observed that "[r]eallocation of legislative seats every 10 years coincides with the prescribed practice in 41 of the States, often honored more in the breach than the observance, however." Id. at 583 (footnote omitted). He noted that, while acting only at the end of a decennial period leads to some imbalance in the population of districts, limitations on the frequency of these efforts are justified by the need for stability and continuity in the legislative system. Id. The Court's eye was, of course, on failures to reapportion, finding no difficulty with reapportioning more frequently than every ten years, while finding less frequent efforts to be suspect. Id. at 583-84; id. at 583 n. 65 ("[T]he constitutions of seven . . . States either require or permit reapportionment of legislative representation more frequently than every 10 years."). Of course, the allocation of seats among the states rests on the decennial count. See 2 U.S.C. § 2a.
We recognize, with Judge Ward, the potential benefits of precluding mid-decade re-redistricting efforts. Judge Ward's citation to the legislative history of the 1976 Census Act amendments, noting that "constantly changing representative districts" is "bad politics, resulting in bad government," is instructive. Providing states with an incentive to obtain their own mid-decade census data to use in redistricting does not appear consistent with Congress's decree that federally-funded mid-decade data could not be used in redistricting because mid-decade redistricting was "bad politics." This history underscores Congress's authority to prohibit or limit discretionary redistricting efforts by the states.
The rule the University Professors and other plaintiffs propose would require us to apply an established doctrine in a novel way, with uncertain basis and effect. We add to these concerns the question of whether the entire contention is outside the mandate of the remand order. While it is true that Larios is relevant to this contention, the Court's order makes no reference to it. We do not wish to overstate any of these concerns viewed singly. Cumulatively, however, they counsel caution. We choose to abide that counsel and decline to adopt the University Professors' proposed rule.
V
The G.I. Forum, Congresswomen, and Texas-NAACP on remand return to their claims that the Texas plan impermissibly burdens minority voters in violation of the Voting Rights Acts and the Equal Protection Clause of the Fourteenth Amendment. Each would tie these claims to partisan gerrymandering. The contention is that these violations occurred in the effort to gain partisan advantage, however else the effort may be flawed. We examined and rejected all of the claims in detail in our previous opinion. As these claims are beyond the scope of the mandate we are not persuaded that we should revisit them.VI
Having reconsidered, we respectfully adhere to our judgment. We deny all relief requested by Plaintiffs and judgment is entered for Defendants.
APPENDIX
The following table was compiled from data provided by the Texas Secretary of State's Office as well as from the Texas Almanac, volumes 1962-1963 through 1992-1993. Year Statewide Strength Congressional Seats R D R D Total Competitive Unopposed 2004 58% 41% 21 (66%) 11 (34%) 32 3 7 2002 57% 41% 15 (47%) 17 (53%) 32 6 9 2000 59% 40% 13 (43%) 17 (57%) 30 2 9 1998 56% 43% 13 (43%) 17 (57%) 30 1 11 1996 55% 44% 14 (47%) 16 (53%) 30 9 1 1994 52% 47% 11 (37%) 19 (63%) 30 5 5 1992 49% 49% 9 (30%) 21 (70%) 30 3 5 1990 47% 51% 8 (30%) 19 (70%) 27 2 13 1988 47% 52% 8 (30%) 19 (70%) 27 2 13 1986 43% 56% 10 (37%) 17 (63%) 27 2 10 1984 51% 49% 10 (37%) 17 (63%) 27 5 9 1982 40% 59% 5 (19%) 22 (81%) 27 4 9 1980 45% 54% 5 (21%) 19 (79%) 24 6 4 1978 43% 56% 4 (17%) 20 (83%) 24 6 4 1976 38% 60% 2 (8%) 22 (92%) 24 4 6 1974 28% 70% 3 (13%) 21 (88%) 24 3 8 1972 45% 51% 4 (17%) 20 (83%) 24 2 11 1970 34% 66% 3 (13%) 20 (87%) 23 2 13 1968 31% 69% 3 (13%) 20 (87%) 23 0 12 1966 32% 67% 3 (13%) 20 (87%) 23 4 17 1964 29% 71% 0 (0%) 22 (100%) 22 2 0 1962 37% 63% 2 (9%) 20 (91%) 22 3 5 We calculated "Statewide Strength" by averaging the percentage of votes garnered in all statewide elections occurring in a given year, including Governor, Lieutenant Governor, Attorney General, Comptroller of Public Accounts, Commissioner of the General Land Office, Commissioner of Agriculture, Railroad Commissioners, Texas Supreme Court, Texas Court of Criminal Appeals, U.S. Senator, and U.S. Congressman-at-Large. For our purposes, this provides a rough approximation of a party's general appeal statewide. We excluded the results for President from these figures.The "Competitive" column tallies those seats that were competitive either in the general election or in the primary. We considered a seat to be "competitive" if it was won with less than 55% of the vote. We first looked to the results of the general election, separating competitive from non-competitive races. For those races deemed non-competitive, we looked to the primary of the party whose candidate prevailed in the general election, and separated competitive and non-competitive races at the primary level. We then tallied all competitive races identified at both the general and primary levels to arrive at an overall measure for a given electoral cycle.
See Richard H. Pildes, supra note 12, at 63 (defining competitive districts as those won with less than 55%) (citing David R. Mayhew, Congressional Elections: The Case of the Vanishing Marginals, 6 POLITY 293, 304 (1974); GARY C. JACOBSON, THE ELECTORAL ORIGINS OF DIVIDED GOVERNMENT: COMPETITION IN U.S. HOUSE ELECTIONS, 1946-1988, at 26).
If a primary race — or a general race in 1996 — went to a runoff, we used the figures from the runoff to determine competitiveness. However, were we to consider the presence of a runoff to be per se "competitive" the number of competitive seats would increase by 2 in 2004, 1 in 2002, 1 in 2000, 1 in 1998, 2 in 1996, 2 in 1984, 2 in 1982, 2 in 1978, and 1 in 1964.
In 1996, thirteen of the districts were redrawn by a court, following a previous decision that some of those districts were the product of racial gerrymandering. See Vera v. Bush, 933 F. Supp. 1341 (S.D. Tex. 1996). Special elections open to all candidates were held and a number of runoffs resulted.
The "Unopposed" column tallies those races in the general election with only a single party candidate — regardless of any third-party candidates.
The table below sets forth in further detail the data on competitiveness, broken down by party in the primary and general elections.
Year General Election Primary Total Seats Competitive Competitive R D R D
2004 32 1 1 0 1 2002 32 1 3 2 0 2000 30 1 1 0 0 1998 30 0 1 0 0 1996 30 2 5 0 2 1994 30 1 4 0 0 1992 30 0 2 0 1 1990 27 0 2 0 0 1988 27 0 2 0 0 1986 27 1 0 1 0 1984 27 3 0 2 0 1982 27 0 2 0 2 1980 24 2 2 0 2 1970 24 1 4 0 1 1970 24 0 3 0 1 1974 24 1 1 0 1 1972 24 1 0 0 1 1970 23 0 0 1 1 1968 23 0 0 0 0 1966 23 0 1 0 3 1964 22 0 1 0 1 1962 22 1 1 0 1
I concur in the result reached by the majority on the question whether any of the plaintiffs have set forth a standard for judging partisan gerrymandering claims which survived the holding of Vieth v. Jubelirer, 124 S.Ct. 1769 (2004). I write separately to explain in somewhat more detail why, if unconstrained by the scope of the Supreme Court's mandate, I would adopt the "one-person, one-vote" arguments initially presented by the Travis County parties and now urged on remand by the University Professors.
I
Justice Harlan warned that the Supreme Court's "one-person, one-vote" rulings might lead to more partisan gerrymandering efforts. He predicted that the Court's holdings requiring mathematical exactitude might encourage more drawing of district lines to maximize the political advantage of the party temporarily in control of the state legislatures. He also observed that "[a] computer may grind out district lines which can totally frustrate the popular will on an overwhelming number of critical issues" Wells v. Rockefeller, 394 U.S. 542, 551 (1969) (Harlan, J., dissenting) Those concerns have been realized. See Vieth, 124 S.Ct. at 1816 (Souter, J., dissenting) ( citing, inter alia, Pamela S. Karlan, The Fire Next Time. Reapportionment After the 2000 Census, 50 Stan. L.Rev. 731, 736 (1998) (noting that "[f]iner-grained census data, better predictive methods, and more powerful computers allow for increasingly sophisticated equipopulous gerrymanders.")).
In Justice Breyer's dissenting opinion in Vieth, he noted that "[b]y redrawing districts every 2 years, rather than every 10 years, a party might preserve its political advantages notwithstanding population shifts in the State" Vieth, 124 S.Ct. at 1827 (Breyer, J., dissenting). By incorporating data from election cycles which post-date the most recent census data, a State under the control of a single political party may repeatedly fine-tune gerrymanders at the expense of the constitutional promise of one-person, one-vote. In Session, the court expressed a similar concern, stating: "[o]ur point is that if the judiciary must rein in partisan gerrymanders, limitations that focus upon the time and circumstance of partisan line-drawing and less upon the `some but not too much' genre of strictures offer the best of an ugly array of choices." Session v. Perry, 298 F.Supp.2d 451, 475 (E.D. Tex. 2004). The rule advocated would require the State to demonstrate that a new redistricting plan does not worsen any population deviations existing among the current districts, denying the benefit of the fiction that the population reflected by the census remains accurate throughout the decade.
The rule is not a blanket prohibition on mid-cycle redistricting, as urged by the State. See State Defendants' Response Brief on Remand at 36 ("Plaintiffs' claims fail because they are an undisguised attempt at a backdoor judicial prohibition on `mid-decade' redistricting, and this Court has already correctly concluded that `mid-decade' redistricting is legal and permissible"). The rule proposed involves a balancing of the State's authority to redistrict mid-decade with the voters' rights to equal representation in the Congress. The court is therefore correct to recognize implicitly that the "one-person, one-vote" and "mid-decade" arguments are distinct creatures. To be sure, the rule proposed in this case would make it more difficult for a State to exercise its authority to alter congressional districts repeatedly over the course of a decade, but it would not preclude it.
II. A.
The majority questions whether the proposed rule (designed to curb the frequency of mid-cycle efforts) might breed even more efforts to gerrymander. The court is surely correct to raise that issue. But the majority does not explain why the rule as distinct from politics might encourage more redistricting. For my part, I am reminded of the practical observations of one court that politicians are not interested in redistricting "because of their innate attachment to mathematical exactitude." Daggett v. Kimmelman, 811 F.2d 793, 801 (11th Cir. 1987). Thus, it is unlikely that a State would engage in mid-cycle redistricting unless the party in power thought it could gain partisan advantage by doing so. Moreover, as a legal matter, the Supreme Court has held that initial plans adopted with the most recent census data comply with one-person, one-vote throughout the decade:
When the decennial Census numbers are released, States must redistrict to account for any changes or shifts in population. But before the new census, States operate under the legal fiction that even 10 years later, the plans are constitutionally apportioned.Georgia v. Ashcroft, 123 S.Ct. 2498, 2515 n. 2 (2003). This legal principle would doom lawsuits directed toward invalidating initial plans drawn with the decennial data. The rule urged by the University Professors does not require states to take any action above what is already mandated by existing Supreme Court precedent.
B.
The majority also suggests that the rule might not serve as a sufficient brake and that this counsels against adoption of the rule: "Despite the assertions of its proponents, the actual bite of a rule denying reliance on decennial data to redistricting efforts voluntarily undertaken is both important and uncertain" The majority is being bashful. No rule designed to police partisan gerrymandering could have less bite than Bandemer. As Justice Scalia noted in Vieth, the one case that upheld preliminary relief under Bandemer did not involve the drawing of district lines. In all cases involving the drawing of district lines, relief under Bandemer was refused. Vieth, 124 S.Ct. at 1778-1778 nn. 5-6 (collecting cases). Contrary to the majority's suggestion, I am convinced that a requirement that a State demonstrate current population figures to comply with one-person, one-vote when replacing a valid plan would serve as a structural brake on partisan redistricting efforts.
The gerrymanderer's paraphernalia is the decennial census data, election results, and a powerful computer. Although the parties do not explicate on the difficulty of deriving more accurate population data, the University Professors suggest (and the State recognizes) that such data do not currently exist. See University Professors' Brief at 4 ("After all, there is effectively no alternative."); State Defendants' Response Brief on Remand at 37 (". . . it requires the State to show an up-to-date data set that the Professors say does not exist. . . .") Neither the State nor the Professors explain that such data could not exist, only that it does not The burden is, of course, on the State to make a good faith effort to achieve population equality among the districts, and it is unlikely that proof of such equality could be met solely with estimates or statistical sampling. See Karcher v. Daggett, 462 U.S. 725, 736 (1983) ("[T]he census data provides the only reliable-albeit less than perfect-indication of the districts' `real' relative population level."); Department of Commerce v. United States House of Representatives, 525 U.S. 316 (1999) (holding that Census Act prohibits use of statistical sampling methods to compensate for undercounting).
A State seeking to replace a valid plan in the middle of the decade might choose to conduct a special statewide census. Properly conducted, this effort would accommodate both the State's constitutional prerogative to redistrict mid-cycle and the voters' rights to districts that comply with one-person, one-vote principles. The State offers no principled reason for subordinating the latter to the former, and I can perceive none. What the State wants in this case is the right to rely on the federal census data to facilitate its efforts at partisan gerrymandering while failing to update the enumeration. This seems to me a perverse use of the data sanctioned by the Supreme Court for use in efforts to avoid the dilution of individual votes cast in congressional elections.
C.
There are practical reasons to require the State to update the census data before embarking on mid-cycle redistricting efforts. Legislators know their districts, and cartographers are aware of what is happening on the ground in terms of demographics. Although the State asserts that "[t]he record does not show that the legislators took into account any post-Census population changes that had already occurred," State Defendant's Brief at p. 44 n. 30, the evidence cited by the Travis County parties from one of the architects of the map demonstrates that new CD4 was designed with an eye toward what was happening on the ground so that it could "grow into" a Republican district during the decade. Brief of Travis County and the City of Austin at p. 10; Tr December 18, 2004, Afternoon Session, at pp. 177-178, Testimony of Phil King ("So this District 4, I think will grow Republican. That's one reason we drew it that way Grayson and Collin counties are fast-growing Republican areas, so we felt like if we didn't win 4 now we'd grow into it"). This is one of the primary dangers of mid-cycle redistricting predicted by Justice Breyer in his dissent in Vieth. The rule proposed by the University Professors would require that districts designed with an eye toward what is happening on the ground be accompanied by proof that those districts were created in a good faith effort to achieve equal population.
In addition, cartographers in Texas who are either unable or unwilling to develop more current data risk diluting the strength of the fastest-growing segment of the population, the Latino population LULAC's brief, for instance, illustrates a comparison of the estimated populations of District 19, a predominately Anglo district, to District 28, a predominately Latino district. District 19 is growing at a much slower pace than District 28. To indulge the fiction of the accuracy of the census data under these circumstances encourages cartographers to use their knowledge of current demographics as well as voting trends exhibited through election cycles when drawing new maps. That legislators in some states might be disposed to use their knowledge of population shifts occurring after the release of the decennial numbers while gerrymandering during the initial round of redistricting is not a sufficient reason to shy away from the rule in this case. As Justice Souter explained in defending his effort to limit extreme gerrymanders: "[t]he most the plurality can show is that my approach would not catch them all." Vieth, 124 S.Ct. 1769 (Souter, J., dissenting).
III.
Although a rule denying the use of the decennial data is rooted in the Constitution, that rule enjoys the benefit of being consistent with what little statutory authority touches on the subject. The 1976 amendments to the Census Act authorized a mid-decade census for use in administering public benefit programs. See 13 U.S.C. § 141(d)-(g). One of the statutory provisions provides that:
Information obtained in any mid-decade census shall not be used for apportionment of Representatives in Congress among the several States, nor shall such information be used in prescribing congressional districts13 U.S.C. § 141(e)(2).
The purpose of the second clause of this provision was explained by its sponsors, who noted that the overall goal of the legislation was:
to provide timely statistical information to assist the Federal Government in administering the allocation of funds under its various programs, assist States and local government in planning, and to assist the business community and the general public.
H.R. Report No. 94-944, at 17-18 (Supplemental Views of Representatives Edward J. Derwinski and Trent Lott). Representatives Derwinski and Lott explained that any issues that were extraneous to the legislation should be removed Id. One of these extraneous issues was the possibility that some might use the federally-funded mid-decade data in redistricting efforts. Citing a 1971 paper prepared by the American Law Section of the Library of Congress, Representatives Derwinski and Lott recognized that Congress had the constitutional power to prohibit discretionary redistricting efforts attempted by the States. Id. In reliance on the 1971 paper, the Representatives concluded:
Therefore, there appears to be a sound basis for the legislative action proposed in our amendment, but it also has the practical effect of avoiding the confusion to the public of having constantly changing representative districts. It is bad politics, resulting in bad government, to promote the continual shifting and drifting of congressional district lines.Id. Representatives Derwinski and Lott therefore sponsored the amendment enacted as part of section 141(e)(2), which prohibited any reliance on mid-decade census data for redistricting purposes. The obvious assumption underlying this amendment was that congressional redistricting would have already occurred once in the decade by the time the mid-decade census was conducted.
Carried to its extreme, the State's logic in this case is that it has the Article I power to alter its congressional districts throughout the decade and it may rely exclusively on the decennial census data to discharge its one-person, one-vote obligations at any time during that period. But it makes little sense for the courts to permit a State to rely on federal decennial data throughout a decade when Congress has explicitly banned reliance on more current data that might be assembled in the middle of that decade A rule denying reliance on the decennial data for voluntary redistricting efforts is therefore consistent with the Congressional policies embodied in section 141(e)(2) of the Census Act.
IV.
The State questions as "bizarre" any holding that would reinstate Plan 1151C over 1374C on the grounds that the former might be more equipopulous than the latter if used in the remaining elections. The State observes that the two plans rely on the exact same census data, and the State's population has shifted equally since the 2000 Census, regardless of whether Plan 1151C or Plan 1374C is in place. The State misses the point, however, of the argument. The former plan was born of necessity to bring the State into compliance with one-person, one-vote principles. The latter statute was born of a "single-minded purpose" of the Legislature to achieve partisan political gain. Session, 298 F. Supp. 2d at 470. The Supreme Court requires States to utilize the census data to facilitate the constitutional voting rights guaranteed by Wesberry. That the data can be manipulated to other ends when used in conjunction with an effort to create equipopulous districts at the beginning of a decade is no justification for extending the protection afforded by the data to state legislatures which voluntarily embark on the task of redistricting for partisan political purposes.V.
Although I concur in the result reached by the court with respect to the question whether the plaintiffs' proposed tests survived the holding in Vieth, I respectfully disagree with the court's suggestion that partisan gerrymandering is not to blame for the decreased competition in congressional races. To be sure, bi-partisan gerrymandering is also a culprit. See Samuel Issacharoff and Pamela S. Karlan, Where to Draw the Line? Judicial Review of Political Gerrymanders, 153 U.Pa. L.Rev. 541, 570-71 (2004). But surely one of the primary goals of Plan 1374C is to elect and entrench Republican congressmen, whatever the voting trends of the state's population might be.
Dr. Alford's declaration explains how gerrymanders of the sort presented in this case contribute to the lack of competitive elections. That declaration, tendered in support of the Jackson Plaintiffs' brief on remand, recounts the utter non-competitiveness of most of the districts under Plan 1374C. As readily admitted by the State, the design of Plan 1374C was to produce a 22-10 Congressional delegation in favor of the Republican Party. As Dr. Alford explains, the one surprise from the 2004 election cycle was how close the map came to producing those numbers in the very first election conducted under the plan. Dr. Alford predicts that in the upcoming 2006 elections, one of the districts won by a Democrat, District 17, will be one of only a dozen or so truly competitive congressional districts in the country. As Dr. Alford has observed, Plan 1374C is not only a partisan gerrymander, but it is a very effective one, no doubt enabled in part by the architects' knowledge of the most recent election returns. Nevertheless, despite my disagreement with some of the language in the court's opinion, I am persuaded that the plaintiffs have not articulated a test which survived the holding of Vieth. I therefore concur in the result reached by the court that Vieth does not compel us to reach a different conclusion in this case.
VI.
In conclusion, I note that the majority and I agree on some very important points. First, the court's opinion observes that the University Professors' argument is a "narrower and seemingly more plausible contention" when compared to other efforts which attempt to measure the elusive concept of the fairness of redistricting plans. I agree Second, the language of the majority opinion ("[w]e do not wish to overstate any of these concerns viewed singly") can hardly be construed as an endorsement of the State's arguments on the merits of this issue. With that view, I also agree. What tips the scale is the scope of the mandate.
The courts have marked time since Bandemer, while the partisan rancor exhibited in the redistricting process has grown to new levels. Justices Stevens and Breyer recently observed that:
After our recent decision in Vieth v. Jubelirer, 541 U.S. 267, 124 S.Ct. 1769, 158 L.Ed.2d 546 (2004), the equal-population principle remains the only clear limitation on improper districting practices, and we must be careful not to dilute its strengthCox v. Larios, 124 S.Ct 2806, 2808 (2004) (Stevens, J. concurring). As the Travis County parties suggest in their brief on remand, this court's prior opinion did not give due consideration to their argument. Nevertheless, an inferior court's desire to end (or at least limit) this "bloodfeud," Balderas v. State of Texas, 6:01-CV-158, slip op., at 10 (E.D. Tex. Nov. 14, 2002), does not justify the exercise of judicial power arguably outside the scope of a higher court's directive.
It has not escaped my notice that the remand order in this case was issued after Larios but instructs this court to reconsider in light of only Vieth. The rule proposed by both the Travis County parties and the University Professors is grounded on principles of one-person, one-vote and is therefore more closely aligned with the holding in Larios. Given free reign to reconsider in light of the court's affirmance in Larios, I would not indulge the legal fiction that the census data remain accurate under these circumstances. I am persuaded that any shortcomings of the rule suggested by the majority are more perceived than real. Respect, however, for the scope of the mandate causes me to concur in the judgment denying relief at this time.
I continue to adhere to the remaining views expressed in my original opinion concurring in part and dissenting in part.