League of United Latin American Citizens [LULAC] v. Perry embraced, in the context of partisan gerrymandering, Felix Frankfurter's conclusion that the Supreme Court should not enter the political thicket of legislative apportionment. Two years earlier in Vieth v. Jubelirer, the Court split 4-1-4 on the justiciability of partisan gerrymandering. O'Conner and the three conservatives held it was nonjusticiable. Each of the four moderate liberals offered a test showing it was justiciable. Kennedy dissented from the conservatives while simultaneously rejecting each of the four tests offered. He announced he was waiting for a better test. When far superior tests were offered in LULAC, he rejected them too.
The newly rejected tests were simplicity personified. Both hinged on the fact of mid-decade redistricting. One claimed that when a state engages in voluntary unnecessary redistricting, it bears the burden of justification. The other avoided justiciability altogether by reverting to the early "one person, one vote" cases. It claimed that a state was bound by Karcher v. Daggett to make a good faith effort to create districts as equal as possible. Because the Texas redistricting was mid-decade, the census data was outdated and therefore the Texas legislature could not know if its plan was equal, better, or worse than the one it replaced. Indifference could not constitute good faith.
LULAC not only offered a clear standard, it offered clean facts. A divided Texas legislature failed to adopt a new redistricting plan in 2001 so a federal court did. Then, with Republicans controlling both houses and the governor, the legislature replaced the court-ordered plan with one of its own in 2003 (on the third try after Democrats thwarted the first two by fleeing to Oklahoma and New Mexico respectively. Who says there is no learning curve for Democrats?). No one disputed that the goal of the mid-decade redistricting in Texas was to defeat all ten incumbent Anglo Democratic Congressmen and to add seven Republicans to the Texas delegation, which after the 2002 elections had a 17-15 Democratic edge even though no Democratic candidate had been elected statewide since 1996. (Additionally the plan shored up the district (District 23) of Hispanic Republican incumbent Henry Bonilla by removing 100,000 Hispanics (a group that overwhelmingly preferred his opponents in recent elections) and replacing them primarily with Anglos. Kennedy, for the majority, held this move to violate the Voting Rights Act. Kennedy, for a different majority, held that the destruction of Marvin Frost's Dallas district, where he had enjoyed huge African American support, did not violate the Voting Rights Act.)
The 2003 plan, demanded by then Republican majority leader Tom DeLay, was a masterpiece that worked. Only three of the ten incumbent Anglo Democrats won reelection as Democrats-surprisingly one in a heavily Republican district-and the Texas delegation shifted to a 21-11 Republican majority. The results gave Republicans an added margin in their narrow majority in Congress. Furthermore, it is all but inconceivable that any Republican incumbent, except Bonilla, could lose in the new districts during the remaining three election cycles before the constitutionally mandated redistricting takes place. It will be necessary to change between four and six district lines to comply with the holding that District 23 violates the Voting Rights Act, but only Bonilla is at risk in this process.
Kennedy rejected the mid-decade trigger because it could not reach all partisan gerrymandering and because it could encourage partisan gerrymandering at the beginning of the decade. The best face that could be placed on the Republicans' plan was that it was tit-for-tat. Kennedy accepted the claim that a Democratically controlled Texas legislature in 1991 adopted a gerrymander that allowed Democrats to keep control of the Texas congressional delegation even as the party was losing control of state elections for the first time since Reconstruction. Democrats, Kennedy reasoned, had gerrymandered to keep a minority party in power; Republicans had gerrymandered in 2003 to give a majority party (a bit more than) its due. The mid-decade test would only allow the latter redistricting to be held unconstitutional. This Kennedy thought was peculiar.
Because the test would have left the 1991 gerrymander alive but invalidated the 2003 one, Kennedy asserted it "does not have the reliability appellants ascribe to it." Kennedy makes perfection the enemy of the "pretty good." Even though partisan gerrymandering is a bad thing, there is probably no way to stop it because, as the four conservatives claim, all districting is gerrymandering and there is no baseline to tell when a party has gone from constitutional padding to an unconstitutional grab. The fact that that problem of gerrymandering cannot be entirely solved does not mean that the problem of voluntary redistricting without state (as opposed to partisan) justification cannot be solved. Placing the burden on the state will sniff out mid-decade partisan gerrymandering. It is perfectly reliable for that issue. In conflating voluntary redistricting with mandatory redistricting, Kennedy has functionally adopted the conservatives' position.
Kennedy also believes that if mid-decade partisan gerrymandering is blocked, then one effect "could be to encourage partisan excess at the outset of the decade, when a legislature redistricts pursuant to its decennial constitutional duty." In other words, if they are prohibited from cheating later, then they will cheat sooner. But when one party controls both the legislature and the governor, it needs no encouragement to gerrymander itself more seats. Computers have allowed politicians to perfect gerrymandering for two decades now. Telling legislators it is now or never is hardly news. After all, until Texas engaged in mid-decade state-wide redistricting, no state in the modern era had done so (either because no one had thought of it, or because initial gerrymanders had proven effective, or because the legislators understood that they had no reliable population enumerations for assuring equal population among the districts).
Arguing in the court below, one of us asserted that the state's theory would allow for a partisan gerrymander in 2009 in preparation for the decennial redistricting in 2011. One of the judges then asked the Texas Solicitor General if that was correct. His answer was yes (but of course no one would do that). There is a difference between a partisan gerrymander at the beginning of the decade and one later on. The later one waits, the easier it is to account for population movement and therefore the ultimate success of the gerrymander. If holding mid-decade redistricting would in fact increase the likelihood of partisan gerrymandering at the beginning (and for reasons just stated we find this unpersuasive), then at least it forces a less reliable gerrymander.
By refusing to close the door on mid-decade redistricting, the Court opened it. Texas was first, but in 2006 Georgia followed suit. A year earlier, California's Governor Arnold Schwarzenegger offered a ballot initiative that would have redistricted that state, but the voters rejected it. Other states have indicated interest. Thus LULAC opens the door to rolling redistricting and not only on the Congressional level. Any body, from the state legislature to a school board, could take advantage of LULAC (except as limited by state law). Politicians will walk through that opening and this will exacerbate the corrosive cynicism that Americans have acquired believing that politics is rigged by and for the politicians.
Early in his opinion Kennedy saluted the replacing of the court-ordered 2001 plan with one drawn by politicians. "[T]he obligation placed upon the Federal Judiciary is unwelcome because drawing lines for congressional districts is one of the most significant acts a State can perform to ensure citizen participation in republican self-governance." Oh? There was no clamor in Texas for new congressional districts in 2003. The demand was from Washington, DC (by a Congressman who, when he would resign in 2006 announced that his home was in Virginia not Texas). Nor is there any basis for believing that the 2003 plan ensures citizen participation. Indeed, quite to the contrary, it was designed to thwart citizen participation. New Republican districts in North Texas stretch like pancakes from the Dallas suburbs eastward. Because party activists are more likely to vote in primary elections, these allow the suburbs to control the choices for the more moderate Republicans and Independents in East Texas. And since the districts were drawn so that no Democrat could have a chance of prevailing no matter whom Republicans nominated, the only election that matters is the Republican primary. If Kennedy believes what he writes, then he is living in his youth when we were taught that voters choose their representatives. The computer has put an end to that. Representatives choose their voters. This was done with care to ensure no fewer than 20 Republicans in the Texas delegation.
For those who believe good lawyering matters (and we do), LULAC is a gem. The Texas Solicitor General, Ted Cruz, out-lawyered the plaintiffs in the district court and by all accounts out-argued them before the Supreme Court. One result was that the Court never understood that the 1991 gerrymander occurred largely in Dallas and Houston with districts that were not at issue in 2003 when Republicans primarily targeted rural Democrats elected by ticket-splitting Republicans or Independents in reasonably compact districts outside these urban areas. Another was that Kennedy actually questioned whether the plan was truly partisan gerrymandering since "partisan aims did not guide every line" and "a number of line-drawing requests by Democratic state legislators were honored." Does he have any conception of how big (and diverse) Texas is? Of course not "every line" was partisan, but that hardly made an ugly gerrymander pretty and the Voting Rights Act (even if violated) was still a constraint (even if the professionals at the Justice Department were overruled by their political bosses).
At least five justices correctly intuited that there was something terribly wrong with what happened in Texas in 2003, but the Court missed its opportunity to preempt the ill-conceived (and anti-democratic) practice of rolling redistricting. The credit goes to Cruz, the blame to Kennedy who, like Justice Sandra Day O'Connor before him, relishes the power of the being the swing vote. O'Connor was the architect of many of the principles that have determined election law for the past quarter-century. Alone among members of the Court, she could draw on real life experience as an elected legislator. To be sure, some of her opinions, like Shaw v. Reno, appear inscrutable, while Kennedy's opinion, by contrast, is understandable even as it has an air of unreality about it. He has wound up where the conservatives are, but he either doesn't know it or doesn't want to admit it. Furthermore his newly found penchant for "compact" minority districts adds nothing to election law jurisprudence (unlike Scalia's with his conclusion that districts can be drawn using race as a predominant consideration when necessary to comply with the Voting Rights Act because compliance is a compelling state interest).
If, as we fear, state and local governments follow the precedent of Texas and engage in unnecessary, wasteful, and acrimonious redistricting using inaccurate population data to benefit partisan interests, we would like to believe the justices will regret their decision in LULAC v. Perry. But we doubt that will happen because there was also an air of unreality about the challenge to the 2003 gerrymander. This is, after all, the Court (and a justice) that assisted George W. Bush into the White House. Could anyone believe it was realistic to expect it (or him) to unseat six Republican Congressmen to assist Nancy Pelosi into the Speaker's Chair?