Do Republicans Have More Control Now than a Decade Ago?

At present, the Republicans are not in a dramatically better position than they were 10 years ago, which is surprising, considering the magnitude of their electoral victory in 2010. The reason for this lack of change is that Republicans were also in a great position 10 years ago. Now, they control only three more critical states—North Carolina, Ohio, and Tennessee—than they did 10 years ago (numerically, they control more states, but these three are the most consequential). One might object that Republicans now control Georgia and Texas, but they have done so since the party conducted mid-decade congressional re-redistricting in these states in 2003. As well, one might scratch one's head about my classification of Ohio as a new acquisition, since Republicans controlled the state government 10 years ago, but at that time, incompetent management of political circumstances necessitated a supermajority vote in the state legislature, and a bipartisan compromise resulted (Riskin Reference Riskin2002).

The Democratic outlook does have a few bright spots. The party controls Illinois, which had split control 10 years ago, and there is also split control in Virginia, which had a unified Republican government 10 years ago. One might object that California—the largest prize—should be counted as a loss for the Democrats, because Democrats controlled California's government 10 years ago and a 2010 ballot initiative has since given redistricting authority to a bipartisan commission. However, a decade ago, the redistricting process resulted in a bipartisan compromise after Republicans threatened to put a redistricting reform amendment on the ballot (Lawrence Reference Lawrence2001). California politicians are clearly aware of how to draw an incumbent protection gerrymander, as only one congressional district and five state legislative seats changed hands during the last decade, despite a high degree of electoral volatility in American politics.Footnote ² The brokered 2001 deal was soon broken, however, when Governor Schwarzenegger placed the first—failed—redistricting reform initiative on the ballot in a 2005 special election. Eventually, state legislative redistricting reform was successful in 2008, and congressional redistricting reform was passed in 2010. Now, a bipartisan commission is likely to draw a congressional plan with a similar mean expected number of seats but greater variance than the current plan.Footnote ³ Thus, in terms of expected number of seats, Democrats are probably in a similar position as a decade ago.

A Florida ballot initiative was also adopted in 2010 to reform redistricting criteria, but not the process itself. One might reasonably suspect that these criteria will be binding, or at least the requirement to “utilize existing political and geographical boundaries.”Footnote ⁴ The reform criteria also require compactness and political fairness, but the courts have proven squeamish in enforcing these vague terms. It is difficult to tell how restrictive these criteria will be until actual maps are drawn, but my suspicion—based on my experience drawing thousands of districts following political boundaries in five Midwestern states (McDonald Reference McDonald2009)—leads me to believe that an unbiased application of the “existing political and geographic boundaries” criterion will produce a congressional plan for the state that is slightly biased toward Republicans, since Democrats are heavily concentrated in the southeastern portion of the state. As a caveat, this tendency of geographic constraints to benefit Republicans may not be universally true, as it depends on a state's geography and politics.

Republicans will gerrymander within this constraint as much as possible. A key to reaching optimality may be to determine how many times a political boundary may be split. Republicans may wish to group all splits of a boundary into a single split boundary, while Democrats may prefer to count each split individually. The former counting method permits Republicans to crack urban areas across several districts, but the latter method is not necessarily optimal for Democrats, as it can still sub-optimally pack their supporters into districts they overwhelmingly win (Cain Reference Cain1984). The upshot for Florida is that 10 years ago, Republicans had free rein over the state, but now, they are somewhat constrained, granting Florida Democrats a slightly better position than they had a decade ago.

Increased Electoral Volatility

A major change of the past decade is that electoral volatility has increased. As a consequence, less certainty exists about what constitutes a safe district. A Republican incumbent will want a seat that can withstand a 2006 or 2008 electorate—years favorable to Democrats—while a Democratic incumbent will want a safe district for a 2010 electorate—a year favorable to Republicans. The optimal strategy for an incumbent facing increased electoral volatility is straightforward: create a safer district. In the face of this increasing volatility, the optimal strategy for a party seeking to maximize either the number of seats won or the probability of winning a majority of seats is also to draw safer districts (McDonald Reference McDonald1999). Therefore, safe seats—safer than a decade ago—are the goal for everyone. However, tension may exist between incumbents and their party, since the optimal level of partisan strength for a partisan gerrymander is more competitive than the optimal level of partisan strength for an incumbent protection gerrymander (Owen and Grofman Reference Owen and Grofman1988).

Optimal Partisan Gerrymandering

Are enough Republican voters available to create safe seats for all Republican incumbents, particularly the newly elected and therefore vulnerable freshmen? Consider that Republicans now have lopsided margins in battleground states such as Florida (20:5 Republican seat advantage, with the 2010 gain of two seats as a result of congressional apportionment), Michigan (9:6 Republican seat advantage, with the 2010 loss of one seat), Ohio (13:5 seat advantage, with the 2010 loss of two seats), and Pennsylvania (13:7 seat advantage, with the 2010 loss of one seat); as well, the large but non-battleground state of Texas has a 23:9 Republican seat advantage after the gain of four seats. Republicans may be beyond optimal partisan gerrymandering in these states and may thus face tensions among their own incumbents as they all jockey for safe seats.

Republicans will have partisan gerrymandering opportunities elsewhere. Perhaps the biggest prize will be North Carolina, where Democrats retained an edge in the state's congressional delegation (7:6 Democratic seat advantage, with no apportionment change in 2010). A Republican gerrymander by the state legislature—which holds sole redistricting authority—could put as many as four Democratic incumbents at risk. In neighboring Tennessee (9:2 Republican seat advantage, no apportionment change), the unified Republican state government should at least be able to secure their three pickups and create a competitive Nashville area district. Democrats may be able to counter Republican mischief in Illinois (10:9 Republican seat advantage, with the 2010 loss of one Democratic seat). Here, Democrats may be able to put as many as four Republicans at risk by shuffling around their seats.

Geographical constraints may limit the political parties from pursuing theoretically optimal partisan gerrymandering. Two states illustrate these constraints. Although Maryland Democrats control their state's redistricting process, they can do very little to Republican Representative Roscoe Bartlett's Sixth District, which is composed primarily of the state's western panhandle. Texas Republicans will be constrained by a confluence of voting rights issues and geographical constraints in the Twenty-Seventh District, which is sandwiched between the Gulf of Mexico and the Fifteenth District, a Latino-majority district mandated by the Voting Rights Act. Here, freshman Republican Representative Blake Farenthold won a narrow victory with less than 50% of the vote, but there is little room to shore up Farenthold's re-election chances through redistricting.

The Voting Rights Act

Another important factor in redistricting considerations is the Voting Rights Act. The few Southern Democrats who remain in Congress predominantly represent safe, heavily Democratic districts required by the Voting Rights Act. Republicans may now face constraints on their optimal partisan gerrymanders in states such as Florida, Georgia (8:5 Republican seat advantage, after a 2010 gain of one seat), and Texas, where dispersing Democrats into Republican-majority districts would otherwise be a viable strategy. Democrats control the Department of Justice for the first time during a redistricting since the passage of the Voting Rights Act. The Department of Justice's judgment regarding what constitutes an effective minority district could affect the overall political character of redistricting plans, and it may be able to bring suit under Section 2—a general prohibition on voting discrimination—of the Act. Republican state governments seeking Section 5 preclearance of redistricting changes may avoid an aggressive Justice Department through an alternative route, the District Court for the District of Columbia. Republicans are likely to unleash a vigorous legal attack on the Voting Rights Act, but they may not want to commit this attack using a statewide redistricting plan as a vehicle, given that a federal court may then draw the map.

In the past, the Voting Rights Act predominantly constrained Democratic gerrymanders by forcing Democrats to inefficiently concentrate their supporters into heavily Democratic minority-majority districts (Grofman and Handley Reference Grofman, Handley and Grofman1998). Now, with more states under Republican control, the Voting Rights Act exerts more constraint on Republican gerrymanders by forcing Republicans to draw Democratic districts where they may otherwise not wish to do so (Schotts Reference Schotts2001). Perhaps the most vivid example of this scenario is freshman Representative Francisco Canseco, who represents Texas's Twenty-Third District, which includes the Big Bend area of the Rio Grande River and was specifically found by the Supreme Court to be protected under the Voting Rights Act. Representative Canseco won this district by less than 50% of the vote and can expect a vigorous challenge in 2012.Footnote ⁵ Perhaps the best outcome that Republicans will be able to achieve is to move Representative Canseco's home into a friendlier district, since shoring up this district with Republican supporters would invite a legal challenge.

New Population Data: Prisoners and the American Community Survey

New York Democrats may find themselves in a more favorable position during the state legislative redistricting process than they were in the last decade. New York is one of three states, along with Delaware and Maryland, that has passed legislation to adjust census population counts for state legislative redistricting by reallocating prison populations from predominantly rural prisons to the location that prisoners consider to be their home. Hawaii and Kansas have similar provisions regarding population adjustments for nonresident students and military. Hawaii's military bases in particular pose a difficulty for state legislative redistricting if these populations are not removed from the district's counts. A little-known fact is that in 1966, the Supreme Court ruled that Hawaii could use voter registration as the population base to conduct a redistricting, although it never did so.Footnote ⁶

These adjustments may prove to be contentious, especially if states attempt to adjust their counts for citizenship considerations. In the past, such an adjustment was typically not possible to make during the redistricting process, since citizenship was reported on the census long form, for which data were often released after the conclusion of redistricting. However, these citizenship data—and much more demographic information—are now available from the annual American Community Survey (ACS).Footnote ⁷ The legal question of using the citizen-voting age population (CVAP) as the population base to conduct redistricting is winding its way through the lower courts but has not yet been addressed by the U.S. Supreme Court(Reader Reference Reader1994).Footnote ⁸ A related issue concerns the intersection of the CVAP and Voting Rights Act requirements; lower courts have ruled that 50% of a district's CVAP is the required benchmark to demonstrate accordance with Section 2, although again, the U.S. Supreme Court has not made a ruling on this matter.Footnote ⁹ Further compounding the legal issues is that the Census Bureau has released measures of uncertainty to accompany the citizenship point estimates and the confidence intervals can be so large that they cover zero. Is a permissible district one for which the citizenship point estimate exceeds 50%, or is it one for which a confidence interval does or does not cover 50%? Is the 90% confidence interval released by the Census Bureau the appropriate level? We can expect further litigation on this subject.

More Public Participation

In the past, a lack of access to expensive redistricting software and elaborate redistricting databases prevented widespread public participation in the redistricting process (Altman and McDonald Reference Altman and McDonald2010). At this time, however, technology can enable anyone with an interest in redistricting to draw maps of his or her community, a district for the area, and even an entire statewide redistricting plan. Online resources are also available to teach people how to draw legal redistricting plans. My colleague Micah Altman and I are part of one such project that aims to freely distribute redistricting tools, data, and educational materials over the Internet.Footnote ¹⁰

One example of the potential public participation that may emerge from these newly available tools is a competition that has been created for teams of faculty and students at Virginia universities to produce redistricting plans for the state legislature. Besides serving as a pedagogical tool, I expect that this competition will engage the broader public as well. As evidence, consider that in 2009, members of the public submitted 14 plans to an Ohio redistricting competition held by the Ohio secretary of state.Footnote ¹¹ More people will likely participate if the maps have a demonstrable effect on the redistricting process.

Broad public participation in this process is a new development, so its effect is difficult to predict. Citizens' commissions in Arizona and California are required to hear public input, so it may be that public participation will most influence the types of plans produced in these states. There may also be a chance for public input into the process in Florida, where the state House is developing an online redistricting tool that will enable the public to submit a map to the legislature at a click of a button.Footnote ¹² Such participation may prove particularly important in the wake of the adoption of redistricting criteria in Florida—if someone in the public can draw and submit a map that best satisfies the criteria, courts may frown if the legislature does not consider his or her redistricting plan. Elsewhere, redistricting authorities may or may not consider the public's input, but where they do not, the media will now have benchmarks to compare plans produced through the political process. In instances involving the courts—and particularly in divided government situations—courts may add public plans to their menu of options under consideration.

Wide public participation may illuminate the possibilities of redistricting reform and the potential tradeoffs among redistricting criteria, which, in the aftermath of redistricting, may better inform future reform efforts and scholarly studies. Reformers learned a difficult lesson in Arizona that seemingly neutral redistricting criteria may have predictable political effects to the detriment of reform goals (McDonald Reference McDonald2006; see also Parker Reference Parker1990). Scholars often do not sufficiently appreciate how a state's geography and partisan distribution meaningfully constrain redistricting. Cursory reflection should lead to the conclusion that states such as Idaho and Massachusetts are dramatically different when it comes to redistricting. Strategic political considerations and how criteria may constrain gerrymandering thus vary among states, even though states are often considered equal units of analysis in scholarly studies. More redistricting plans will enable scholars and the public to better evaluate what is possible within a state, as well as the trade-offs between potentially competing goals. For example, the 2009 Ohio redistricting competition revealed that a significant tradeoff does not exist among the Voting Rights Act, equal population, partisan fairness, competitive districts, compactness, and respecting political boundaries. Twelve of 14 submitted plans equaled or exceeded the current congressional plan on all of these criteria.