The Effect of Redistricting Commissions on Electoral Competitiveness in U.S. House Elections, 2002-2010

This research examines the various methods of electoral districting in the post-2000 Census years in order to determine the variation in competitiveness for subsequent elections to the U.S. House of Representatives. The evidence from this period suggests that “backup” and independent commissions resulted in more competitive districts, as measured by margin of victory and challenger win rate, after controlling for partisan and contextual factors.


Introduction
Most advocates for electoral reform in the U.S. emphasize the lack of competition in U.S. House races, and subsequently argue that redistricting commissions are the optimal way to remedy the problem of incumbency advantage in U.S. House elections.Gary Jacobson notes in The Politics of Congressional Elections (2009) that competition in US house races is declining, as measured by incumbency strength, reelection rates, and margins of victory.Additionally, as Mann (2005, 92) notes, "….recent congressional contests suffer from an unusually high degree of incumbent safety, (and) a precipitous decline in competitiveness."Groups such as Common Cause (2013), the Brennan Center for Justice (2013), and the Annenberg Center (2013), argue that independent commissions are more likely to promote competitiveness and less partisan gerrymandering.
As a response to these trends, more states have turned to a variety of measures, including constitutional amendments, rules changes, as well as the establishment of redistricting commissions.By the time of the 2010 census, eleven U.S. states had adopted a type of commission, as compared to the more traditional legislative method, to conduct their redistricting process -representing 25% of all U.S. House districts.

Previous Literature
While scholars have debated the virtues and vices of electoral competitiveness, recent research has emphasized a need to change the current state of decreasing House competitiveness.Cain, McDonald, and McDonald (2005) have asserted that specific language is needed to promote and prioritize competition.They point to Arizona, Iowa, and Washington as examples of states that have competition mandated through their laws, saying that they "provide examples of the types of redistricting institutions that may foster the creation of competitive districts."Several legal scholars, such as Confer (2004) and Kubin (1997) have argues that redistricting commissions are a key factor in achieving the goal of more competitiveness.
The previous literature on this topic has been mixed in its findings about this assumed effect of redistricting commissions.Early research by Erikson (1972) and Ferejohn (1977) rejected the hypothesis that the redistristing was a major factor in decreasing marginal elections to the U.S. House.More recently, Mann (2005, 110) notes that commissions do not automatically lead to more competition, citing New Jersey and other states where commissions drew districts that protect incumbents "as efficiently as the normal legislative process."He goes on to outline the best practices in shaping commissions that would lead to more competition and responsiveness, these include, "specific instruction to promote competition…priority to…standards of partisan fairness and competitiveness (and) a fully transparent process" (Mann, 2005, 110-111.)Similarly, Abramowitz et al. (2006) argue that incresaed polarization of the electorate and incumbency advantage are more likely explanations for declining competitiveness in U.S. House elections.Masket et al.'s (2012) recent research shows very little difference in the levels of competitiveness in a commission-led process and the more traditional, legislature-led process.Winburn's (2011) analysis of the "hypothetical single-year swing ratio" of the 2002 elections suggests that the results are the same regardless of who is responsible for the redistricting process.Similarly, Masket et al. (2012) do not find evidence of clear long-term trends in levels of competition across the various redistricting methods.
However, it is difficult to parse out whether incumbency advantage is also affected by a more partisan redistricting process.Cain, Mac Donald, and McDonald (2005) and McDonald (2006) suggest that the protection of incumbents by state legislatures is one strong explanation for the lack of competition in elections.Similarly, Carson and Crespin (2004) suggest that commissions produce more competitive elections than when state legislatures controlling the redistricting process.

Research Design and Methodology
Since a number of states established these commissions after the 2000 census, it is now possible to re-examine the findings of this previous research by examining all five post-2000 Census elections as a time series.Our purpose is to determine if there are distinguishable differences in the rate of competitiveness across redistricting methods.This research uses a number of different measures of competitiveness to look for any differences by who drew the districts, starting with margin of victory, including a cross-sectional time series GLS regression to analyze the entire decade's worth of elections, controlling for the variation across elections.This methodological approach is more effective is addressing this effect of redistricting commission on competitiveness, as we are not restricted to one election year.This analysis thus provides a "window" of the past decade of redistricting, across a variety of methods, and the consequences, if any, for electoral competitiveness.

Hypothesis
H1: States with districts that are drawn by commissions will have more competitive races as shown by closer margins of victory.

Results
In the post-2000 round of redistricting, 28 states used the traditional legislative process to draw their congressional districts (261 districts); three used an independent backup committee when the legislature failed to draw suitable districts ( As shown in Table 1, a majority of House districts were drawn using the traditional legislative process.Another 15% were court-drawn districts, which Cox and Katz (2002) point out, often ultimately rely on partisan maps).After the 1.6% for single district states, the remaining 23.4% were drawn by some form of commission.).The medians also show the same pattern with traditional, advisory, and partisan methods with median margins of 35 each, and court-drawn districts equal to the overall median margin of 33.Finally, the use of backup commissions had the lowest median margin at 21, followed by single district states (24) and independent commissions (26).Such results, however, are limited by the unknown effect of other factors on competitiveness, such a state's overall partisan balance or uncontested seats in a particular election year.As such, we turn to a multivariate analysis, as shown in Table 3.
Table 3 is a cross-sectional time-series regression analysis of margins across redistricting method, controlling for open seats, and the absolute value of the PVI (a measure of the size of the pre-existing partisan balance), showing a y intercept of 22.52.The time series analysis allows us to distinguish variance that is caused by our hypothesized effect, and controls for variation in average margins across elections.The results show that backup commissions (CT, IA, and IN) had the largest effect on margin with a predicted 11.66 reduction in the average margins.Also, independent commissions (AZ, ID, and WA) had a major reduction in margins of around 10 points.Single district states had margins that were on average 12.31 closer.Districts drawn with court intervention had 3.59 point closer margins on average.Advisory and partisan commissions did not have statistically significant impacts on the margin of victory.Open seats were 14 .71points closer, and each point of absolute value of PVI (Partisan Balance) leads to elections with 1.5 point larger margins.The overall r-squared of 33.84 indicates that about 34% of the variation in the margins in U.S. house races from 2002-2010 was explained by the model.This supports the hypothesis that commissions, both independent and backup ones, have a statistically significant impact on the margins in U.S. House races, lowering them on average by 10-12 points.

Conclusion
The past decade has provided an opportunity to assess the impact of the various types of redistricting commissions, and this analysis supports the hypothesis that commission-led restricting may lead to closer U.S. House races.A key characteristic is to ensure that these commissions are independent, and that they require competition as a criterion in the redistricting process.Partisan and advisory commissions do not seem to have an effect on the margin of victory.The independent/backup commissions found in Arizona, Connecticut, Indiana, Idaho, Iowa, and Washington have led to margins of victory that are on average over 10-12 points closer than those districts redrawn under the traditional legislative process.In contrast to the more "null hypothesis" conclusions of previous researchers (often based on a single election), our analysis of the entire decade suggests that the type of redistricting process does indeed have a noticeable effect.
These results also suggest that if reformers are seeking to reduce incumbency advantage, it would be worth their effort to encourage redistricting using some form of independent commission, tasked with creating competition in the districts.While this change is likely to lead to only modest improvements in the degree of competitiveness, these closer races could lead to more accountable legislators.Even if incumbents continue to maintain their advantage, as we expect they will, this reform of the redistricting process could have a positive and measurable effect.

Table 1 .
Responsibility for the 2002-2010 US house districts

Table 2
This table shows the highest mean margins for the traditional legislative process (40.1%), advisory commissions (40.2%), partisan commissions (36.5%), and court-created districts (37.2%).The districts with below average margins were drawn by backup commissions (22.3%), and independent commissions (26.6% examines the mean margin of victory by who drew the districts, and shows that the average margin for the entire decade was almost 38%.The average margin was 26.7% for single district states that are not subject to gerrymandering.

Table 2 .
Margin of victory (MOV) across redistricting method

Table 3 .
Cross-sectional time-series regression of margin of victory U.S. house elections, 2002-2010