Author: Lukáš Kindl


Gerrymandering is a phenomenon accompanying some elections in the United States, since most of the state legislatures are responsible for the drawing of districts in the respective states. Many critics of gerrymandering argue that it distorts electoral results and some even add that it contributes to the polarization in the House of Representatives (HoR).

This paper focuses on gerrymandering in the case of elections to the U.S. House of Representatives and evaluates the real its impact on electoral results and on the polarization in this institution. The paper concludes that this effect is considerably less significant than widely anticipated.

The paper is divided into six parts. The first part describes the substance of gerrymandering and mentions its assumptions and constraints; the second part then outlines its various types. The third part deals with some other factors that might override the effect of gerrymandering. The fourth part specifies possible limitations of gerrymandering and discusses its effectiveness. The fifth part deals with the impact of gerrymandering on the polarization in the HoR. The last part focuses on majority-minority districts, which constitute a special case of gerrymandering for they contradict the findings from the fifth part.

Definition and Conditions of Gerrymandering

As the U.S. congressional elections are conducted in single-seat districts, it is more than probable that the distribution of popular vote between the two parties will not correspond to the distribution of seats among them. However, if only one party benefits from these distortions, we can talk about gerrymandering. In other words, in the absence of gerrymander, if one party wins a certain number of seats for a certain percentage of votes, the other party should also win approximately the same number of seats if it manages to gain the same percentage of votes in next elections.

Gerrymandering means deliberate shaping of districts by the majority party in order to maximize its electoral gain at the expense of the minority party. As a result, gerrymandered districts might have bizarre shapes.

In forty-four states, state legislatures are responsible for drawing borders of their districts for U.S. congressional elections.[1] [2] Thus, when one party controls both the legislature and the governorship, or, when the dominant party in the legislature is able to override the veto of the governor from the opposition party, it is able to redraw the districts as it sees fit and such redistricting is, therefore, called partisan. On the other hand, when each chamber of the state legislature is in different hands, or, when one party controls the governorship and the other one the state legislature, it results in bipartisan redistricting, which is considered as beneficial to both parties.


The states can redraw their districts whenever they wish so. However, redistricting usually occurs after a national census that is conducted every ten years, since there is a need to redraw the districts to see if states lose or gain seats in the U.S. Congress. Moreover, redistricting following the changes in apportionment offers greater opportunities for partisan gerrymandering.

Up until the 1960s, election results could have been distorted because of malapportionment. It meant that state legislatures could create districts of very unequal size, so that some of them contained two or three times the populations of other districts in the state. As a result, sparsely populated rural or small town districts were overrepresented and heavily populated urban districts were underrepresented, as all of them had just one representative. However, in 1964, the Supreme Court ruled in the Wesberry v. Sanders decision that “one person, one vote” principle shall be applied when drawing the districts.[3] Since this ruling, the districts have been required to be of equal size, which reduced the opportunities of state governments to gerrymander the districts.

There are also other constraints limiting the scope of gerrymandering. The districts have to be compact, which means that there should be some regularity in district shaping, even though there is no agreed definition of such “compactness”. Thus, it is up to the U.S. Supreme Court to evaluate the compactness of allegedly oddly-shaped districts. In addition, according to the Department of Justice, the redistricters have to prove that they do not deliberately split districts, where minority groups constitute a majority. However, additional constraints are imposed by some states. Twenty states require congressional districts to be contiguous, twenty states have a condition that district boundaries follow as near as possible the lower administration units (counties, cities, municipalities), and, finally, nine states require that the new districts preserve the cores of the previous districts.[4]

Forms of Gerrymandering

Majority parties in state legislatures use various types of gerrymandering in order to maximize their gains in elections to the U.S. Congress. These include packing, cracking, and, according to some, also kidnapping.

Packing means that the dominant party, which oversees the redistricting process, packs the electorate of the opposition party into as few districts as possible hence making these districts “super-safe”. As a result, the opposition party achieves a sweeping victory in these districts, but a large number of votes are wasted. These votes are then missing in the surrounding districts, which enhances the chances of the majority party there.

In case of cracking, the majority party splits the opposition party’s strong base in that it dilutes its voters into several districts where they constitute a minority. Although this disadvantages the majority party candidates in these districts as they are expected to achieve victory by lower margins, the majority party manages to deprive the minority party of its safe district.

Finally, the occasionally used term of kidnapping[5] refers to a redistricting technique directed against opposition party’s incumbents. The majority party redraws the districts in such a way that it moves the incumbent’s established electorate out from under him or her. As a result, the incumbent is situated in new district where he or she has to face another incumbent from the same party.


Gerrymandering has different impact in various types of districts. Donald Ostdiek[6] identifies four types of districts: 1. safe majority party districts, 2. marginal majority districts, 3. safe minority districts, and 4. marginal minority districts. Types 1 and 3 are generally safe districts for the majority or minority party, respectively. Types 2 and 4 are quite competitive, but leaning toward the majority or minority party, respectively.

According to Ostdiek,[7] the majority party controlling the redistricting tends to weaken its own incumbents in their safe districts (type 1) in order to help their candidates in less certain districts (type 2) in that it moves some portion of its electorate from safe to marginal districts. They also move some portions of their voters from safe districts to those that are leaning towards the opposition party (type 4) in order to reverse the balance in their favor. In a similar way, the minority party’s safe districts (type 3) are weakened. In states where apportionment increases, the electorate from the majority party’s safe districts might be used to create new districts.

How to Identify Gerrymandering

If we watch the actual outcomes of the elections, it might not be possible to identify intentional gerrymanders simply by comparing the results of several consecutive elections. Erikson argues[8] that as there are also other factors affecting the vote shifts, such as national trends, funds available for campaign, incumbency advantage, and higher or lower turnout, it is necessary to distinguish the net effect of gerrymandering from the remaining factors. These might either hide the gerrymander, if they advantage the opposition party, or, accentuate it even more, if they advantage the majority party.

Thus, what might, according to the election results, seem as an end of gerrymander might be, in reality, only a consequence of some of the above mentioned trends. In such a case, if the gerrymander really disappeared, the election results would be even more positive for the party disadvantaged by gerrymandering.

Erikson[9] suggests that we might distinguish the effect of gerrymandering by excluding the circumstances specifically related to congressional elections in the particular districts (incumbent advantage or funds available for campaign) in such a manner that we use the results of presidential elections in the respective districts. He also points out that we have to take into account eventual over- or underrepresentation caused by lower or higher turnout in some districts. If, for example, the turnout in Democratic districts is lower than in Republican districts, the net effect of this will be that the Republican share of seats is lower than its share of the popular vote.

Ostdiek[10] proposes an approach that looks retrospectively and simulates what the results of the last elections before redistricting would be under the new redistricting conditions in order to see the shifts caused by gerrymandering. This is, he argues, a way how to discover the real intention the legislators had when they redrew the districts.

As expressed by both authors, when the partisan legislatures redraw the districts, they do so according to the data available at that time and with an eye on the previous election results. Thus, they cannot take into account changes and trends that may occur afterwards.

The Limits of Gerrymandering

Despite the popular wisdom that gerrymandering can significantly alter the election outcomes, the choices of partisan state governments are very limited and all of them have their disadvantages.

First, it should be noted that a unified state government, the necessary condition for partisan gerrymandering, occurs less frequently than a divided government. Furthermore, in states with only one congressional district, gerrymandering is obviously impossible, and similarly, states with more than one, but still very few districts, do not encourage gerrymandering either. Thus, the actual number of states, where partisan gerrymandering is feasible, is quite limited.

During the last round of the redistricting process in 2000, twenty-three states had unified governments and, with regard to the above mentioned constraints, only thirteen of them had the potential for partisan gerrymandering. In 1990, there were only twenty states with unified state government, whereas sixteen of them had this potential.[11]

Partisan redistricters have to decide whether they seek incumbency protection and, therefore, need to reinforce the safety of marginal districts or, whether they intend to capture more seats, hence creating as many marginal districts – where they maintain only slight majority – as possible. Michael Lyons and Peter Galderisi[12] suggest that the partisan redistricters usually resort to a compromise between these two contradictory goals. They also argue that bipartisan redistricting provides better protection for incumbents than partisan redistricting does. Concentration on incumbent safety is, obviously, an acceptable agreement for both parties, because any redistricting aimed at capturing additional seats by either party is unthinkable.

Lyons and Galderisi[13] also evaluate the impact of reapportionment on partisan redistricting. They find out that it is much easier to capture new seats in states that gain seats, since divesting incumbents from safe districts of some portions of their electorate, which is necessary to create slight majorities in new districts, is less risky as the total size of the constituencies usually decreases anyway. On the other hand, in states that lose seats, gerrymandering aimed at capturing of new seats is very unlikely, because it would endanger the incumbents of the majority party, who already have to deal with some portions of unfamiliar constituents, which are added to their districts as a consequence of the decline in the number of districts. However, in this case, gerrymandering can effectively capture the seats held by opposition incumbents, so that the lost seats will be the opposition party’s seats.

Nicholas Seabrook[14] deals with the long-term negative effects of gerrymandering on the party that controlled the redistricting process. Since the goal of gerrymandering is to secure more seats for the redistricting party at the expense of their safety, while packing the opposition electorate into as few districts as possible, the newly captured districts are, according to Seabrook, likely to be lost over time, as the districting party enjoys in these districts only marginal majorities and they hence become more competitive. When some major shift occurs, it is, thanks to the gerrymander, much easier for the opposition party to capture these districts. Such shift may result from changes in voting blocs, migration patterns, or swings in national popular vote. It means that even though the redistricting party is expected to achieve some gains in the short term, this effect is likely to disappear in the long term and might even result in backlash.


The following data prove this theory. In the 2006 midterm elections, the Democrats gained thirty seats, whereas ten of them came from states where the Republicans controlled redistricting in 2000. In 2008, the Democrats won additional ten seats from these Republican states, while in states where they had controlled the redistricting process, they lost two and gained three. On the other hand, twenty-one out of the fifty-four seats the Republicans gained  in the 1994 midterm elections came from states where the Democrats controlled the redistricting process in 1990.[15]

However, this data also shows that the long-term loss of gerrymandered districts occurs only when there is a swing in favor of the opposite party within the respective decade. Although both parties were able to control the 2000 redistricting cycle in a similar number of states,[16] only Republicans were significantly hurt by the long term backlash. Thus, I would argue, that if there is a major national swing only towards one party within a decade, the gerrymanders produced by the party that will benefit from this swing might be effective for the whole decade.

It should be also pointed out that redistricting does not necessarily have to take place only after a decennial census. There are some examples of mid-decade redistricting, a trend that has risen in the 2000s. The 2003 Republican redistricting in Texas, which enabled the Republicans to gain six seats in the 2004 elections, or in Georgia in 2005, with no gains in the 2006 midterm elections, are cases in point. In both cases, the redistricting occurred shortly after the Republican Party managed to win majority in both chambers of the state legislature.[17]

Given this new phenomenon, I would argue that the parties might deal with the negative long term impacts of gerrymandering through the midterm redistricting, but only as long as they enjoy the control of the state. If the majority party, for example, sees its preferences declining nationwide, it might decide to redraw the districts so that they take a similar shape as before the previous districting.

Gerrymandering and Polarization

Given the ever-increasing polarization of the U.S. House of Representatives, manifested in the growing partisan difference in congressional voting, and gradually decreasing electoral competition, manifested in the higher number of reelected incumbents as well as in the lower number of districts with close results, many people, especially journalists, argue that this development results from increasingly more sophisticated techniques of gerrymandering.[18] They argue that because of the high number of safe seats, politicians do not have to reach out to moderate voters and move instead more to the right or left in order to avoid being replaced by more extreme candidates. While this might be true, it does not necessarily mean that it results from gerrymandering.

As described in this paper, gerrymandering is usually aimed at capturing additional seats, thereby decreasing the safety of the majority party’s districts. Thus, it has actually the opposite effect, as it encourages competition in marginal majority districts rather than helps incumbents to retain their seats. Ironically, bipartisan redistricting is more likely to produce polarization, since it seeks to protect incumbents from both parties, as mentioned above.

In their analysis of the impact of gerrymandering on polarization, McCarty, Poole and Rosenthal[19] suggest that gerrymandering cannot explain the polarization of the House of Representatives, since the Senate, which is not affected by gerrymandering, has become evenly polarized. They also point out that the distribution of the presidential vote across congressional districts resembles that across counties, and that, according to the statistics, new Representatives tend to vote more extremely than the incumbents and, similarly, legislators from safe districts tend to have less extreme voting records than their counterparts in marginal districts.

McCarty, Poole and Rosenthal[20] argue that the polarization stems from two factors: intra-district divergence and sorting. Intra-district divergence means that Republicans and Democrats representing the same district have increasingly more different voting records. Sorting means that the congressional districts become more homogenous for which there is an increased congruence between the district’s characteristic and its legislator. However, they claim that this homogeneity is based on geographical lines and natural distribution of both parties’ electorates, whereas gerrymandering might play only a modest role.

They demonstrate this on a model which simulates random redistricting and resulting polarization. In this model, if the districts are created from a random nationwide sampling that ignores state borders and other geographical, legal or political constraints, the resulting level of polarization is by 33% lower than if the districts are created from random sampling within each state. It means that differences between states account for one third of the polarization. If the authors include criteria of compactness and contiguity into the model, the results produce even higher level polarization.[21] [22]

Majority-minority Districts

However, there is one exemption. In case of majority-minority districts, gerrymandering might have significant effect on the results and contribute to polarization. After the 1990 census, the Department of Justice, acting under the mandate of the 1965 Voting Rights Act, issued directives to states to create congressional districts for minority groups. However, the Supreme Court, in a set of decisions, declared that if race was the dominant factor in drawing of the districts, the act is unconstitutional.[23] Thus, in practice, only districts that were not oddly shaped were usually able to survive.

Since majority-minority districts ensure a safe election for minority candidates who are usually left-leaning Democrats, they move the Democratic Party more to the left and polarize the climate in the House. If the majority-minority districts are super-safe, they actually produce two winners: the minority Representatives from the Democratic Party (usually members of the Black Caucus) and the Republicans, who benefit from the packing of minorities into as few districts as possible. The main losers are white Democrats, whose total number of seats might decrease.[24]

The opponents of majority-minority districts argue that it reestablishes segregation, as it discourages candidates who try to build multi-racial coalitions, and that it protects the incumbents and hence contributes to their radicalization. In addition, white voters are discriminated. On the other hand, the supporters suggest that it is the only way how to give minorities some Representatives. They argue that, for example, black candidates would otherwise have no chance under the first-past-the-post system, because white voters would not vote for them simply because of their color.[25]


Given the geographical, legal and political constraints, gerrymandering offers only limited possibilities how to distort the election results. Even if it produces some short term benefits for the redistricting party, this effect is often likely to erode over time due to some major shifts in the electorate. However, gerrymandering might help the redistricting party if it is able to redraw the districts in case of these shifts.

The real impact of partisan gerrymandering on the polarization in the House is negligible as its primary goal usually is to capture more seats, not to increase their safety. On the other hand, bipartisan gerrymandering might contribute to the polarization, since it often focuses on protection of both parties’ incumbents. In addition, majority-minority districts are supposed to contribute to the polarization, but such impact is limited given the number of these districts. As exemplified in this paper, geographical distribution of voters and regional differences account for much of the polarization.


[1] “Electoral Boundaries in America, Time to Bury Governor Gerry,“ The Economist, October 9, 2010, 12.

[2] The number of states where districts are drawn by state legislatures has now decreased to forty-three as voters in California passed Proposition 20, which handed over this responsibility to the California Citizens Redistricting Commission. See: Congressional Redistricting Initiative, (Accessed December 1, 2010).

[3] Schmidt, Steffen W. – Shelley II., Mack C – Bardes, Barbara A, American Government and Politics Today (Belmont: West/Wadsworth, 1997), 406.

[4] Seabrook, Nicholas R., “The Limits of Partisan Gerrymandering: Looking Ahead to the

2010 Congressional Redistricting Cycle,” The Forum 8, No. 2 (2010), Article 8: 3-4, (Accessed December 14, 2010).

[5] For instance, see: Peck, Don and Casey, Caitlin, “Packing, Cracking, and Kidnapping,” Atlantic Monthly 293, No. 1 (Jan/Feb2004): 50.

[6] Ostdiek, Donald, “Congressional Redistricting and District Typologies,” The Journal of Politics 57, No. 2 (May, 1995): 535.

[7] Ibid., 539-542.

[8] Erikson, Robert S., “Malapportionment, Gerrymandering, and Party Fortunes in Congressional Elections,“

The American Political Science Review 66, No. 4 (Dec., 1972): 1236-1241.

[9] Ibid.

[10] Ostdiek, Donald, “Congressional Redistricting and District Typologies,” 537-538.

[11] Seabrook, Nicholas R., “The Limits of Partisan Gerrymandering: Looking Ahead to the

2010 Congressional Redistricting Cycle,” 4-5.

[12] Lyons, Michael and Galderisi, Peter F., “Incumbency, Reapportionment, and U. S. House Redistricting,” Political Research Quarterly 48, No. 4 (Dec. 1995): 859-860.

[13] Ibid., 861-862.

[14] Seabrook, Nicholas R., “The Limits of Partisan Gerrymandering: Looking Ahead to the

2010 Congressional Redistricting Cycle,” 2-3.

[15] Ibid., 10-11.

[16] The Democrats controlled the redistricting process in nine states, whereas the Republicans in eight states. See: Seabrook, Nicholas R., “The Limits of Partisan Gerrymandering: Looking Ahead to the

2010 Congressional Redistricting Cycle,” 9.

[17] Seabrook, Nicholas R., “The Limits of Partisan Gerrymandering: Looking Ahead to the

2010 Congressional Redistricting Cycle,” 9.

[18] For example: Peck, Don and Casey, Caitlin, “Packing, Cracking, and Kidnapping.”

[19] McCarty, Nolan – Poole, Keith T. – Rosenthal, Howard, “Does Gerrymandering Cause Polarization?” American Journal of Political Science 53, No. 3 (July 2009): 667-668.

[20] Ibid., 667.

[21] Ibid. 673-676.

[22] This finding is similar to Erikson’s earlier observation of the era between 1950s and 1970s, in which he identified something what could be called a natural Republican gerrymander in the North. It resulted from the fact Republican voters were more efficiently distributed than the Democratic ones, who were clustered together in large cities. See: Erikson, Robert S., “Malapportionment, Gerrymandering, and Party Fortunes in Congressional Elections,“ 1242-1244.

[23] Schmidt, Steffen W. – Shelley II., Mack C – Bardes, Barbara A, American Government and Politics Today, 407-409.

[24] Miller, John J., “Segregation Forever?” National Review, May 14, 2001, 24-26.

[25] Stanton, Robert, “Minority Districts: Justice for Some, Gerrymandering for Others,” Headway 8, No. 3 (March 1996): 10-11.



Schmidt, Steffen W. – Shelley II., Mack C – Bardes, Barbara A. American Government and Politics Today (Belmont: West/Wadsworth, 1997).

Journal Articles:

Erikson, Robert S. “Malapportionment, Gerrymandering, and Party Fortunes in Congressional Elections.” The American Political Science Review 66, No. 4 (Dec., 1972): 1234-1245.

Lyons, Michael and Galderisi, Peter F. “Incumbency, Reapportionment, and U. S. House Redistricting.” Political Research Quarterly 48, No. 4 (Dec. 1995): 857-871.

McCarty, Nolan – Poole, Keith T. – Rosenthal, Howard. “Does Gerrymandering Cause Polarization?” American Journal of Political Science 53, No. 3 (July 2009): 666-680.

Ostdiek, Donald. “Congressional Redistricting and District Typologies.” The Journal of Politics 57, No. 2 (May, 1995): 533-543.

Peck, Don and Casey, Caitlin. “Packing, Cracking, and Kidnapping.” Atlantic Monthly 293, No. 1 (Jan/Feb2004): 50-51.

Seabrook, Nicholas R. “The Limits of Partisan Gerrymandering: Looking Ahead to the

2010 Congressional Redistricting Cycle.” The Forum 8, No. 2 (2010), Article 8: 1-16. (Accessed December 14, 2010).

Stanton, Robert. “Minority Districts: Justice for Some, Gerrymandering for Others.” Headway 8, No. 3 (March 1996): 10-11.

Newspaper Articles:

“Electoral Boundaries in America, Time to Bury Governor Gerry.“ The Economist, October 9, 2010, 12.

Miller, John J. “Segregation Forever?” National Review, May 14, 2001, 24-26.


Congressional Redistricting Initiative, (Accessed December 1, 2010).