In the beginning our Founding Fathers knew that England had a significant problem concerning the boundaries of election districts for the House of Commons. The district boundaries had evolved historically, but by the 18th century, those boundaries no longer reflected (if indeed they ever had reflected) the number of people living within each election district. It became so distorted that some election districts, known derisively as “rotten boroughs”, had fewer than 100 electors, and so those few citizens could elect (almost appoint) a Member of Parliament. The Founding Fathers sought a permanent fix.
As a reaction to England’s history, our Framers decreed that the election districts for the House of Representatives would be redrawn every 10 years based on a decennial census. Hence each state has a process for redrawing Congressional Districts every 10 years. In most states, the governor and state legislature control this process. If the same political party controls both the governorship and the legislature, then that party controls the process. Therein lies the problem.
The Practice of Gerrymandering
“Gerrymandering” is the practice of drawing the boundaries of an election district to favor particular candidates or a particular political party. In 1812, Massachusetts Gov. Elbridge Gerry engineered adoption of an obtusely shaped election district to favor his own political party. Cartoonists thought the shape looked like a salamander, and so the term “gerry-mander” was born – a conflation of the governor’s name plus “salamander”. Through much litigation over the next two centuries, courts have generally permitted strangely shaped election districts, as long as
- populations of each district are roughly equal,
- every district consists entirely of contiguous land areas, and
- no protected class of citizens (such as a racial minority) is unfairly marginalized as a result.
However, generally those districts that clearly favor one political party over another have been allowed.
In modern times gerrymandering has become a science in which incumbents draw boundaries that perpetuate the party in power. This is certainly one of the reasons that our current House of Representatives cannot get anything done, why incumbents are almost always re-elected, and why Congressional leaders cannot get Members of Congress to go along with any consistent program – no matter what the people want. It is also a significant factor in explaining why state legislatures are so resistant to change. The fact is that nearly every incumbent resides in a “safe” district, one in which the current Member of Congress has little need to worry about the opposition party. In fact, the only threat that many incumbents fear is a primary challenge from their own party. Gerrymandering is effective for both major parties in almost all districts in almost all states. For example, in the 2014 elections, 416 of the 435 members of the House of Representatives sought reelection, and 393 (or 94%) of them were reelected.
Partisan gerrymandering generally comes in two forms, called “cracking” and “packing”, both designed to disenfranchise voters in the minority party.
Cracking occurs when the dominant party in a state targets a community with a high concentration of voters from the minority party. To prevent that party from winning the election that that community might normally be expected to win, the dominant party draws congressional district boundaries that break up (or “crack”) that community into two or more partitions assigned to different CD’s in which the state’s dominant party still has a majority.
Packing occurs when the minority party might have a competitive advantage in several adjoining districts. To prevent the minority party from winning multiple elections, the dominant party artificially skews the district boundaries so that the minority party’s voters are packed into one district, where the minority party candidate will win by an overwhelming margin, but the dominant party’s candidates will win all the other adjoining CD’s.
In both cracking and packing, the result of partisan gerrymandering is that the power of the minority party’s voters is weakened or extinguished.
CD boundaries are also sometimes artificially rigged to reward a favored candidate or incumbent politician or to punish an unfavored one. The politicians in charge of drawing the boundaries have created unnatural districts so as to include or exclude the residence of a particular person, or to force two incumbents into the same district. Since voters often insist that their Congressperson live within the CD (while the Constitution only requires a Member of Congress to live within the state), a re-districting scheme that places two incumbents in the same district obviously disadvantages one or both of them.
In addition to disenfranchising large groups of voters, we can also point to the negative effect on individual voters. Ongoing research at the University of North Carolina demonstrates that citizen engagement of individual voters lessens significantly when a community is split into two or more congressional districts. Citizens in such split communities are less likely to 1) know who their Congressperson is, 2) contact or meet their Congressperson, or 3) believe that their Congressperson represents their own views or the views of their community.
Americans sometimes kid themselves into thinking that this is only a peripheral issue, and that most Congressional Districts are drawn pretty fairly with only a few disputable boundaries. So let me give you a few examples of egregious gerrymanders from the 21st century.
Following the 2010 Census, many state legislatures that have a single party controlling both houses of the state legislature as well as the governorship opted for extreme partisan gerrymandering in order to solidify the dominant party’s control of seats in the U.S. Congress. Here are three examples from North Carolina (Republican), Maryland (Democratic), and Pennsylvania (Republican). These three CD boundaries were all used in the Congressional elections of 2012 and 2014.
The U.S. Supreme Court threw out North Carolina’s 12th CD before the 2016 elections, stating that this was an unconstitutional racial gerrymander. North Carolina had concentrated black populations across the North Carolina Piedmont into this one district, so that the black community could elect one Congressman, but would not affect elections anywhere else.
Maryland’s 3rd CD, a truly contorted figure, is still in use in 2018. Democrats in the state legislature drew this district to protect John Sarbanes, the incumbent since 2007. Court challenges to these district boundaries continue apace.
Pennsylvania’s 7th district (like all Pennsylvania congressional districts) was redrawn by order of the Pennsylvania Supreme Court for the 2018 elections.
Compare these three modern examples with Gov. Gerry’s original “salamander-gerrymander”. We have obviously gotten much more sophisticated at drawing crazy boundaries to suit a particular political purpose.
Here is a rather dense mathematical analysis of gerrymandering, presenting one simple but extreme hypothetical example of how gerrymandering works and how bad it can get.
- In this fictitious example, a state’s 10,000 voters elect a legislative assembly with 100 members from 100 election districts with 100 voters per district. In an election held yesterday, 5500 voters (that is, 55%) voted for Party A, and 4500 voters (45%) voted for Party B. So Party A won the popular vote across the whole state by 10 points. What might the legislative assembly look like after this election?
- Without gerrymandering, Party A can be expected to win 70 seats, give or take. With a 10 point advantage in the popular vote, you might expect it to win every district by 10 points and thus win all 100 seats. Or, since Party A won 55% of the vote, you might expect it to win exactly 55% of the seats. But neither of these scenarios is at all likely for these reasons:
- In the first scenario Party A wins every seat, which can only happen if Party A’s majority is consistent across every district in the state. But such an even distribution rarely occurs because voters in both parties probably reside throughout the state with pockets of strength in certain areas. Further a party that loses every election will not remain a party for long, while a party that wins every election will soon split into factions and become two or more parties.
- In the second scenario Party A wins exactly 55 seats and wins 55% of the total vote. Party A would need to win those 55 seats by much more than an average of 10 points so that it could lose 45 seats by a small margin. If one party wins 55 seats, it’s much more likely that it will have ~51% of the popular vote.
- So the most likely outcome, without gerrymandering, is that Party A will win ~70 seats, give or take. It will win many seats by about 10 points, will win fewer seats by more than 10 or by less than 10 points, and will narrowly lose some number of seats. How many it wins depends on how evenly distributed Party A’s voters are across all the districts.
- With gerrymandering, consider what happens. Note that a political party wants to win not only a majority (in this case, at least 51 seats) but also, if possible, a veto-proof majority (in this case, 67 seats). Remember that Party A has a 55% to 45% popular majority, so Party A’s objective in drawing district boundaries is to maintain that 10 point spread in its favor in every district. That’s probably impossible, but it’s not improbable that Party A could concentrate Party B’s 4500 voters in 10 districts. Let’s say Party B wins those 10 districts 63% to 37%, thereby accounting for 630 of Party B’s voters. Then Party B’s remaining 3870 voters are distributed across the remaining 90 districts, giving it 43 votes in each of those districts on average. As a result, Party A wins a solidly veto-proof 90 seats in the legislature.
- But let’s say that the minority party, Party B, controls the election districts for whatever reason (perhaps Party B was the majority party the last time the election districts were drawn up). Party B has arranged these districts so that Party B’s 4500 voters are concentrated in 80 election districts, and Party B wins each of those contests by a small majority (52 votes to 48). In the other 20 districts, Party A receives 83 votes while Party B gets 17, so Party A elects just 20 representatives. As a result, Party B wins a solidly veto-proof 80 seats in the legislature. What a surprising result!
- The table below shows these results side by side.
|None||by Party A||by Party B|
|Districts won by Party A||70||90||20|
|Party A votes per district||65||57||83|
|Subtotal Party A votes||4550||5130||1660|
|Party B votes per district||35||43||17|
|Subtotal Party B votes||2450||3870||340|
|Districts won by Party B||30||10||80|
|Party A votes per district||32||37||48|
|Subtotal Party A votes||950||370||3840|
|Party B votes per district||68||63||52|
|Subtotal Party B votes||2050||630||4160|
|Total Party A votes||5500||5500||5500|
|Total Party B votes||4500||4500||4500|
This might be even more dramatic as a graph. Even though Party A wins 55% of the vote, if Party B controls the district boundaries, Party B could end up with 80% of the seats:
Why and How We Need a Fix, and What Clearly Will Not Work
Unfortunately, gerrymandering has become the common practice whenever either party controls the redistricting process in any state. By the way, we should not blame politicians for this state of affairs. For any species, the two most important biological imperatives are replication (that is, preservation of the species) and survival (that is, preservation of the individual). Politicians behave like any other species. We may hire them to do the public’s business; but, left to their own devices, they will focus first on their party’s and their own political survival. We the people are unlikely to change the nature of politicians, nor can we expect them to instigate change themselves, nor can we solve the problem by exchanging current politicians for new ones.
In earlier times, several states had multi-seat Congressional Districts. The reasoning behind multi-seat CD’s at that time was this: In some states, the majority party or ideology might hold a significant advantage in many parts of a state (let’s say 60% to 40%), but was the minority party in one area, often an urban area. If the state had single-seat CD’s, the majority party would win wherever it held the majority, but would lose in the one area where it was a minority. But if the state fashioned a multi-seat CD that included that one area where the majority party was in the minority, the majority party could win all the seats in that multi-seat CD.
The multi-seat CD’s came to an end with a 1965 reform, which mandated single-seat CD’s. The motivation for that reform was the desire to give minority populations the opportunity to elect at least one of their own. As we will see in Part II, we have a better alternative today.
To get back to the original issue: The systemic problem is that the process of redrawing Congressional districts after each census is largely in the hands of politicians, who stand the most to gain from gerrymandered districts.
A very good summary of the state of redistricting reform was published in 2015 by US News & World Report (https://www.usnews.com/news/articles/2015/12/01/redistricting-reform-gains-steam).
|An added thought indirectly related to gerrymandering: While politicians gerrymander districts to favor themselves or their own party, race, or ideology, they also attempt to inhibit voting among the groups they do not like. Voter suppression efforts include Voter ID laws, fewer polling places, limited voting hours, and similar restrictions – just like the poll tax of yesteryear, which we eliminated by the 24th Amendment to the Constitution. But what if we used the number of votes cast in each state in previous elections rather than population figures as a basis for apportioning House seats? Then states which inhibit voting would end up with fewer seats in Congress – probably a reasonable penalty for states that disenfranchise their own citizens.|
Applicability to State Legislative Districts
Gerrymandering may be even more significant when drawing district boundaries for state legislatures. This is true for three reasons: 1) State laws often affect citizens more directly than federal laws. Hence the fair representation of citizens in state legislatures is doubly important. 2) State legislatures usually control the process for drawing the boundaries of federal Congressional Districts. 3) Most members of Congress served first in their state legislature, so the state legislatures serve to a considerable extent as Congress’ J.V. Your local state delegate has a reasonable chance of becoming your future Congressman.
Because state legislative districts are so carefully drawn to favor the incumbent politician/party, many would-be candidates choose not to run for elected office at all. They can see that “the fix is in”, so there is little point in competing against a rigged system.
Fixing the gerrymander challenge with respect to state legislative district boundaries would go a long way toward repairing our damaged politics.
 Curiel, J. and T. Steelman, “How to measure the impact of partisan gerrymandering”. Presentation at Reason, Reform & Redistricting Conference, Duke University, January 2019.