Introduction

This essay examines the electoral system we know the least:  proportional voting.  As noted previously, John Stuart Mills, along with several other political theorists, devised this system in the 19^th century.  It became widespread as liberal parties in Europe struggled to maintain relevance with the growth of socialist and worker parties.  Proportional voting allowed liberal parties as well as conservative parties to win seats without having to attain a majority of votes.  Beginning in the early 20^th century, the U.S. saw a similar interest in proportional voting.  The Progressive Movement made it a priority on its policy agenda, and other organizations such as the Proportional Representation League promoted the idea.  A number of U.S. cities adopted a form of proportional voting beginning in 1915 but efforts waned over time.

Unlike the single round simple majority system, proportional voting is a relatively recent arrival to democracy.  It lacks the intuitive appeal of majority voting.  We will see that from an operational standpoint, it performs as a near mirror opposite of majority systems.  It encourages and thrives on multiple parties.  It does not typically result in decisive electoral outcomes that lead to majorities, but instead, it requires coalition building.  For this reason, critics worry about the ability of proportional voting to generate functioning governments.  By encouraging voters to express their preference rather than to vote strategically (i.e., to prevent a less desirable outcome), proportional voting produces certain advantageous social behaviors in a democratic system.  These behaviors provide a strong foundation for democracy borne out by the fact that democracies using proportional voting function well in today’s environment as compared to majority ones.

Duverger’s Second Law

As described in the last essay, Maurice Duverger is credited with identifying the connection between the simple majority single round voting and two-party systems.  Known as Duverger’s Law, this theory says that majority systems cause voters to select candidates who may not be preferred but are the most likely to defeat the candidate they like the least.  This “psychological factor” encourages a two-party system by polarizing the electorate, which undercuts the rise of third parties.  Duverger’s second law gets much less attention.  It relates to proportional voting systems.  While a simple majority system tends to generate and sustain a two-party system, a proportional voting system tends to generate and sustain a multi-party system.

In Political Parties, Duverger recounts the example of Belgium.  Socialists made great strides at gaining seats in the 1890s to the detriment of the Liberal Party.  The conservative Catholic Party could see what was coming under the existing majority voting system:  within a few short election cycles it would face the Socialist Party on its own.  Even worse, if the Socialist Party gained a majority of seats, the Catholic Party would lose any seat at the table.  In response, the Catholic Party instituted proportional voting, helping the Liberal Party to make a comeback.  That allowed supporters of the Liberal Party to avoid having to make the difficult choice imposed by a majority voting system between the Socialists and Catholics.  Proportional voting ensured that several parties had a seat at the table.

Duverger does not provide much detail about the relationship between proportional voting and multi-party systems.  He simply points out that multi-party systems arise naturally in the absence of strategic voting.  He writes:

The polarization of the single-ballot system is pointless under proportional representation where no vote is lost (at least in theory); hence we have the opposite process of ‘depolarization.’  The first effect of proportional representation is therefore to put an end to any tendency towards a two-party system; in this respect it may be considered as a powerful brake.

He continues that with proportional voting:

There is no encouragement to parties with similar tendencies to fuse, as their division does them little or no harm.  There is nothing to prevent splits within parties, for the total representation of the two separate factions will not be mechanically reduced by the effect of the vote; it may be psychologically, by the confusion it sows among the electors, but the ballot plays no part in this.

In essence, political parties react to an environment that does not penalize them for receiving less than a majority of the vote.  They no longer have to form broad, unstable coalitions to win.  They can afford to shed certain constituencies that strain their core philosophy or identity.  Duverger notes, “The only attenuation of the fundamental tendency to preserve an established multi-partism comes from the collective nature of proportional representation: the party must have organization, discipline, structure.”  In other words, parties in a proportional system still must compete as a viable enterprise with other parties – they simply do not have to maneuver within a context of polarity between two dominant factions.

Types of Proportional Voting Systems

The idea of proportional voting systems arose after the implementation of majority systems.  As European nations gradually moved toward democracy in the mid-19^th century, they benefited from observing and thinking about how electoral systems worked in practice.  Some political theorists thought deeply about the meaning of representative government, as we will see in the next essay, and this drove experimentation with new ideas.  These thinkers could see the benefits of a proportional system.  However, given the complexity of this approach, it took trial and error and significant theoretical analysis to understand how a proportional system could be implemented in an election.  By the latter part of the 19^th century, many of the details were resolved. Ultimately, it took the self-interest of party leaders under threat from rivals to spur the adoption of these systems at the dawn of the 20^th century.

Fundamentally, proportional voting seeks to translate a party’s share of the vote into a corresponding share of seats in the legislature.  In other words, a party that receives less than a majority or plurality of votes may still gain seats in proportion to its share of votes (e.g., a party receiving 30% of the votes wins 30% of the seats).  To work, proportional voting requires multi-member districts.  Typically, the number of votes needed to win a seat in a multi-member district, referred to as a threshold or “quota,” is a function of the vote total divided by the number of seats.  For example, if 100,000 votes are cast in a district with ten seats, a party must win at least 10,000 votes to win a seat (i.e., the quota is 10,000).  This formula ensures that the number of winners does not exceed the number of seats in the district, assuming that votes are evenly distributed among all parties or candidates.  Of course, most elections do not result in the even distribution of votes.  Usually, there will be leftover votes above quota for some candidates and parties.  Therefore, proportional systems must include a mechanism to allocate leftover votes until all seats are filled.

System designers refer to the number of seats in a district as “district magnitude.”  District magnitude has several operational impacts on elections and the performance of a democracy.  The larger the district magnitude, the more seats to fill in an election.  An entire country can form a single district with many seats to fill.  This is the case in Israel and the Netherlands, which have 120 and 150 seats respectively.  The greater the district magnitude, the greater the proportionality.  The greater the district magnitude, the easier it is for parties to make quota and gain seats.  Therefore, the greater the district magnitude, the more political parties represented in government.  For example, in Israel or the Netherlands, it is feasible for a party to win a seat with as little as 1.5% of the vote, allowing smaller parties to win seats.

There are two main types of proportional voting systems:  the Single Transferable Vote (STV) and the list proportional representation (List PR) system.  The STV system is only used in a few instances – mostly in countries with a connection to Great Britain.  On the other hand, the List PR system is the most popular system in the world with most new democracies utilizing it as well as most countries in Europe, Latin America, and Africa.

Single Transferable Vote

Several political theorists created the STV system independently during the 19^th century.  Even though it predated the List PR system and is a favorite of political scientists, its use remains limited.  The Republic of Ireland provides the best-known example.  Other applications include Malta, the Australian Federal Senate and a handful of local jurisdictions in New Zealand, Scotland, Northern Ireland and British Columbia.   Thomas Hare, a political philosopher and member of British Parliament, is most associated with the STV system.  Hare wrote Treatise on the Election of Representatives between 1859 and 1873.  In the preface, he wrote that proportional representation would “end the evils of corruption, violent discontent and restricted power of selection or voter choice.”  Hare was a contemporary of John Stuart Mills, who extolled the virtues of STV, describing it as “the greatest improvement of which the system of representative government is susceptible; an improvement which … exactly meets and cures the grand, and what before seemed inherent, defect of the representative system.”

In essence, STV employs a key aspect of the Alternative Vote (AV) system (also known as preferential or ranked choice voting). Like the AV system, voters rank candidates on the ballot according to preference.  However, STV does not require a candidate to receive a majority of the votes to win a seat.  Instead, STV uses multi-member districts and a quota, which makes it a proportional system.  Voters rank individual candidates listed on the ballot.  Candidates who reach the quota fill a seat.  If a round of counting produces no winner, the candidate with the lowest number of votes is dropped and counting continues until all seats are filled.  STV scores well with political theorists because preferential voting allows seats to be filled by counting ballots in successive rounds rather than using a formula to fill seats as is necessary with the List PR system.  But voters are not required to rank all of the candidates on the ballot.  This can result in the disqualification of ballots during subsequent rounds of counting.  In some cases, it can result in candidates winning seats even though they have not achieved the quota.

As noted, very few countries use STV and only two use it for their lower house:  the Republic of Ireland and Malta.  Both of these countries have small populations.  The Republic of Ireland’s population is less than half that of North Carolina’s, and Malta has about 500,000 people.  The story behind the adoption in Ireland informs our thinking about STV systems.  The British pushed for a proportional system at independence in 1922 to ensure the Protestant minority would have a voice in the parliament or Dail.  As previously stated, the British never adopted proportional voting so had little familiarity with List PR.  Instead, STV was selected given its status among British political scientists.  There are around 166 members of the Dail and roughly 40 constituencies or districts.  That means each district has four or five seats.  Consequently, the district magnitude sustains four or five parties that win most of the seats in the Dail.  The largest party, Fianna Fail, initiated a referendum in 1959 and 1968 to replace STV with a majority system, reflecting the natural tendency of larger parties to eliminate competition from smaller parties.  Both referenda failed.

Through Ireland’s post-independence history, two parties have dominated its politics:  the centrist, Fianna Fail, and center-right, Fine Gael.  In recent election cycles Sinn Fein has made significant gains at the expense of Fianna Fail.  Given that voters can choose among candidates within a party, Ireland’s STV system creates intense intra-party competition.  Critics claim that incumbents focus on constituent service at the expense of broader policy matters that effect the nation. One reason for this is the ratio of representatives to population.  It is 1:20,000 in Ireland as opposed to 1:50,000 for state house districts in North Carolina and 1:750,000 for members of the U.S. House.  This ratio maintains a close connection between members of parliament and voters.  Given the size of Ireland (less than half the population of North Carolina), it is difficult to draw too many conclusions.  Regardless, the Republic of Ireland is ranked among the top 10 democracies in the world by the Democracy Index, which will be described in more detail later.

List PR System

Like other proportional voting systems, List PR originated in the 19^th century as political theorists looked for alternatives to majority voting.  The first clear description of the List PR system came from Victor D’Hondt in Belgium.  He described the system in 1878, and Belgium adopted the system for its parliamentary elections in 1900.  It quickly expanded in Europe for the reasons described previously.  Now, the List PR system is the most popular electoral system in the world with about 35% of democracies using it (in contrast, 24% of democracies use a first-past-the-post system).  It is also interesting to note that among the approximately 30 nations that have undergone electoral reform in the last 30 years, most have moved from a majority system to a List PR system or a system with more proportional elements.

In its most basic form, each party provides a list of candidates for the seats in a multi-member district.  Electors vote for a party’s list.  Parties are allocated seats based on their vote total.  Lists can be open or closed.  With an open list, voters can select among candidates on a party’s list.  With a closed list, voters must select a party’s list as presented.  System designers find that districts with between three and seven seats work well.  These systems sustain a manageable number of political parties.  Some countries by law set a minimum threshold to win a seat.  For example, Germany and New Zealand require that a party win at least five percent of the vote nationwide to gain a seat in parliament.  This rule is intended to reduce the role of fringe groups in government.

The List PR system requires technical support not present with majority voting systems.  A common issue concerns the assignment of leftover votes or votes not used to meet the quota for a seat.  There are several ways electoral system designers translate leftover votes into seats.  The most popular includes the Highest Average method and the Largest Remainder method.  The Highest Average method requires the number of votes received by each party be divided successively by a series of divisors.  This produces a table of averages.  The table allocates seats based on the divisor until no open seats remain vacant.  This method tends to favor larger parties because it skews upward the threshold required for attaining seats. Other systems use the Largest Remainder method.  When some seats remain available because no party exceeded a required threshold, the remaining seats are awarded under this method to parties in order of the number of left-over votes they have.  The approach can help smaller parties pick up seats.

Experience with List PR in New Zealand

New Zealand provides an interesting example of a variant of the List PR system.  As a member of the British Commonwealth, New Zealand inherited a majority voting system.  New Zealand followed a similar pattern to Great Britain.  A rising Labour Party supported proportional voting at the dawn of the 20^th century.  The Liberal Party resisted reform until it was too late, and the Labour Party replaced it as the second major party.  Concentrated in urban areas, the Labour Party suffered from a disproportionate number of wasted votes.  Finally, the nation experienced highly distorted outcomes in two successive national elections in 1978 and 1981.  In both instances, the conservative National Party retained an absolute majority of seats in the House of Representatives even though the Labour Party won more votes.

These anomalous outcomes led the Labour Party to establish a Royal Commission on the Electoral System after gaining power in 1984.  The Commission studied a number of electoral systems and issued a report in 1986, recommending adoption of a Mixed Member Proportional (MMP) system similar to Germany’s.  As the Labour Party saw its electoral fortunes improve in the late 1980s, it backed away from the recommendations.  The National Party saw a political opening.  In its 1990 election manifesto, it promised a referendum on the Commission’s recommendations.  The National Party regained a majority in 1990, and then it too tried to back away from reform.  It took substantial public outcry during a recession to pressure the National Party to deliver on its promise.  The government presented two referenda.  The first asked voters whether they supported “a change to the voting system.”  It passed with nearly 85% in 1992.  The following year, the public was presented with four options to replace the majority system.  A strong majority approved the proposed MMP system.  There was one final binding referendum in 1993 in which both major parties vigorously opposed the MMP system.  It still passed easily, and the new system finally went into effect in 1996.

Under the MMP system, electors have two votes.  First, they vote for a candidate in a single member district under a majority voting system. Second, they vote for a political party at the national level.  The MMP system takes the results from the national vote and allocates seats to the respective parties proportionally.  For example, if a party wins 25% of the party votes, it should receive 30 seats in the 120 member Parliament.  If that party has won 20 seats through the single member district vote, the MMP system allocates an additional 10 seats to the party to achieve parity with its party vote.  Like some other proportional systems, New Zealand imposes a threshold.  To win a share of the party votes, a party must exceed 5% of the national vote or win at least one single member district.  Since adoption of the MMP system, multiple parties have met this threshold – usually five or so.  Importantly, the proportionality index fell from an average of 11% prior to reform to an average of 3% since the reform (with 0% being perfectly proportional).

Politics in New Zealand has now adapted to the proportional system.  No longer does one party gain a majority of seats in government.  While the two major parties continue to garner most of the seats, they must form coalitions with other parties to reach a majority.  This requires compromise and collaboration. The accommodation has generated results.  For example, parliament has had no problem passing budgets.  After decades of deficits, the country has run fiscal surpluses.  Since the reform, no government has suffered from a no confidence vote.  Voters report greater satisfaction with government.  The National Party did manage to force one more retention vote on the MMP system in 2011.  It passed with nearly 60% of the vote.  The Democracy Index now ranks New Zealand as the fourth strongest democracy in the world.

Participation

Under Rousseau’s framework for democracy, proportional voting systems achieve high marks.  Interestingly, the attributes of proportional systems tend to be nearly the mirror opposite of majority systems.  As noted previously, Rousseau laid out a vision for democracy where the will of the people is revealed through participation, formation of majorities, shifting coalitions, equality and choice. These attributes make for a healthy democracy.  Surveys show that voter turnout in proportional systems is higher than majority ones.  With fewer wasted votes, voters believe their vote has a greater chance of helping their candidate reach the quota needed for a seat.  In addition, parties play a more prominent role in List PR systems.  As we saw earlier with the “calculus of voting,” parties drive turnout by reducing the cost of voting.

Voters in proportional systems also say in surveys that elected officials are more responsive to their interests, which may also drive-up participation.  With multi-member districts, parties seek votes wherever they can get them.  A vote for the conservative party matters just as much in the city as it does in the countryside.  The key is to exceed the quota needed for a seat.  The need to pursue votes wherever they are enhances voter satisfaction and participation in elections.  In contrast, majority voting systems more often have noncompetitive races because single member districts enable the manipulation of district boundaries through gerrymandering.  For example, only about 10% of the legislative races in North Carolina have been competitive in recent election cycles.  When the election is predetermined, voters have less motivation to turn out.

Of course, the ballots for proportional voting systems are more complex by virtue of multi-member districts.  Rather than voting for one candidate per office, a voter can be faced with multiple choices for a district.  In addition, the STV system adds a level of complexity due to the ranking system.  After the fall of the Berlin Wall, Estonia established an STV voting system.  The country abandoned it after one attempt in 1990, finding it confusing and instead switched to the List PR system.  In sum, proportional systems, particularly the List PR system, encourage strong participation despite the increased ballot choices presented to voters.

Formation of Majorities

The major criticism of proportional voting centers around its performance post-election.  As noted in the last essay, a majority system claims the ability to produce law as envisioned by Rousseau when a single party wins in a two-party system. In a proportional voting system, a single party rarely wins a majority of seats in an election.  Therefore, these systems require an additional step to create a governing majority.  Parties must form coalitions after the election to achieve a majority that can enact laws.  Critics claim this leads to legislative gridlock or undue influence by minor parties needed to reach a majority of seats.  There are certainly examples where this happens, particularly in countries new to democracy or with districts that have a large magnitude, encouraging a multitude of parties.  However, in mature democracies with a proportional system, parties typically form majorities with ease.  Typically, any coalition government includes one of the two larger, centrist parties, which provides stability and continuity.

More importantly, these governments function at a high level.  One measure is reflected in the Democracy Index published every year by The Economist.  This analysis considers accountability to the electorate between elections, checks and balances, transparency and openness.  It is notable that this Index also prioritizes the functioning of the legislative branch.  As the branch that makes law, it plays a central role in any representative democracy.  Nearly all of the highest performing democracies are proportional ones.  Since parties require coalition building to reach a majority in a proportional system, voters expect compromise and collaboration.  Seldom are these democracies marked by campaigns declaring the election as “the most important in history.”  Voters do not expect an election to result in the final vanquishment of their odious enemy.  The election simply marks another opportunity to see more favored parties gain a stronger hand in negotiating with coalition partners to form a government.  This approach allows for the implementation of policies that require long-term political support in government.  It is no surprise that the Democracy Index now places the U.S. in the category of “Flawed Democracy” mostly due to low marks for functioning of government and political culture.

Shifting Coalitions

Similar to the formation of majorities, critics of proportional systems point to the stasis of coalitions from one election cycle to the next.  Rousseau cited the dominance of one faction as anathema to the general will.  Democratic societies require a dynamism that sees viewpoints and interests rise and fall as the needs of society change.  With the need to form coalitions, do proportional systems find it difficult to respond nimbly to voters as reflected in election outcomes?  Some scholars have noted that the third largest party in a proportional system often control government by acting as the perennial key to majority rule.  Additionally, a party that performs weakly can retain seats over multiple cycles because it is easier to exceed a quota than reach a majority of votes.  Also, it is more difficult to target any particular party when campaigning in a multi-party system.

Despite the lack of dramatic swings and life or death campaigns, proportional systems do produce new coalitions from cycle to cycle.  Note the recent rise of Sinn Fein in Ireland, which has shifted government from center-right to center-left in recent years.  Since parties can shed constituencies and still win seats, they can afford to be more coherent and consistent on policy.  For example, New Zealand has experienced the rise of new parties such as United Future, which advocates socially conservative but economically centrist policies.  ACT New Zealand advances a libertarian agenda, which is socially liberal and fiscally conservative.  The success of these parties signals to a coalition government which policies require attention between elections.  In the U.S., the Republican Party and Democratic Party both express, uncomfortably, conflicting philosophies in and out of office.  Needing to hold together broad, unstable constituencies, our major parties find it difficult to make tough policy choices that could alienate a faction of the party.  Proportional voting provides spaces for parties to compete on policy, knowing they will still have a role in government.  Consequently, coalition governments will shift over time in response to the clarity resulting from an election cycle.

Equality

The proportional voting system was devised specifically to address concerns over the lack of equality associated with majority systems.  Political theorists recognized majority systems treat voters unequally in two ways:  first, they can permanently lock minority groups and minority perspectives out of any representation; and second, they can exacerbate the number of wasted votes so that some voices count more than others.  Proportional voting cures these ills.  John Stuart Mills wrote that without proportionality, “there is not equal government, but a government of inequality and privilege; one part of the people over the rest.”  Proportional voting remedies this outcome by ensuring minority perspectives are represented equally in proportion to their support.  For those jurisdictions where one party enjoys a sustained majority, a proportional system ensures minority views also have a voice.  While minority parties may not have the ability to drive legislation, they can participate in the legislative process and influence it.

More importantly, proportional voting obviates the scourge of wasted votes.  As noted in the last essay, gerrymandering captures the worst element of wasted votes where one party manipulates district boundaries to dilute the number of seats by the opposing party.  It concentrates supporters of one party in a few districts so that another party can win many more districts by closer margins.  Consequently, gerrymandering artificially inflates the power of the party drawing districts.  Proportional voting dramatically reduces the chance of gerrymanders through multi-member districts.  Such districts help ensure the percentage of votes received by a party or candidate corresponds to the seats won.  As noted, the most extreme examples are the Netherlands and Israel, which both have one national district so there is no mechanism available to manipulate districts.  Regardless, even districts with four or five seats have far fewer wasted votes than majority systems.

Choice

While proportional voting scores well for participation and equality, choice offers its greatest attribute.  The last essay noted that majority systems undermine choice when more than two choices appear on the ballot in several ways;  first, they can cause outcomes inconsistent with voter preferences, second, they can suppress opportunities for minority and female candidates; and third, they can distort voter preferences through strategic voting.  Proportional voting systems avoid these problems.  As noted, proportional voting lists multiple candidates for multiple seats on the ballot.  Therefore, this system does not have to manufacture a way to whittle multiple candidates to two before selecting a winner.  Proportional voting does not completely escape the need to employ formulas as noted with leftover votes, but those mechanisms are better equipped to translate voter preferences to outcomes than those associated with majority systems.  It is no surprise that Condorcet, who exposed the flaws of majority systems, devised a voting system with proportional aspects.

Representation by Minorities and Women

As noted in the last essay, the simple majority voting system tends to diminish the representation by women and minorities in legislative bodies.  For this reason, it unnecessarily limits choice.  This phenomenon – known as the “most broadly acceptable candidate” theory – occurs in a two-party system that relies on homogeneity to appeal to disparate coalitions for success.  Proportional voting obviates this tendency by encouraging parties to advance lists with a diverse array of candidates.  Not needing a majority to win seats, parties can appeal to a range of constituent groups.  The Electoral System Design Handbook cites ample research showing that racial and ethnic minorities as well as women win many more seats in a proportional system as compared to a winner-take-all voting system.  14 of the top 20 nations at electing women are List PR systems.

Freedom of Choice

Lastly, the proportional system affords latitude to voters in choosing a candidate or party.  By doing so, it has a significant psychological effect in the way they behave within the political system.  Importantly, proportional voting allows voters to express their preference rather than to vote strategically.  As described in the previous essay, strategic voting can contribute to a negative feedback loop of polarization as voters are increasingly motivated by hatred of an opposing party rather than support for a preferred choice.  With proportional voting, there is no penalty for voting for a candidate who cannot win a majority of votes.  Candidates merely have to achieve quota.  Therefore, proportional voting encourages support for a variety of parties, which depolarizes the electorate.  Since proportional voting encourages the free expression of preferences, it reveals the diversity of choices expressed by voters, unlocking the collective brain.  While proportional systems often result in coalition governments, those coalitions can more accurately represent the will of the people, leading to decisions over public goods reflective of the diverse interests of the populace.

The Promise of Proportional Voting

Proportional voting arose out of a conscious search for a system that addressed certain defects of majority voting.  Political theorists could see that majority systems unduly squelched minority perspectives from participating in government.  They also understood how majority voting encouraged the manipulation of districts to entrench certain interests.  In response, theorists envisioned a system that gave a voice to minority interests.  However, it took time to devise a voting system that generated winners in proportion to their share of the vote.  More importantly, it took unique political circumstances at the dawn of the 20^th century to permit reforms that diminished the power of the major parties and allowed others a seat at the table.  The experience in New Zealand shows that public pressure can lead to reform when electoral outcomes expose the flaws of majority voting.  Perhaps the greatest attribute for proportional voting relates to the characteristic identified by Duverger:  it “depolarizes” voters by encouraging multiple parties.  At a time when polarization poses an existential threat to democracy, such an attribute has tremendous value.

Mack Paul is a member of the state advisory board of Common Cause NC and a founding partner of Morningstar Law Group.

Parts in this series:

Introduction: Building Democracy 2.0

Part 1: What Is Democracy and Why Is It Important?

Part 2: How the Idea of Freedom Makes the First Innovation Possible

Part 3: The Second Innovation that Gave Rise to Modern Democracy

Part 4: The Rise and Function of Political Parties – Setting the Record Straight

Part 5: How Political Parties Turned Conflict into a Productive Force

Part 6: Parties and the Challenge of Voter Engagement

Part 7: The Progressive Movement and the Decline of Parties in America

Part 8: Rousseau and ‘the Will of the People’

Part 9: The Dark Secret of Majority Voting

Part 10: The Promise of Proportional Voting

Part 11: Majorities, Minorities and Innovation in Electoral Design

Part 12: The Misdirected Attempts at Electoral Reform in the U.S.

Part 13: Building Democracy 2.0: The Uses and Abuses of Redistricting in American Democracy