With the 2016 general elections right around the corner, the island nation of Taiwan will once again exercise its hard-won democracy in a contest heralding exciting political change. On January 16, each citizen of voting age (20 years or older) with a valid household registration in the “Free Area of the Republic of China (ROC)” will be presented with three paper ballots:

  • a presidential ballot to decide who will become the next ROC president, taking office on May 20
  • two separate ballots to determine the makeup of the Legislative Yuan, the unicameral legislature that will be seated on February 1

In this primer, we take a closer look at the Taiwanese electoral system, including how the multiple legislative ballots are counted and how seats are assigned to lawmakers.

Composition of the Legislature

Reforms enacted in 2005 set the size of the Legislative Yuan at 113 seats. Upon arriving in the capital of Taipei, all 113 lawmakers are equal in terms of their rights and responsibilities. (One will eventually be chosen as the speaker of the legislature.) However, in the process of gaining office, not all of them are elected using the same method.

There are three main ways to join the lawmaking body: 73 legislators take seats representing geographically-based single-member districts (SMD); six seats are reserved for aboriginal legislators; and 34 seats are filled by proportional representation (PR), in accordance with a nationwide party vote.

1. Geographical Constituencies: majoritarian single-member districts (SMD)Geographical constituencies in Taiwan are single-member electoral districts, akin to congressional districts in the United States. A single winner represents the entire district by receiving a plurality of the votes, i.e. the most votes, often referred to as a “first-past-the-post” contest.

Geographic constituencies make up 73 seats, or nearly two-thirds of the Legislative Yuan. Every major administrative division—including special municipalities (直轄市), counties (縣), and county-level cities (市)—is guaranteed representation by at least one electoral district. At present, there are 22 municipalities, counties, and county-level cities in Taiwan.

Map of the 22 administrative divisions of Taiwan. Even smaller counties on outlying islands with fewer than 100,000 residents command at least one electoral district, which skews representation in the Legislative Yuan. 

Most electoral districts are drawn to contain roughly 320,000 registered residents. (Note: Electoral districts are not to be confused with city districts, which are political boundaries for local administration.) Boundaries were originally decided in 2007, as part of the implementation of the current electoral system. With 73 electoral districts mapped across the nation, the democratic intent was for each district to represent a similar number of citizens in the legislature.

However, some areas have seen major population booms, while others have experienced outflows of residents, which is not reflected in current electoral districts. For example, Hsinchu County’s population as of November 2015 is about 0.54 million, which has surpassed both Nantou County (0.51 million) and Chiayi County (0.52 million)—yet it is represented by one lone legislator, while the latter two counties have two representatives each. As a result, Hsinchu County’s residents are comparatively underrepresented in the national legislature. Future redistricting efforts could alleviate these imbalances.

Numerous small districts are clustered in urban areas because of higher population density, whereas geographically large districts carved out on the less-populous east coast represent roughly the same number of people.Political scientists consider the first-past-the-post SMD system “majoritarian.” In an electoral district, a candidate only needs to receive the bare majority of the vote (or simply a plurality when multiple candidates are running) to win office. This “winner-take-all” model, utilized in Great Britain and the United States, readily forges legislative majorities in parliament.

However, a substantial number of voters may not feel represented by the winning candidate, who becomes the legislator for the whole district. They did not vote for the candidate, and their voices appear to be discounted. Taiwan corrects for this problem with another set of “proportional” seats (see below), engineered to balance out the majoritarian features of the SMD-plurality system with another mode of representation.

2. Aboriginal Seats: representation for minorities

Aboriginal communities in Taiwan, numbering about 550,000 persons, have distinct status and thus hold a number of “reserved” legislative seats. The six at-large seats for aborigines are selected via a multi-member district electoral system. In effect, there are two aboriginal “districts,” each encompassing the nation: three seats representing the “Plains aborigines” (平地原住民) and three seats representing the “Mountain aborigines” (山地原住民). Multiple candidates from the same political party may run, but every voter marks only one preferred candidate on the ballot. The top three vote-getters in the Plains and Mountain categories all win seats.

The classification of whether an aboriginal person is a Plains aborigine or a Mountain aborigine (sometimes referred to as “Lowland” and “Highland” aborigines) is somewhat arbitrary. It does not take into account ethnic or tribal affiliation, and is instead pegged to where one’s family household registration was located (i.e. in the “plains” or “mountain” administrative regions) prior to 1945, the year the ROC occupied and took control of Taiwan from the Japanese.

Note: The official definition of “Plains aborigine” (平地原住民) as described in the Status of Indigenous People’s Act must be distinguished from the common phrase of “plains tribes” (平埔族 or 平埔原住民), used in English to describe the indigenous communities historically living in the western lowlands of Taiwan. To a great extent, members of the “plains tribes” (and their present-day descendants) were Sinicized and did not register as aborigines. They are not recognized by the ROC government as one of Taiwan’s 16 official aboriginal groups today, though a movement exists in some quarters to reclaim an aboriginal identity.

Persons with aboriginal status vote for aboriginal candidates instead of voting for candidates in a particular geographical single-member district. Because the aboriginal population is spread out, and much smaller relative to the Han population, it is feared their votes would be diluted in any one geographical district—hence the mechanism of “reserved seats” to ensure indigenous representation.

Under this system, Taiwan’s aborigines are indeed guaranteed decent representation in the country’s legislature. Though comprising only 2.34% of the total population (Taiwan’s population is about 23.5 million), the aboriginal communities’ six reserved seats form 5.31% of the legislature. With two additional aboriginal legislators likely to be elected in the separate PR party-list vote (see below), indigenous representation will potentially grow to 7.08% in the next session of the Legislative Yuan. Hopefully, this empowers aboriginal legislators to promote laws and policies supported by their community, such as statutes regarding autonomy.

Two other issues arise from this system: given the nationwide scope of the aboriginal seats, candidates are typically chosen from the more populous tribes. Smaller tribes are often not represented at all. Second, voting still takes place at local polling stations around the country, so in places with very few aboriginal residents, the secrecy of the vote may be compromised when the “aboriginal vote” for that station is tallied and reported.

3. Proportional Representation: choosing seats through party lists

In addition to voting for either a single-member district or aboriginal candidate, every voter receives a second legislative ballot featuring a list of political parties. This ballot, commonly referred to as the party ballot (政黨票), determines the number of seats awarded to political parties in a third pool of seats for legislators-at-large.

These 34 seats are assigned based on the nationwide proportion of votes the parties obtain, making it a “proportional representation” or PR system. Electoral districts do not matter here—only popular support across the country. The intent of PR is to make the legislature more representative, counteracting the “majoritarian” tendencies of the SMD seats. Though a voter may not feel represented by her local legislator because she voted for an opponent, that voter’s voice is “heard” at the national level when she casts a ballot for a political party she does support.

Many other countries also use PR methods to fill part, or all, of their legislatures, including Sweden, Germany, Israel and Uruguay, etc.

(A) Party Lists & Electoral Thresholds

Once seats have been proportionally allocated to the political parties, each party then fills them according to its party list. Each party determines its list of candidates internally, making this a closed-list PR system.

Candidates are listed in order of priority. For example, if a party wins enough votes to be awarded two PR seats, then the #1 and #2 candidates on the list would become legislators, while candidate #3 or anyone ranked below would not gain office.

With 34 PR seats, theoretically any party winning more than 2.94% of the national party votes would be entitled to one legislative seat (100% / 34 seats = 2.94% per seat). However, Taiwan only awards seats to parties that win at least 5% of votes. Smaller fringe groups cannot grab seats in parliament unless they truly command support from a substantial part of the populace.

As a consequence, voters who choose parties that do not pass the threshold are not represented in this pool of seats. If two smaller parties receive only 3% votes each, they would not meet the minimum threshold and would receive no PR seats. Meanwhile, the parties that do cross the threshold, accounting for 94% of total ballots cast, would split the 34 seats according to their share of these remaining votes.

A sample PR ballot from 2008 showing the list of parties. The number associated with each party was randomly selected. (Source: Central Election Commission)

(B) Hypothetical Example for 2016

How the process works in accordance with the procedures laid out in the Civil Servant Election and Recall Act is illustrated by the following chart. The “% valid votes” in this hypothetical matchup are actual projections for the 2016 general election drawn from a popular election prediction website, rather than a past election.

The PR electoral threshold prevents too much fragmentation of the legislature and limits spoilers from entering the political system. Internationally, Taiwan’s 5% threshold is in the middle of the pack: the Netherlands has no formal threshold (but it is effectively .66%, the minimum vote required to win one of 150 legislative seats), whereas Turkey has an illiberal 10% threshold that regularly disenfranchises many voters who support legitimate and sizable minor parties.

Since PR lawmakers do not represent a particular electoral district, they are not tied to any one locality or population. As a result, their motivations and electoral incentives are distinct from those of SMD legislators. Loud or controversial candidates may seek a national reputation to secure a higher ranking on the party list and improve their chances for election. They are also less likely to engage in retail politics and connect with the residents of a city or county, instead placing their exclusive focus on national issues. By the same token, the fate of PR lawmakers is more closely tied to the party’s overall popularity, so they do not have local goodwill to rely on in times of political difficulty.

 (C) Forming a Party List: female representation required

The creation of the party list and the ordering of candidates within that list is a political game unto itself. The following chart shows the lists of candidates submitted by the four parties likely to win PR seats in 2016. Each cell indicates the gender of the legislative candidate (red = female, blue = male) as well as the candidate’s qualifications to hold office. The order of the candidates indicates, to some degree, the esteem of that politician in the party—candidates higher on the list are more likely to be elected, because seats are filled starting from the top—as well as the relative importance of the issues they specialize in.

Some of the qualifications include labor, women’s rights, environmental issues, youth advocacy, human rights, and education. Some members of the PR lists even represent the interests of Overseas Chinese communities. Before the 2005 electoral system reform, a number of seats in the legislature were specifically set aside to represent the overseas community, as inscribed in the ROC Constitution. Those interests are now accounted for as part of the broader PR pool of seats.

While parties may submit lists of up to 34 candidates, only the candidates with a meaningful chance to enter the Legislative Yuan next year are shown in the chart.

It is no accident that female candidates are well represented in these lists. Article 134 of the ROC Constitution mandates female political representation, though it does not specify a quota. Accordingly, the architects of the current electoral system required that female legislators fill at least half of the PR seats awarded to any party. If a party were to win only one PR seat, that single legislator would have to be a woman.Using the numbers from the projected scenario above (where the DPP wins 17 of the PR seats, KMT: 12 seats, PFP: 3 seats, NPP: 2 seats), and then inspecting candidates’ gender on the party lists, we see that for the KMT, PFP and NPP, women lawmakers represent at least half the winning candidates for each party. In the case of the DPP, only 8 women are elected to fill the DPP’s 17 allocated slots—slightly under half. In this case, the final male candidate on the list (#17) is bumped in favor of the next female candidate (#19) to achieve the 50% quota for women.

The 34 PR seats are meant to counterbalance the more majoritarian tendencies of the 73 SMD seats, but there are questions as to whether this does enough, or if increasing the number of PR legislators would allow for more proportional outcomes in the Legislative Yuan overall.

An Evolving Exercise in Democracy

The election in 2016 marks the third time the Taiwanese people will vote under this legislative regime. With multiple types of legislative contests paralleling the presidential campaign, and three separate ballots issued to each voter, the electoral system in Taiwan is no simple affair. Yet it reflects the need for more sophisticated approaches to democracy, particularly in multiethnic societies seeking to balance greater inclusion with satisfactory representation and effective governance. The system is by no means perfect, and suggestions for political and electoral reform continue to abound. Regardless, with another free and fair election to put under the belt, and the expectation that this upcoming election will precipitate historic change and renewal, whatever the precise final apportionment of PR votes, aboriginal seats, and SMD constituencies next month, Taiwan’s people and their modern interpretation of democracy will be the victors.


About the Authors

Kevin Fan Hsu is Lecturer in Urban Studies at Stanford University, where he co-founded the Human Cities Initiative (humancities.stanford.edu). His studio, Skyship Design, develops educational experiences with a social mission, including the online courses “Democratic Development” (with Larry Diamond) and “International Women’s Health & Human Rights” (with Anne Firth Murray). Find out more at skyshipdesign.net.

Lucien Wei Hickman grew up in Taoyuan, Taiwan and now studies geography in Washington state. When not exploring the intricacies of paper and digital maps, he’s outside navigating landscapes and covering new terrain in person. His work can be found at Lucien.Earth (www.lucien.earth).

Links to PDF Files

Legislature Overview


PR Procedure: Seats to Parties


PR Party Lists for 2016


(Feature photo of Doris Yeh, whose husband Freddy Lim is running for legislator in Taipei, from New Power Party press release.)


Kevin Hsu

Kevin Fan Hsu is Lecturer in Urban Studies at Stanford University and co-founder of the Human Cities Initiative. He crafts open online courses and designs other educational experiences with a social mission at Skyship Design (www.skyshipdesign.net)