A survey of 1,213 Taiwanese conducted in December 2022 found that the Democratic Progressive Party’s William Lai has a narrow lead in the upcoming presidential election.
In January, Taiwan will have its closest presidential election since 2008. With the failure of Taiwan’s opposition parties to coordinate on a single presidential candidate, the Democratic Progressive Party’s William Lai has maintained a narrow lead across most polls, while considerably underperforming compared to that of outgoing incumbent Tsai Ing-wen. I ask how this election would look differently under different electoral rules, namely variations of a rank order system, and to what extent supporters for the Kuomintang’s Hou You-ih or Taiwan People’s Party’s Ko Wen-je would choose the other as their second choice.
Many presidential systems use a runoff system, especially in Latin America, with the proposed benefits typically focusing on generating majority support and limiting spoiler effects and strategic voting, while encouraging remaining candidates to appeal beyond their traditional partisan bases. In contrast, Taiwan’s lack of a runoff allows for the situation where a majority of voters would have preferred a different winner, as seen most clearly when the Democratic Progressive Party’s Chen Shui-bian won the 2000 presidential election with only 39.3% of the vote due to the splitting of the “pan-blue” vote by Lien Chan and James Soong. Before 2000, the Kuomintang even opposed a runoff system believing it unnecessary to secure electoral victory.
In contrast, few presidential systems use a rank order system (e.g. Kiribati), although legislative and local offices frequently use variations. In Australia’s alternative vote system, if no candidate wins an outright majority in the first round, the worst performing candidate is dropped and their second-preference votes are reallocated for remaining candidates, and repeated if necessary until a candidate receives a majority vote. Under the Borda count system, similar to that used by Major League Baseball to determine the Most Valuable Player, candidates receive points based on their preference rankings with the highest point total winning.
A cursory glance at existing polls shows both systems would presumably benefit opposition candidates currently and lead to a Lai loss. Yet no existing survey asks Taiwanese to rank order the presidential candidates in order to measure the extent of preferences.
I surveyed 1,213 Taiwanese via web survey December 1-11, conducted by Macromill Embrain, using quota sampling for gender, age, and geographic region.
I first asked respondents who they planned to vote for in 2024. Here Lai’s support (28.77%) edges out Ko and Hou (26.05%, 21.68%). Yet, nearly a quarter of respondents (23.5%) stated they did not plan to vote, likely capturing also those who have yet to decide. Admittedly, a web survey may end up over-representing younger respondents, precisely those that the Ko campaign most strongly appeals, while failing to pick up shifts in opposition support towards Hou since the collapse of coordination talks.
I next asked respondents to rank the candidates with their most preferred candidate first and least preferred third. Overall, first choice preferences show Lai with a small lead over Ko (38.99% vs. 34.21%), with Hou trailing by over ten percent (26.79%). Among second preference choices, both opposition candidates favor better with nearly identical rates (Hou: 39.24%, Ko: 38.58%), while Lai’s third preference ranking nearly matching his first preference support clearly indicates opposition preferences.
As expected, those who first preferred Lai showed little preference over the opposition candidates, with 50.11% choosing Hou over Ko. In contrast, 71.38% of those choosing Hou as the first preference ranked Ko second. However, Ko supporters showed a much lower likelihood of choosing Hou as their second choice (57.59%), consistent perhaps with Ko’s appeals to younger voters, appealing to voters with less entrenched political views, and Ko’s own comments about the KMT. The pattern also suggests that if Ko had dropped out earlier, the net benefit to Hou’s campaign may be less pronounced as anticipated.
The survey also asked about vote choice in 2020. Unsurprisingly, over two-thirds of those who claimed to have voted in 2020 for the DPP’s Tsai Ing-wen (65.84%) or KMT’s Han Kuo-Yu (70.59%) stated they would stay with the same party vote in 2024. Meanwhile, a thinner majority (57.5%) of those who supported James Soong in 2020 stated their first preference for Ko. Moreover, pluralities if not outright majorities of those who did not vote in 2020 or were ineligible (e.g. not old not to vote) stated their first preference as Ko.
Returning to 2024 preferences, using the alternative vote would lead to Hou’s third place finish and the dropping of his candidacy, reallocating votes to second preferences, which would result in a Ko victory (53.33%) in the second round. If the alternative vote would generate a majority, the Borda Count would produce a winner with an even narrower margin than the current system. Tallying up preference votes (3 points for first place down to 1 point for third place) would result in a nearly evenly split vote across candidates, with Ko eking out a victory with 34.5% of the points (Lai: 33.36%, Hou: 32.14%). Both alternatives show the extent in which Lai benefits from the split opposition, but also that rank ordering options can produce unintended outcomes as well. While the focus in runoff and alternative vote systems is to produce a winner with clear majority support, the Borda count system provides no such guarantee and in this example may exacerbate the problem it sets out to avoid.
READ NEXT: President Yoon’s Dismal Approval Ratings Create Challenges Going Forward
TNL Editor: Kim Chan (@thenewslensintl)
If you enjoyed this article and want to receive more story updates in your news feed, please be sure to follow our Facebook.