After the Nov. 24 series of referendums, the international media reported that Taiwan’s push for same-sex marriage referendum had been defeated. Although this is disappointing, there was another piece of news that was not picked up as widely: a majority of Taiwanese voted in favor of providing legal protections to same-sex relationships, albeit not as “marriage.”

You see, in the referendum held two Saturdays ago, there were actually three referendum questions on same-sex marriage. They were:

  • Question 10: Do you agree that Civil Code regulations (constitution) should restrict marriage to being between a man and a woman?
  • Question 12: Do you agree to types of unions, other than those stated in the marriage regulations in the Civil Code, to protect the rights of same-sex couples who live together permanently?
  • Question 14: Do you agree that the Civil Code marriage regulations should be used to guarantee the rights of same-sex couples to get married?

Questions 10 and 12 came from anti-marriage equality groups, while question 14 was proposed by pro-equality advocates.

The anti-equality groups had collected more than 677,972 signatures for their questions to be put on the ballot for the referendum, though only 565,676 (83 percent) were considered valid.

For the pro-equality groups, they managed to collect 330,000 signatures within the first 37 days after they started obtaining signatures, which they said was a record in itself in Taiwan, done in twice the speed as the anti-equality groups. Eventually, they collected 474,777 signatures, of which 432,329 (91 percent) were considered valid. Their proposal was also approved at a later stage, which they said tipped the level-playing field in favor of the anti-equality groups, as they were allowed to begin campaigning earlier.


Credit: AP / TPG

Participants attend the 2017 Taipei Pride parade on Oct. 28, 2017.

In Taiwan, referendum proposals need to collect signatures from 1.5 percent of the electorate (or 281,745 for the last referendum) for them to be put to a vote.

The anti-equality groups had campaigned on asking voters to vote yes to their questions (10 and 12) while the pro-equality groups had campaigned for voters to vote no to the anti-equality questions and voting yes to their question (14).

By now, it is known around the world that the pro-equality had their referendum defeated. Only 32.74 percent of the voters supported amending the constitution to legalize same-sex marriage.

Globally, attention had been paid to the question by the pro-equality groups. But the equally important question by the anti-equality groups was largely neglected. No doubt the intentions of the anti-equality groups were questionable, but nonetheless, on question 12, there were actually 61.12 percent in support of legalizing same-sex relationships, albeit under a separate law.

This is actually a win. It is not the win that same-sex marriage advocates were hoping for, but it is a win nonetheless for Taiwan, and for Asia.

The question on same-sex marriage was moot in the first place. The support was already there.

Coming from Asia, and from Singapore where sex between two men is still illegal, seeing that 61.12 percent of Taiwanese voters choose to to legalize same-sex relationships in some form was something to shout for.

Perhaps it was easier as a foreigner living in Taiwan watching in to look at it differently, or perhaps I saw the cup as half-full. Or perhaps, coming from a country where same-sex rights are still infringed upon, this would be something to celebrate for us.

But there is also another interesting point – there were 32.74 percent of the Taiwanese who voted in support of codifying same-sex marriage into the constitution while 61.12 percent voted in favor of enacting a separate law to legalize same-sex relationships.

Do these voters overlap? I asked some Taiwanese what they thought.

“Of course they might overlap,” a person ventured, “but I believe a huge proportion do not overlap.”

“By right, they should not overlap,” he added.

Another person said: “I don’t think a rational person would do that [vote yes to both questions].”

It is impossible, of course, to prove the amount of overlap between “yes” votes on the various questions. Many voters went to the referendums with an incomplete understanding of what was a complex and confusing process. However, most people I spoke to felt that most voters would not have voted “yes” to both questions, because for people who were in support of amending the constitution to enable full marriage equality for same-sex couples, they are unlikely to be willing to settle for enacting a separate law – this has been a debate that has been going on over the last few years in Taiwan.

Conversely, for those who gave their support to enacting a separate law, it is seen that they would not be likely to vote in favor of amending the constitution.

Of course, there could be stray votes. A friend of a friend said that he voted “no” to all three questions on same-sex marriage because he thought that questions of human rights should not be put to a vote. Maybe this was illogical, or perhaps he voted otherwise, but this was his spoken standpoint.

There were actually 10 questions that were put to a vote on this referendum, and analysis by Professor Austin Wang from the Department of Political Science at the University of Nevada Las Vegas showed that there is a strong positive correlation between the referendum questions 14 and 15. Question 15 is the other question by the pro-equality groups on ensuring gender equality education in schools: “Do you agree that gender equity education as defined in the Gender Equity Education Act should be taught at all stages of the national curriculum and that such education should cover courses on emotional education, sex education and gay and lesbian education?” It garnered 34.01 percent of the votes.

There is also a strong positive correlation between questions 10 and 11 and a strong positive correlation between questions 11 and 12, and a relatively strong positive correlation between questions 10 and 12.

Question 11 is the third question by the anti-equality groups, on disallowing homosexual-related education in schools: “Do you agree that the Ministry of Education and individual schools should not teach homosexual-related education, as detailed under the Enforcement Rules for the Gender Equity Education Act, in elementary and middle level schools?” It got 67.44 percent of the votes.

Could the strong correlation between the questions by the pro-equality groups suggest that the voters who voted “yes” to the question would be similar? Could the same suggestion apply for the voters who voted “yes” to the anti-LGBT questions as well?

Similarly, there is a strong negative correlation between those who voted “yes” to questions 10 and 11 by the anti-equality groups with those who voted “yes” to questions 14 and 15 by the pro-equality groups. Would this suggest that the voters who voted “yes” to the two sets of questions are largely distinct voters? We cannot make any firm conclusions by way of correlation, and we cannot be sure that they would be distinct voters unless we could have conducted an exit-poll or a separate survey, but this is certainly food for thought.

In any case, there are people, and likely a majority, who voted in support of codifying same-sex marriage into the constitution who did not vote on the question on enacting a separate law. If so, this would mean that the proportion of voters in support of legalizing same-sex marriage in some form would undoubtedly be higher than the 61.12 percent who voted “yes” for question 12 (on a separate law) and would be somewhere between 61.12 percent and around 90 percent.

It is no mean feat, if you ask me. The Taiwanese did not vote against legalizing same-sex relationships. The debate was over the form, in spite of the unfairness that admittedly still exists.

This is not the win that same-sex marriage advocates were hoping for, but it is a win nonetheless for Taiwan, and for Asia.

This may seem hard to believe at face value, but the referendums, after all, cannot be isolated. They must be seen as part of the historical polls that have been conducted with the Taiwanese on this issue in the past, and the polls in Taiwan have shown themselves to be quite accurate on assessing voters’ attitudes, when you compare with how the local elections have panned out.

In a survey in 2012, 52.5 percent of Taiwanese had agreed that “homosexuals should have the right to marry.”

In 2013, a higher 53.7 percent of Taiwanese had agreed that same-sex couples should be allowed to marry.

In 2014, the support increased to 67.5 percent who agree that same-sex marriage should be legalized.

And in 2015, in a survey of more than 310,000 Taiwanese conducted by the Ministry of Justice, the support increased further to 71 percent of Taiwanese who believe in the right of same-sex couples to marry.

As such, it is clear that support for same-sex marriage has been increasing year-on-year.

In fact, in 2016, in a poll by the Kuomintang (KMT), there was still also a majority who supported legalizing same-sex marriage – 51.7 percent who agreed as compared to only 43.3 percent who disagreed.

Even in a survey conducted by a conservative-leaning party like the KMT, support for the legalization for same-sex relationships was still high!

Therefore, the question on same-sex marriage was moot in the first place. The support was already there.


Credit: AP / TPG

Costumed participants pose at the 2018 Taipei pride parade on Oct. 27, 2018.

This is remarkable in Asia, where India decided just three months ago to decriminalize sex between two men and harsh laws against same-sex relations still exist in countries like Singapore and Malaysia.

A survey in September showed that 55 percent of Singaporeans supported the law that criminalizes sex between two men (with only 12 percent opposed), and it was only four days ago where another survey showed that 42 percent of Singaporeans agreed that the law should be retained even if it is not enforced (with only 19 percent disagreeing), even as 51 percent did not think that religion should influence Singapore’s laws.

In Indonesia, the city of Pariaman has passed a by-law to ban LGBT activities on the basis of public disorder, whilst imposing a fine of up to 1 million rupiah (NT$2,144).

Perhaps the only other bright spot is Thailand, where the government is expected to introduce same-sex civil partnerships early next year.

Meanwhile, even though Hong Kong’s courts have ruled that overseas same-sex partnerships have to be recognized, its legislature has rejected a motion to legalize same-sex unions for locals.

Breaking it down by region

Another interesting question on Taiwan’s referendum came on question 12, proposed by anti-equality groups, on enacting with a separate law to legalize same-sex relationships. Even though 72.48 percent of the voters voted “yes” to restricting the Civil Code to heterosexual relationships under question 10, a lesser 61.12 percent were in favor of enacting a separate law to legalize same-sex relationships under question 12.

The 27.52 percent who voted “no” to restricting the Civil Code to discriminate same-sex couples would indefinitely have voted in favor of marriage equality. This would leave about 11 percent of the voters who voted to restrict the Civil Code to only heterosexual relationships and who did not also want to legalize same-sex relationships in anyway, if we could combine the percentages to both questions.

Professor Wang opined: “Many church leaders only support Q10 and Q11 but oppose Q12.”

What is clear, though, is that there is an obvious pattern in how voters voted.

A group of netizens compiled the results from the referendum and mapped them according to the voting districts.

On question 10 on restricting the Civil Code to heterosexual marriages, Taipei and Hsinchu cities stand out as having lower-than-average “yes” votes – there were fewer voters who discriminated against same-sex relationships. The same pattern can be seen for the support for question 12 (to enact a separate law) and question 14 (to enshrine same-sex marriage in the constitution).


In Hsinchu City, the lowest support for the anti-equality questions came in East Hsinchu City, as did the highest support for the pro-equality questions. On question 10, on retaining the Civil Code for heterosexual marriages, there was a higher 71 percent support in North Hsinchu City and 72 per cent support in Hsinchu’s Siangshan township, while there was a lower 67 percent in East Hsinchu City.

On question 14 on amending the Civil Code to legalize same-sex marriages, 36 percent of the voters in East Hsinchu City voted in support of full same-sex marriage equality, while there was a lower 33 percent in North Hsinchu City and 32 percent in Siangshan township.

An analysis by CM Media also showed that the Taipei has the highest support for marriage equality, at 38.13 percent of the votes.

On question 12, Yilan and Yunlin counties also stand out as having lower support for enacting a separate law.


When you look closer at Yilan County, the same Datong and Nan-ao townships where more people voted in favor of restricting the constitution were also more in favor of enacting a separate law (darker pink areas), and less in favor of codifying same-sex marriage into the Civil Code (lighter yellow areas).

Figure: Voting pattern for Question 10 in Yilan County. Source:
Figure: Voting pattern for Question 12 in Yilan County. Source:
Figure: Voting pattern for Question 14 in Yilan County. Source:

We can see a somewhat similar pattern for other cities like New Taipei and Taoyuan and counties like Hsinchu, Hualien and Taitung, but the percentages would need to be broken down even further and correlated for a better analysis.

If, indeed, the voters seemed to have aligned themselves quite clearly in voting “yes” for either the questions by the anti-equality groups, or for the pro-equality groups, then it does lend weight to the hypothesis that only about one-tenth of voters in Taiwan are vehemently against legalizing same-sex marriage. This would make those who are in support of some form of legalization of same-sex relationships in the high majority in Taiwan.

If so, Taiwanese are actually not the conservatives we make them out to be.

Again, we cannot make any firm conclusions, but the different variations of the questions on same-sex marriage and their alignment to the pro-/anti- groups, and how voters therefore voted, do give us something to think about.

While it is disappointing that whereas Taiwan could have achieved marriage equality, that the referendum has put a spanner in the works, Taiwan has still achieved something that no other Asian country has to date.

How the referendums were gamed

Taiwan should have achieved equality.

The Constitutional Court ruled in May last year that same-sex marriage has to be legalized either by amending the Civil Code or enacting a separate law. It said that if that this was not done within two years, that same-sex marriage would automatically be legalized.

After the referendum, Taiwan’s Judicial Yuan has also come out and said that the ruling cannot be overridden because the constitutional ruling is the most respected in Taiwan’s legal hierarchy.

After the referendum, Taiwan’s Premier William Lai (賴清德) also said that a separate law to legalize same-sex marriage will be drafted within the next three months. Pro-equality advocates consider this discrimination, as same-sex marriage will still not be codified into the constitution, and would make LGBT people second-class citizens.

So the battle is not over.

In fact, the anti-equality groups have even now said that they want the legislation to be about registering same-sex couples even as “family members.”

But this can be considered a departure from their referendum question which did not mention relegating the relationships of same-sex couples to just being legalized as “family members”.

Even so, there is the idea that those who support question 12 on coming out with a separate law might not actually support the legalization of same-sex marriage.

“The voters could have voted yes because the Constitutional Court has already ruled that same-sex marriage have to be legalized in some form, and people therefore voted in accordance to that,” a woman I spoke to suggested.

“Yes, but they still voted in support of legalization. They could have chosen to vote no, like about 10 percent did,” I replied.

“They might just be showing the least respect [that they think they should give on the issue],” another person suggested.

Another person told me: “They might want to discriminate against gay people so they still agreed only to having a separate law.”

This is true in both cases. Nonetheless it is still an agreement that same-sex relationships should be legalized.

But it is most likely that voters simply didn’t care. From the conversations I have had with people on this issue, the majority agree on this point. Let gay people live and let live: This seems to be the attitude among the Taiwanese. I have walked on the streets holding hands, and I have seen gay and lesbian couples holding hands all the time, and this has never been an issue. I could never do this in Singapore without feeling out of place.

If so, the referendum results might be a reflection of how voters were simply going with the flow, and following the loudest voice, without thinking through their votes. This in itself would be problematic. But this would also show how voters do not have a strong stance on the issue, and that their vote became largely symbolic for other reasons – which also goes to show how human rights should not have been put to a vote, if their approval would be swayed based on political sentiments. This was the point that marriage equality advocates in Taiwan tried to make when referendum questions by anti-equality groups were first approved.

In fact, Professor Wang showed as well that there is a correlation between the “yes” votes for the questions by the anti-LGBT groups and the gains by the KMT.

Here’s the other thing: The referendums were actually held alongside the local mid-term elections, and because they were held in conjunction, voters may have used the referendums as a more figurative referendum against the current governing party, the more progressive-leaning Democratic Progressive Party (DPP), many without questioning what the referendums were about.

If, according to the people I have spoken to, most people do not care about the issue of same-sex marriage, it may have been the case that voters were willing to use their votes in the referendum for same-sex marriage as protest votes against the current government, in expression of their unhappiness over its governance. This aligned with the KMT, which campaigned in opposition to full marriage equality.

Perhaps we can look at the support of the Taiwanese for both the DPP and KMT as an indication of voters’ attitudes towards same-sex marriage. As of September this year, support for the DPP stood at 24.7 percent while, for the KMT, it was 23.1 percent. 34.5 percent of voters did not support any party.

In fact, support for the DPP went from a high of above 50 percent in mid-2016, just after they won the presidential election, to less than half today. On the other hand, support for the KMT consistently hovered below 25 percent.

The KMT decided to use the referendums as a tactic to embarrass President Tsai’s government.

Therefore, the voters who no longer supported the DPP were up for grabs. From the analyses that have been written, it is clear that voters who therefore swung towards voting for the KMT did not actually support the KMT – instead, they were voting against the DPP. The consistently low support of less than 25 percent for the KMT said as much. As such, the correlation between the “yes” votes for the questions by the anti-equality groups and the gains by the KMT could be attributed to the protest votes against the DPP.

Of several people I spoke to, many voiced their concerns that voters did not even know what they were voting for at the referendum.

Indeed, it was not only the referendum questions on same-sex marriage that were used as a protest vote against the DPP.

The KMT had themselves put up three questions for the referendum, along with another pro-nuclear question put up by groups in alignment with them. Professor Wang also showed that the “yes” votes to these questions showed a high correlation with one another.

Professor Wang also did up a chart to show how the “yes” votes for the anti-LGBT referendums (questions 10, 11 and 12) were more aligned to the conservative KMT values in the top right quadrant in the chart below, and how the “yes” votes for the marriage equality referendums (questions 14 and 15) were more aligned to the progressive DPP values in the bottom left quadrant. My read of this is that, on the surface, we might assume that voters are innately conservative and therefore voted in alignment with the KMT. But if voters’ intentions were to choose not to vote for the marriage equality referendums as a protest vote against the DPP, then it might not suggest that they have become more conservative. It is only the “perception” that has been created that they are more conservative. As it is, surveys have shown that support for the KMT has never risen above 25 percent over the last two years.

Credit: Austin H. Wang

If anything, it showed that the KMT was successful in using referendums as a political tool to turn the voters against the current government.

The irony is that it was only this year that the DPP government had made it easier for the referendums to be conducted by drastically reducing the threshold to enable referendums to be initiated and passed, for example, by reducing the threshold for proposals to be passed from 5 percent of the electorate to 1.5 percent. They had perhaps not calculated that the KMT would be able to latch onto the relaxing of the requirements to manipulate the voters.

Remember, the referendums have not been proven to be a reliable indicator of public attitudes. We should recall that a 2015 survey by the Ministry of Justice showed that 71 percent of Taiwanese believe that same-sex couples should have the right to marry.

However, the KMT had also asked respondents to choose between amending the Civil Code (or constitution) or introducing a separate law. And to this question, 53.3 percent supported a separate law while 32.2 percent supported revising the Civil Code. But this is an about turn from the Ministry of Justice’s survey done just a year earlier, where a majority had supported revising the Civil Code.

By 2016, however, there was already widespread support for marriage equality. President Tsai Ing-wen (蔡英文) of the DPP had also campaigned on the platform of legalizing same-sex marriage and even garnered the second-highest margin of victory ever since democratic elections began to be held in 1996. Therefore, support was there for same-sex marriage. Or, at least, it was a non-issue for voters.

It is clear that the Taiwanese do believe that equal rights should be enshrined in some way.

But it was perhaps during this time that anti-LGBT campaigners also began to mobilize themselves and picked up steam. The KMT, being the more socially conservative of Taiwan’s two major parties, also decided to swoop in to divide the conversation and steer it in a more conservative direction.


Credit: Reuters / TPG

Some KMT candidates co-opted anti-marriage equality sentiments during their 2018 campaigns.

In effect, they forced the hands of the Taiwanese by making them choose between legalized marriages codified into the constitution or legalized unions via a separate law.

The KMT might have calculated that voters could be easily swayed on this issue, knowing that voters did not have a firm preference over the issue of same-sex marriage, and thus they knew that there was steam left on the conversation to manipulate.

As such, together with the other referendum questions they put out, the KMT decided to use the referendums as a tactic to embarrass President Tsai’s government. And it worked.

Editor: Nick Aspinwall (@Nick1Aspinwall)

If you enjoyed this article and want to receive more like it in your news feed, please be sure to like our Facebook page below.