新加坡政府親自做紅娘： 諾貝爾獎演算法配對，仲送禮券催拍拖｜ 愛情顧問解說

📰 What Happened

At the end of September, Singapore's Government Technology Agency (GovTech) launched an official dating service called FirstDate, initially open only to single civil servants aged 21 to 35, with applications closing on 5 October. Applicants must first log in through Singpass, Singapore's national digital identity system, to verify their identity; singles, widowed and divorced people are all eligible. The design itself makes one thing clear: it exists to screen out fake accounts and married people.

The matching process works like this: users complete a questionnaire covering interests, lifestyle habits, values and preferences, which may include religion, age, and even "deal-breakers." The system uses the Gale-Shapley stable matching algorithm and pairs you with just one person per cycle, along with a compatibility score, a short profile and a personal message from the other person — and note, a cycle does not always produce a match. Both sides receive an SMS and must decide whether to accept within 72 hours; only if both say yes are contact details exchanged, and personal information is visible only to the person you are matched with.

The origin of the service is actually rather interesting: it grew out of a GovTech annual hackathon, where the team was researching the question "what difficulties do people face when meeting someone new?" The original idea was to offer free meals, but the formal terms were changed to a "Date Quest" — a dating task such as trying a dish that neither person has eaten before, which may earn a third-party voucher on completion. However, the value and number of vouchers are decided by GovTech and may be allocated on a first-come, first-served basis, so a free meal is not guaranteed.

The background numbers are really the heart of this story: Singapore's resident total fertility rate fell from 0.97 in 2024 to 0.87 in 2025, and the number of births in 2025 fell below 30,000 for the first time in 60 years. This is not the government's first time playing matchmaker — it set up the Social Development Unit back in 1984 to encourage university-educated singles to socialise.

Public reaction has been polarised: some say the service discriminates against non-civil servants, while others criticise the age cap. Whether it will open up to the public after the trial is still undecided.

On the same day, Hong Kong's Legislative Council Panel on Manpower discussed 11 pro-natalist measures from the Policy Address. One lawmaker suggested partnering with the Jockey Club or commercial organisations to launch a "fertility lottery," under which babies born over the course of a year could win prizes of HK$1 million to HK$4 million. Deputy Chief Secretary for Administration Cheuk Wing-hing responded that prizes of hundreds of thousands to millions do carry some motivational pull, but that the resources of both the government and the Jockey Club are limited, and the commitment would be an "astronomical figure." Hong Kong's total fertility rate in 2024 was 0.733, and the number of births from mid-2025 to mid-2026 was 29,700.

👩‍💼 The Advisor's Take

① Governments playing matchmaker is nothing new — what is new is stepping directly into "matching"

Governments acting as matchmaker is hardly a novelty. Singapore set up the Social Development Unit back in 1984; Japan has government- and locally-subsidised kon-katsu (marriage-hunting activities); and South Korea and the Communist Youth League in mainland China have both run official matchmaking events. So the question is not "should the government do it," but "how deep should it go."

Where is the difference? The old model was about "creating occasions" — running social mixers, subsidising matchmaking fees — and in the end, who you picked and whether it worked out was still your own business. FirstDate is different: it steps directly into "screening and matching." A questionnaire decides who you are paired with, and the algorithm decides that you have only one choice this cycle. The government shifts from "venue provider" to "matchmaker."

Behind this shift lies an economic assumption — it treats singleness as a "market failure." In other words, being single is not because you don't want to date, but because "matching efficiency is too low": search costs are high, information is asymmetric, and suitable people are scattered across different circles, never seeing each other. So the government steps in and builds a central matching platform.

This is where we need to be clear about the essence of the Gale-Shapley algorithm. It was proposed by Gale and Shapley in 1962, and Shapley won the 2012 Nobel Memorial Prize in Economic Sciences for this contribution (together with Alvin Roth). Note carefully: what it pursues is not the "happiest match" but the "stable matching" — defined as a state where, once matching is complete, no pair of a man and a woman would both rather choose each other over their current partners. This mathematical property is beautiful, but it is something entirely different from "happiness."

So the first tension emerges: the government is using an algorithm designed to "prevent people from breaking apart" to tackle the question of "how to make things last."

② An algorithm can match "meeting" — it cannot match "being together"

The most worthwhile thing to unpack in this story is that "match success" and "relationship success" are two entirely different layers. The former is a technical problem, the latter a psychological one — and FirstDate only covers the first layer.

Psychological research consistently points to the same conclusion: the strongest predictor of whether a relationship lasts is not "how well the personality checklists align," but "how conflicts are handled" and "the quality of everyday interaction." Gottman's long-term tracking research is the classic example — from just a few minutes of videotaped interaction between a couple, he can predict with considerable accuracy whether they will divorce, based on the frequency of the "Four Horsemen" (criticism, contempt, defensiveness and stonewalling), rather than on whether their interests align.

So even if FirstDate achieves extremely high matching efficiency, at most it compresses the "meeting" step; the "being together" step — aligning values, attitudes to money, family expectations, conflict patterns — still has to be walked by the two people themselves. This also explains why so many couples "break up after a few dates": they didn't fail to find someone, but the compatibility score was high while their relationship skills were not yet ready.

More crucially, there are several flights of stairs between "matching success" and "fertility rate." To get from "matched successfully" to "a baby is born," you still have to pass through, at minimum: being together → deciding to marry → deciding to have children after marriage → how many children to have. Each step has its own attrition rate. Singapore's difficulty is not only that "singles can't find a partner," but also delayed marriage age, delayed childbearing after marriage, high child-rearing costs, and the large opportunity cost of women having children.

Breaking it down by case, the people this app actually helps are fewer than you might imagine:

• Case one: genuinely want to date, but their circle is too narrow. Engineers, healthcare workers, shift workers, expatriates — their problem really is "not meeting people." Official matchmaking helps, because it addresses precisely the search cost.
• Case two: date successfully, but don't want children. Matching cannot help at all. Their difficulty lies at the "having children" step, not the "meeting people" step.
• Case three: don't even want to date. Economic pressure, prioritising personal freedom, seeing marriage as a burden — this group isn't "unable to find someone" but "not planning to look," and no matchmaking tool can reach them.

The government's policy assumption is that "dating → marriage → children" is a straight line; in reality these are three relatively independent decisions. Using an upstream tool to solve a downstream problem is a classic mismatch.

③ What Hong Kong should copy is not "playing matchmaker" but "lowering search costs"

Now look at Hong Kong. Hong Kong has no universal, real-name social identity platform — even "iAM Smart" has yet to achieve the penetration and trust of Singpass. And romantic preferences are among the most private kinds of data — religion, age limits, "deal-breakers" — so once the government holds this data, the threshold of public trust would be very high. The path of "the Hong Kong government building its own matchmaking app" therefore faces no small resistance.

But what this story is genuinely worth Hong Kong's reference is its underlying approach: a falling marriage rate cannot necessarily be solved by "handing out money" alone; you can also start from "matching efficiency." Hong Kong in fact has a large number of semi-official and non-profit organisations — social welfare agencies, professional bodies, alumni associations — and what they can do is "networking": not having the government hold personal data itself, but building a standardised, trustworthy, low-threshold channel for meeting people.

Compare this with Hong Kong's "fertility lottery" discussed the same day: the two approaches target two entirely different groups. Handing out HK$1 million to HK$4 million affects people who already intended to have children, because the money arrives only after they have made the decision; official matchmaking affects people who "genuinely want to date but can't find anyone." And neither scheme reaches the largest group — those who "do the maths and decide not to have children." What this group wants is not prize money, but housing, childcare, working hours and workplace protection for women.

To sum up in one line: Singapore's move is an "honest but limited" experiment. It acknowledges the pain points of singles, but it is using a "matching" tool to respond to a problem of "cost and values."

💡 What You Can Learn

First, don't mistake "being matched" for "being solved." The government or an app can help you meet someone, but no algorithm can help you be together. A high compatibility score only means the starting point is smooth; it does not mean you will go far. What truly determines the length of a relationship is how you argue, how you repair, and how you face reality together.

Second, too many choices numb you, but too few also causes problems. FirstDate's "one match per cycle" design targets swipe-culture fatigue — in psychology this is the "paradox of choice": with too many options, people find it harder to decide and less satisfied. But conversely, if the matching logic is mismatched with your real needs, you will mistakenly think "no one is right for me," when in fact "the system can't find the kind of person I want."

Third, be clear about what you want. If you are someone with a narrow circle who genuinely wants to date, this kind of official or semi-official matching is worth a try, because your problem really is "not meeting people." But if you are just following the crowd, you might as well save the spot for someone who genuinely needs it.

Finally, this story from Singapore actually reflects one thing: when a place has to mobilise the government, algorithms and vouchers to "nudge" people into dating, it shows that dating has shifted from something "natural" to something that "needs to be pushed." That signal is the most worthwhile thing for Hongkongers to pay attention to.

Talk to a LoveWise love advisor — if you also feel that it's not that you don't want to date, you just don't know where to meet someone, we can help you see clearly what you want 💕