Asia's Ultra-Low Fertility: Is It a Solvable Problem?

18 years. 378.1 trillion won. And a total fertility rate of 0.80.

Looking at the numbers alone, it’s strange. Common sense suggests that pouring policy funds into a problem should improve results. However, South Korea’s budget for responding to low birth rates has grown 4.7 times since 2006, while the total fertility rate has plummeted from 1.23 to 0.80—a decline of nearly one-third. The budget and the results did not move in the same direction. They moved in opposite directions.

People’s first reaction upon seeing these numbers is generally the same: the policies are wrong. They argue that cash support was insufficient, that budgets were scattered across ministries, or that politicians were only concerned with votes. All of this might be true. However, these explanations share a common premise: that low birth rates are a “malfunction” that can be fixed. It is a belief that if we identify the cause accurately and implement the right policies, the birth rate will rise again.

It is not just South Korea that held this belief. Taiwan, Hong Kong, and Singapore have also tried to fix this “malfunction” in their own ways for decades. And all three have failed. To be precise, the results are so ambiguous that it’s hard to even call it a failure. Despite having policies in place, birth rates continued to head downward, and at some point, the question, “Was this ever really a solvable problem?” began to occupy a more fundamental place than the debate over policy failure.

Countries Suffering from the Same Disease

As of 2025, the figures are as follows: Hong Kong 0.75, South Korea 0.80, Taiwan 0.87, Singapore 0.97, and China is estimated by demographers to be around 0.70. Japan is relatively higher at 1.15, but this is also the lowest figure in its history, recorded alongside the lowest number of births in 126 years of statistical tracking. The OECD average is 1.43. Not a single one of these six East Asian countries and regions exceeds this average.

What these regions have in common is not just the birth rate figures. They share compressed industrialization, high zeal for education, cramped living spaces, and an extreme concentration of educational investment per child. While Hong Kong and Singapore differ from Korea and Taiwan in that their populations are still growing due to immigration despite low birth rates, for Korea, Taiwan, and Japan, a declining birth rate translates directly into population decline. However, when faced with the question of “Why aren’t people having children?” the entire region points in the same direction.

There is another reason why this commonality is important. If Korea’s low birth rate were due to uniquely Korean policy failures—for instance, failed real estate policies or the incompetence of a specific administration—there would be no reason for Taiwan or Singapore, which have different political systems and different real estate markets, to show similar numbers. Yet, they do. And they do so following almost the same trajectory, albeit with a few years of time lag.

Why Policies Fail to Move the Numbers

Here, we need to dig one layer deeper. If the policies failed, what did they fail at?

An analysis released by the National Assembly Budget Office in 2024 points to an interesting observation. The fundamental issues that the Basic Plan for Low Fertility and Aging Society should address are divided into four categories: delayed marriage, declining marriage, declining birth, and the concentration of the population in the capital region. However, the actual policy tools were filled with projects that were either unrelated or only loosely connected to these four issues. The budget carried the label of “low birth rate,” but the contents were a hodgepodge of different policy goals, such as youth employment, housing, childcare infrastructure, and balanced regional development.

This alone remains a bureaucratic explanation for the lack of effectiveness relative to the budget. The truly interesting question comes next: Why did policy designers miss the core of the problem every time? Was it just accidental incompetence, or was this problem the kind that couldn’t be touched by such policies in the first place?

At this point, it is necessary to borrow a lens: coordination failure, as discussed in game theory. Coordination failure refers to a situation where each participant makes a perfectly rational choice from their own perspective, but when those individual choices are aggregated, they result in an inferior outcome that nobody wants. A bank run is a classic example, where rumors of a bank’s insolvency lead all depositors to withdraw their funds simultaneously, actually causing a sound bank to collapse. The individuals did nothing wrong; they were only trying to protect their own money. Yet, when those rationalities are combined, it becomes a disaster.

If we view low birth rates through this frame, the picture changes. There is no irrationality in the process of an individual deciding, “I will not have a child now.” It is a very rational decision that accurately accounts for housing costs, private education expenses, the risk of career interruptions, and the imbalance in domestic labor sharing with a spouse. The problem arises the moment those rational individual decisions are aggregated across society. If everyone decides “not now,” the sum of those judgments becomes a national demographic cliff. Optimal for the individual, worst for society. This is the very definition of coordination failure.

This lens is useful because it explains why policies lack power. Cash support or vouchers only slightly reduce one cost item in an individual’s calculation. They do not change the entire calculation. And in a state of coordination failure, a policy that only touches one item in the equation will not cause a majority of people to shift to a different equilibrium simultaneously. The equilibrium itself only shifts if everyone changes their judgment at the same time—thinking, “Ah, I can have a child now”—but policy is not a tool that can create such collective, simultaneous movement.

The 400 Million Won Newborn, and the Item No One Calculates

In 2023, the budget poured into responding to low birth rates in Korea exceeded 48 trillion won, and 235,000 newborns were born that year. If divided, the budget is large enough to pay 200 million won per newborn. The actual “First Meeting Voucher” provided is 2 million won for the first child and 3 million won for subsequent children. It’s not that the math doesn’t add up, but that most of the low birth rate budget doesn’t go to individuals as cash. It is scattered across the construction of childcare facilities, local government projects, and various programs across multiple ministries.

This fragmentation is a natural result within the coordination failure framework. The government has approached the problem by lowering individual cost items—childcare, housing, education—in the calculation. However, a survey conducted by the Korea Institute for Health and Social Affairs (KIHASA) targeting dual-income couples revealed items that could not be converted into money. When asked about satisfaction with the division of household chores, the average level of dissatisfaction among dual-income wives was more than four times that of their husbands. For single-income wives, the dissatisfaction was more than five times that of their husbands.

This single figure reveals the blind spot in policy design. Even if the government supports childcare costs and lowers housing loans, the way labor is actually divided within the home after having a child is an area that policy cannot touch. Demographer Peter McDonald theorized this structure back in 2000. Ultra-low fertility in societies that have achieved economic growth is the result of a mismatch: while men and women have become almost equal in the public sphere—education, employment, legal rights—the institutions in the private sphere of the family have not kept pace with that equality. Women are trained as equal participants in the workplace, but after marriage and childbirth, they are still expected to adhere to pre-modern gender roles. The greater the gap between these two worlds, the most rational response a woman can make is not to choose between them, but to avoid the situation that creates the gap in the first place—by delaying marriage, not getting married at all, or choosing not to have children even if married.

There is a simple reason why the government has failed to include this gap in its calculations: policies are executed via budget, and budgets are allocated only to measurable items. The imbalance in household chores between spouses is not an item that can be mandated by law or bought directly with a budget. Such items are the hardest to deal with in a coordination failure game. If social norms that are not monetized—and which do not change well within a single generation—account for half of the equation, no matter how much you tweak the other half that can be converted into money, the results will not change significantly.

Different Numbers Even Within the Same Country

Here, a counter-argument might arise: does this mean policies have no effect at all? That is not the case. The fact that regional gaps are widening significantly within the same country is proof of that.

In the 2025 provisional statistics, Jeonnam ranked first in the nation with 1.10, while Seoul recorded the lowest at 0.63. Within a country that uses the same laws, the same national policies, and the same currency, a 1.75-fold gap exists.

This gap does not show the uselessness of policy, but rather the regional variations of coordination failure. Seoul is a place where housing costs, private education competition, working hours, and population density are compressed and concentrated. The cost items in the calculation are piled up much more heavily than in Jeonnam. Conversely, the absolute values of those cost items are lower in Jeonnam. While the policy was applied equally, the underlying conditions—housing costs, competition intensity, work culture—differ by region, so the results appear differently. This does not mean that policy is meaningless, but that policy alone is insufficient to bridge the depth of coordination failure that varies by region.

If we move this gap from numbers to an individual’s choice, the picture becomes clearer. Imagine an office worker in their early 30s working in the Gangnam area of Seoul. This person is not an exceptional case; they are close to the most common type according to statistics. The “jeonse” (key money deposit) price is many times their income, they have to worry about daycare waiting lists the moment they have a child, and from the moment their child enters elementary school, the academy schedule of the neighbor’s child becomes their own child’s schedule. They already know through many precedents that if they apply for parental leave at a workplace where overtime is taken for granted, they will be quietly pushed off the promotion track. When they run the calculator on all these conditions, the conclusion of “not now” is not an incompetent judgment, but the most accurate one based on the information they have. For a couple of a similar age in a rural area of Jeonnam, the calculation they face has a different number of items. As housing costs and competition intensity are lower, the answer pointed to by the calculation also comes out differently. Even though they are citizens of the same country, they are essentially playing different games.

---

Up to this point, this could be read as a story within one country, Korea. But the moment Taiwan and Hong Kong follow the same trajectory, the story moves beyond the explanation of the Korean government’s incompetence.

These three societies commonly underwent compressed industrialization in the latter half of the 20th century. This means they skipped from agricultural societies to high-tech manufacturing and financial societies in just one generation. The changes that Western societies passed through over 100 years—urbanization, expansion of education, and women’s entry into the labor market—these three societies experienced compressed into 30 to 40 years. The problem is that institutions and norms could not keep up with that speed. While women’s education levels and labor market participation rates quickly caught up to Western levels, gender role norms within the family, long-working-hour cultures, and entrance exam competition structures inherited the frames of the pre-industrial generation. The gap created between the speed of modernization in the public sphere and the speed of modernization in the private sphere. This is the debt that compressed modernization inevitably leaves behind.

Policy failure is a symptom of this debt, not the cause. No matter how much budget is poured in, the norms themselves cannot be changed in a few years. Norms move on a generational scale. On the other hand, the term and budget cycle of a policy is 5 years. Essentially, we have been putting two things with different time scales on the same stage and asking why the policy couldn’t beat the norms.

Countries That Spent More vs. Countries That Spent Differently

Korea is not the only country that has conducted this experiment. Hungary is considered the country that has pushed the most aggressive pro-natalist policies for 16 years. They even went as far as providing interest-free lump-sum loans if a couple promised to have two children, and charging penalty-like interest if they failed to keep that promise. The result was a brief rebound from 1.25 in 2010 to 1.59 in 2020, but it then slumped back to 1.31 in 2025. A woman in Budapest interviewed by the BBC regarding this case answered that what is needed is the improvement of education and medical systems, not cash tied to childbirth itself.

Within the same Europe, Sweden has walked a completely different path. The principle stated by a Swedish government official is simple: there is no separate “pro-natalist policy” in Sweden. Family and childcare policies are simply naturally included within the social welfare principle that all employable people participate in labor. A significant portion of parental leave is mandatorily allocated to fathers, and childcare facilities exist as a default, not an exception. Germany followed a similar path. When the birth rate dropped as women’s social advancement increased in the early 2000s, Germany referred to the Swedish model and expanded full-day schools where children could be left until 4 PM. The proportion of full-day schools, which was 16.3% in 2002, increased to 71.5% in 2020. Germany’s total fertility rate subsequently escaped the 1.3 range and rose to the 1.5 range.

The difference between the two continents and two approaches is clear. Hungary and Korea adjusted the cost items of the calculation—cash, loans, vouchers. Sweden and Germany changed the calculation itself. They chose to forcibly redistribute care to society and male spouses so that the act of having and raising a child does not exclusively consume a woman’s career and time. Translating this into the language of the coordination failure game, the former provided participants with slight incentives on top of the existing equilibrium, while the latter moved the equilibrium point itself to a different location.

Whether this difference can be applied to Asia as it is, is another matter. The interventions in Sweden and Germany worked in a way that institutions supported already existing, relatively egalitarian gender norms. In contrast, what Korea, Taiwan, and Hong Kong inherited is the gap itself between norms and institutions left by compressed modernization. This means that even if the same prescription is used, if the soil in which the prescription is placed is different, the results can be different. Korea already surpasses many European countries in the generosity of its parental leave system on paper. The problem is whether that system can actually be used, whether there are disadvantages at work when used, and whether men use that system as naturally as women. The existence of a system and the operation of a system are stories on different levels.

Between Solving and Adapting

If policy cannot overturn coordination failure, there seem to be two options left. One is a radical intervention that shakes the equilibrium point of coordination itself—for example, making parental leave mandatory for both men and women, or a complete reorganization of working hours combined with a 4-day workweek; an intervention that changes the structure of the calculation itself, not just the cost items. The other is to redesign the social system based on the premise of population decline. Extending the retirement age, re-employment of the elderly, supplementing the labor force through increased immigration, and administrative efficiency through regional consolidation. It is not “solving” the low birth rate, but “adapting” to the population structure created by the low birth rate.

What is interesting is that these two options are not mutually exclusive. Rather, the policy failures to date are closer to the result of mixing the two and doing neither properly. They were neither radical enough to shake the equilibrium point of coordination, nor did they seriously embark on structural redesign based on the premise of population decline. It has been 18 years of waiting for results to change while keeping the equilibrium as is, by slightly tweaking a few lowering cost items.

This is why there are conflicting interpretations of the 2025 rebound—from 0.75 to 0.80—between those who say the policy effect is finally appearing and those who say the size of the population at the prime marriage age was just temporarily large. Whichever is correct, no one can be sure at this point whether this rebound is the result of shifting the equilibrium point of coordination itself, or just a coincidental fluctuation on top of that equilibrium point.

Individuals who decided not to have children never made a wrong calculation. They simply chose the most rational answer within their respective conditions. The number that the state calls a loss is the sum of the accurate judgments made by the countless individuals who constitute that state.

Then, the question we really need to ask might be this: Who decides the boundary between what we call a “problem to be solved” and a “condition to be adapted to”? And are the people who set those boundaries calculating the same conditions as the people who actually live within those calculations?

---