Japan Lost 3.1 Million People in Five Years
The 2025 Census Shows What Comes Next

Japan's population decline is no longer new. What the 2025 census changes is the scale, speed and geography of it.

According to the Statistics Bureau's preliminary count, Japan had 123,049,524 residents on October 1, 2025. That was 3,096,575 fewer people than in 2020—a decline of 2.5% in five years. During the previous five-year census interval, the decline had been only 0.7%. In raw numbers, the loss accelerated from roughly 949,000 people to nearly 3.10 million.

A century of censuses now points decisively downward

Japan conducted its first national census in 1920, when the population was 55,963,053. Except for the disruption of the war years, the long twentieth-century story was growth.

Industrialization, urbanization and rising longevity carried the population above 100 million and then 120 million. Census population reached 126.93 million in 2000 and peaked at 128.06 million in 2010.

By 2015 it had fallen to 127.09 million, marking the country's entry into sustained census-to-census decline. The 2020 count was 126.15 million. Five years later, it was 123.05 million.

The important change is the slope. Japan lost about 963,000 people from 2010 to 2015 and about 949,000 from 2015 to 2020. From 2020 to 2025, it lost almost 3.10 million.

Japan's demographic question is no longer when population decline begins. It is how quickly institutions built for growth can be redesigned for contraction.

Tokyo gained 199,000 people. Hokkaido lost 239,000.

The national average hides enormous regional variation. Tokyo, the largest prefecture by population, reached 14,246,219 residents and gained 198,621 from 2020, an increase of 1.4%. Okinawa gained just 740 people, or 0.1%.

Every other prefecture shrank. Hokkaido recorded the largest numerical loss, down 239,195. Shizuoka lost 164,357; Hyogo 141,177; Niigata 132,796; and Fukushima 121,215.

Percentage changes show the strain more clearly. Akita fell 8.1% in five years, Aomori 7.9%, and both Iwate and Kochi 7.0%. A decline of that magnitude means losing roughly one resident in thirteen within a single census interval.

The policy problems at the two ends are almost opposites. Tokyo must manage housing demand, congestion and service capacity. Shrinking regions must decide how to sustain schools, stores, hospitals, buses, roads and water systems for fewer people.

Thirty percent of Japan now lives in the Tokyo region

Tokyo, Kanagawa, Saitama and Chiba together held 36.986 million people, 30.1% of the national population. The region as a whole grew by 71,000 from 2020.

That increase came despite Kanagawa, Saitama and Chiba each recording a population decline; Tokyo's gain outweighed them. The pattern suggests that concentration is becoming increasingly centered on the capital itself rather than broad suburban expansion.

The census count alone cannot explain who moved and why. The Statistics Bureau plans to release detailed internal-migration tabulations by December 2026. Those data should clarify the age structure and origins of the flows.

Population fell, but the number of households rose

Japan counted 57.125 million households in 2025, up 1.294 million—or 2.3%—from 2020.

That can seem contradictory: 3.1 million fewer residents but more households. It makes sense if households are becoming smaller. More one-person households, older couples living alone and households without children can increase the number of residential units even while total population falls.

This is one reason demographic contraction does not translate neatly into lower infrastructure cost. Fewer people do not automatically mean proportionally fewer homes, electricity connections, garbage routes, water pipes, delivery stops or administrative cases.

If people remain geographically dispersed, the cost per resident of maintaining networks can rise.

The engine of decline is births minus deaths

Population changes through two broad channels: natural change—the difference between births and deaths—and migration.

In the most recent rolling 12-month preliminary vital-statistics period, April 2025 through March 2026, Japan recorded 706,153 births and 1,559,047 deaths. That was a natural decrease of 852,894 people.

That one-year flow cannot be used to reconstruct the entire five-year census change. But it shows why Japan's contraction cannot be explained simply by young people moving from rural areas to Tokyo. At the national level, deaths substantially exceed births.

The number of children has fallen for 45 consecutive years

The Statistics Bureau estimated 13.29 million children under age 15 on April 1, 2026, down 370,000 from the previous year. The child population has now declined for 45 consecutive years and accounts for 10.8% of the population.

This is not a distant future problem. Falling child numbers are already reshaping school consolidation, after-school activities, pediatric access, retail, tutoring and public-transport demand.

Closing schools may improve operating efficiency, but it can increase travel distance and require new school-bus spending. A school can also be a community hub, not merely a classroom building. Demographic efficiency and local accessibility often pull in opposite directions.

At the same time, the 75-plus population is still growing

The September 2026 population estimate put the population age 65 or older at 36.211 million. Within that group, people age 75 or older numbered 21.531 million, up 479,000 from a year earlier.

That matters because demand for health care, home care, emergency services and mobility support does not necessarily fall when total population falls. In some communities it can rise.

This is why “fewer people means fewer public services” is not a workable formula. The composition of demand changes at a different speed from the population total.

The working-age population is down to 73.34 million

The same September estimate put the population age 15 to 64 at 73.336 million, down 222,000 from a year earlier.

The effect is larger than a generic labor shortage. Japan needs fewer workers to serve a smaller population in some sectors, but more workers in others. Care workers, bus drivers, construction trades, municipal staff, medical professionals, teachers and logistics workers are themselves drawn from a shrinking labor pool.

Services losing customers and services gaining elderly demand therefore coexist—and both must be operated by fewer working-age people.

Regional buses are losing both passengers and drivers

Japan's transport ministry has repeatedly described the double squeeze on regional public transport. Long-term population decline and aging reduce ridership, while driver shortages make it harder to maintain routes even when communities still need them.

A bus company can therefore lose revenue and labor at the same time. Cutting frequency can make the service less useful, pushing more riders away. Yet older residents who no longer drive may depend on the same bus more than ever.

That is why regional rail and bus policy increasingly looks less like a normal private-business question and more like essential-infrastructure policy.

Water pipes do not get shorter when the population shrinks

Much of Japan's road, bridge, water, sewer and public-facility network was built during decades of population and economic expansion.

Population decline does not physically shorten those networks. If settlement becomes less dense, infrastructure length per resident can actually rise. Fee revenue and tax bases weaken while pipes, bridges and buildings still reach replacement age.

This mismatch is one reason MLIT has promoted broader regional approaches to infrastructure management, grouping assets across municipalities and sectors rather than assuming every town can maintain every system independently.

Foreign population growth is cushioning part of the decline

Japan's September 2026 population estimate put the foreign population at 3.917 million, up 301,000—or 8.3%—from a year earlier. The Japanese population, by contrast, was 118.921 million, down 860,000.

Foreign residents are already supporting labor supply across manufacturing, care, construction, restaurants, logistics, IT and other sectors, and their growth partially offsets national population loss.

But immigration is not a one-variable demographic fix. Housing, Japanese-language education, schools, health care, social insurance, family settlement and community integration determine whether foreign workers become long-term residents and whether local services adapt effectively.

Shrinking municipalities will become the norm, not the exception

Japan's National Institute of Population and Social Security Research projects that 1,610 municipalities—93.2% of the total in its standardized geography—will lose population between 2025 and 2030. By the 2045–2050 interval, the projection rises to 1,709 municipalities, or 98.9%.

Those are projections, not census observations. Births, deaths and migration can diverge from the assumptions. But the planning implication is powerful: government can no longer treat a shrinking municipality as an unusual case requiring a special program.

Outside a limited set of growth centers, schools, hospitals, transit providers and businesses need models that work without assuming a larger customer base every year.

A projection is a design condition, not destiny

Demographic projections are conditional. They apply assumptions about fertility, mortality and migration and calculate what follows if those assumptions hold.

It is therefore too simplistic to say “rural Japan will disappear” or “more immigration solves the problem.” The same population loss can produce very different outcomes depending on settlement density, transport access, universities, hospitals, industry, housing and migration.

A town that concentrates residents around services may preserve access with fewer people. A town with the same population scattered across a large territory may face much higher costs.

The next stage of demographic policy is therefore about geography as much as headcount: where people live, how old they are, what households they form and how far they must travel for essential services.

Local economics depends on density, not just population

Supermarkets, clinics, gas stations, banks, restaurants, tutoring centers and delivery networks all need some viable customer catchment.

If a region loses 10% of its population but the remaining residents cluster around a center, some services may remain viable. If the same number of people is spread thinly across mountains, islands or extensive suburbs, travel and delivery distances grow.

That is the logic behind Japan's “compact plus network” planning concept: concentrate some housing and essential functions while linking them through public transport. It is an attempt to redesign service geography when restoring population growth is not a realistic near-term strategy.

Shrinking and declining are not the same thing

Population loss constrains economic growth, tax revenue, housing markets and social insurance. It does not automatically mean living standards must fall in lockstep.

Vacant housing can be consolidated. Former schools can become multi-service centers. Demand-responsive transport can replace lightly used fixed routes. Telemedicine can connect local clinics to larger hospitals. Infrastructure management can be shared across municipalities. Automation can stretch scarce labor.

The mistake would be to preserve every institution exactly as it was built for a growing population and wait for population growth to return.

The policy challenge is to build systems that remain functional with fewer people.

“A city disappeared” is the wrong metaphor for 3.1 million fewer people

A loss of 3.1 million people in five years invites the comparison that a large city has vanished. But that is not how Japan's demographic contraction works.

The residents disappear gradually across thousands of communities. Households remain. Older people continue to need transport and health care. Roads and water pipes remain in place. Schools and hospitals still have geographic coverage obligations. At the same time, Tokyo continues to attract people.

That is what makes the problem difficult.

The challenge of a shrinking Japan is not simply to close places after people leave. It is to decide which services can be consolidated, which choices should remain locally available and how far a person should have to travel to reach the things required for ordinary life.

The 2025 census's 2.5% decline is the number that says this debate can no longer be conducted in the future tense.