Eve is here. The mill towns, or parts of the towns where I lived that had factories as important employers, have shrunk, although they are not yet ghost towns. Escanaba, Michigan’s population was over 15,000 in the 1970s, but is now less than 13,000. Chillicothe, Ohio’s population was 29,000 in the 1970s, but today it’s less than 22,000. And when it comes to aging, it’s hard to beat Bailey Island, Maine, where my father’s family has lived since they first settled there. Median age was 59 years. Grumpy Yankees!
Written by Elisa Giannone, Junior Researcher, Center for International Economic Research (CREi) Yuhei Miyauchi, Assistant Professor, Department of Economics, Boston University. Xinle Pang, assistant professor at Buffalo School of Economics; Yuta Suzuki, Antai School of Economics and Management, Shanghai Jiao Tong University, and Nuno Miguel Marquez da Paixao, Senior Economist, Bank of Canada. Originally published on VoxEU
Policy discussions about aging and population decline often treat the issue as a national problem. In this column, I will introduce evidence that Japan is also having a significant spatial influence. Aging and population decline will hit rural areas first. Population outflows from regions that are currently shrinking will reduce the number of births tomorrow, resulting in further aging of the original regions. However, policies that reduce spatial inequality also reduce aggregate efficiency, creating a trade-off in which policy makers are faced with the challenge of choosing between different goals.
Governments are once again asking how much to invest in regions that are aging, shrinking and with fewer young people. Japan has restarted the task of regional revitalization by clearly linking regional policies to the declining birthrate and aging population and the declining birthrate and aging population (Cabinet Secretariat 2024). Spain has mobilized more than €10 billion across 130 measures to address the “demographic challenge” (Government of Spain 2021). Italy’s Interior National Strategy (OpenCoesione 2026) and the European Commission’s Long-Term Vision for Rural Areas (European Commission 2021) similarly place essential services, connectivity and rural development at the center of their policies.
The central question is not simply whether shrinking regions deserve support, but what such support would look like and at what cost. Economists have warned that shrinking municipalities may face reduced amenity as public services exhibit economies of scale (Heinemann et al. 2007), and that aging populations may weigh on growth (Kotschy and Bloom 2023). The missing piece is geography. Who leaves, who stays, and how these choices reshape the financial and economic foundations of communities.
Giannone et al. (2026) study this question using Japan as a laboratory. Japan is useful not only because it is one of the world’s oldest and fastest shrinking countries, but also because its demographic future overshadows problems already seen in Europe and parts of East Asia. By 2015, 26% of Japan’s population will be aged 65 or older, and this proportion is projected to reach 37% by 2050. Japan’s aging wave can be considered a cautionary case for the social and financial pressures associated with population aging (Stawasz et al. 2018). But these national numbers hide striking spatial patterns. While the proportion of elderly people in some municipalities is already close to 50%, metropolitan areas remain young and densely populated.
The first fact is simple but important. This means that population decline and aging have not even spread to space. From 1980 to 2010, Japanese municipalities, which were already relatively old, lost population and became older faster. Municipalities that were young in 1980 tended to grow and remain young. Figure 1 visually illustrates this historical difference.
Figure 1 Historical differences in demographics in Japanese local governments from 1980 to 2010
Note: In 1980, older municipalities lost population and became even older, while younger municipalities tended to grow and stay young.
Source: Giannone et al. (2026), Figure 4.
This also highlights an important timing point. In other words, although Japan’s overall population decline began around 2010, regional aging and population decline were already underway long before that. This discrepancy is not simply due to differences in birth rates depending on location. This is also a story about migration. Young people continue to move to big cities, especially Tokyo, and their children are born there rather than where their parents left. The result is a cumulative population mechanism. In other words, out-migration today reduces the number of births tomorrow, which in turn causes regions of origin to become older.
This mechanism is important because regional economies are built around the regional scale. Stores, clinics, schools, childcare centers, public transportation, and city halls all depend on enough users and taxpayers to survive. Using municipality-level data and an instrumental variables strategy based on historical migration links, we find that areas with declining working-age populations experience declines in several measures of local amenities, such as retail trade and health services. At the same time, the per capita cost of local public spending increases. The cost of road networks, school buildings, and city halls does not decrease proportionately as the population declines. This is a financial calculation for a ghost town. There are fewer residents, fewer workers, fewer children, and higher costs per person.
To study where these forces direct us, we build a dynamic spatial model in which people of different ages choose where to live, taking into account wages, housing costs, amenities, migration costs, pensions, and taxes. Because dense locations support productivity and services, wages and amenities are likely to rise with the local population. When more people want to live in the same place, housing costs rise. Local public services also cost more per person in smaller places. The model is calibrated to Japanese data and used to predict future regional outcomes.
Baseline predictions are sovereign. Tokyo’s share of Japan’s population will rise from about 10% in 2015 to about 26% over the next two centuries. The combined population share of the five oldest prefectures (Kochi, Shimane, Tokushima, Tottori, and Yamagata) will drop from approximately 3% to less than 1%. While the proportion of elderly people is approaching 60%, in Tokyo it remains at nearly one-third. Blocking internal migration or fixing the spatial distribution of newborns at 2015 levels would eliminate much of this regional variation. In other words, the future geography of aging will not be mechanically determined solely by a country’s birth and death rates. It largely depends on the spatial selection of young populations and where the next generation is born.
Figure 2 Prediction of population proportion and proportion of elderly people in Tokyo and the five oldest prefectures
Source: Giannone et al. (2026), Figure 9.
This finding connects aging to the broader urban economics literature on agglomeration and regional divergence (Moretti 2012, Diamond 2016). Large cities can offer higher productivity and more amenities, but they also come with higher housing costs. In our simulations, Tokyo’s productivity and amenity benefits dominate the effects of residential congestion. As the population becomes more concentrated, the attractiveness of a city increases, while areas in decline lose size. This reinforces regional inequalities in access to distribution, especially for working-age residents. Although older people are partially protected by pensions provided by the state rather than the local government, they remain dependent on local services and amenities.
However, there are some difficult developments. The same redistribution that exacerbates spatial inequality can increase overall efficiency. By concentrating more people in areas with higher productivity and lower costs, average labor incomes increase and the national financial burden of providing local public services is reduced. Blocking migration or implicitly freezing the geography of birth will keep more people in smaller, less productive areas where the per capita cost of public services is higher. Therefore, the policy issue is not a simple choice between “good” support and “bad” abandonment of declining regions. It is a trade-off relationship with stock efficiency.
Place-based policies sit right on this border. Recent research on place-based industrial policy shows that subsidies can have important spillover effects and have only a modest impact on regional inequality (Atalay et al. 2023). Another recent study shows that retirement migration in France can bring economic benefits to poorer rural areas (Badilla-Maroto et al. 2026), reminding us that not all migration flows strengthen metropolitan concentration. But a central force in Japan’s aging population is the movement of young workers and future parents away from older areas.
We simulate one transparent policy. It is a transfer to residents of the five oldest prefectures, funded by Tokyo residents’ taxes. This is not intended to duplicate any particular program. This is a benchmark for understanding the economic forces behind regional revitalization policies. In this context, the duration of the policy is important, as migration, fertility, and regional-scale effects unfold slowly. One-off subsidies may temporarily cushion the economic downturn, but long-term policies change expectations about where it’s worth living, working, and raising children.
Therefore, if a transfer equivalent to 5% of income is maintained for 100 years, the effect will be large. By 2065, the population of the five oldest prefectures will almost double compared to the baseline, and the proportion of elderly people will fall from about 45% to about one-third. The increase in real regional income is greater than the transfer itself, as more residents increase their productivity and amenities through the regional scale.
But politics is not free. The same 5% transfer reduces total per capita labor income by more than 1% and increases total per capita fiscal expenditure by about 0.5%. These costs arise because relocation leaves many people in places where productivity is low and public services are more costly to provide. Although this policy reduces spatial inequality, it also reduces aggregate efficiency.
Figure 3 Impact of relocation to the five oldest prefectures on population share and elderly share in 2065
Source: Giannone et al. (2026), Figure 13.
The main implication is that national discussions about aging and population decline require a spatial lens. Countries may shrink while the largest cities expand their population share. Aging societies can lead to further regional inequality, even as people rationally move toward opportunity. Subsidies can slow the hollowing out of rural areas, but they can also reshape where people live, work and have children. That is, the benefits and costs accumulate over decades.
The important question for policy makers is not whether all villages can or should return to their previous population. It’s how we choose between a variety of goals, including protecting access to critical services, protecting communities, supporting migration, encouraging births and settlements in areas of population decline, and maintaining aggregate productivity. Our results suggest that these goals cannot be maximized all at once. Living in a ghost town is not just a demographic issue. This is a spatial economics problem and requires policies that directly confront the geography of aging.
See original post for reference
