OK, my first Japan post isn't my intended introduction to "ordinary" Japan. Instead, it’s from the bookend, because it is one of several areas in which change in Japan presaged rather than followed that in the rest of the developed world.
To start: a "crisis" is something suddenly salient. Absent famine, war and pestilence, that's not the case with demographics.
That's because we know today how many (potential) mothers there will be a quarter century hence, because they have already been born. We also know roughly how many children they will have, because reproductive behavior changes but slowly. So we can look to 2070 with considerable accuracy at the younger age end of the Japanese population.1 While most of those looking at the postwar Japanese growth miracle didn't realize it, the same was true of 1970: an eventual decline in population was already locked into place. Hence the headlines about a “crisis” with another fall in population in 2023.
Japan’s Population Decline
Looking back, Japan still had a young population in 1970, the blue data in the chart below. The impact of the baby boom of the immediate postwar period is clearly visible, and the "echo" from when those early post-WWII women began to have children. Fifty years later (the 2020 data in pink) and the "echo" babies of 1970 are already a decade past their childbearing years. When we look at the 2070 data, in gray, we see that despite the comparatively large size of the echo cohorts, continued low fertility meant the echo of the echo is barely visible. It's also delayed, because by the turn of the millennium women were having children later in life. In another fifty years the overall population will be smaller, the number of potential future mothers will be in rapid decline, and the population will be much older, with the largest single cohort age 70. That demographic momentum is irreversible.
Japan Population Pyramid, Women Only, 1970, 2020, 2070
Source: Author's graph, based on data from ipss.or.jp2
Women are having children later
Now the focus of the news stories is on the total population. I generate the graph below using the same data. As an economist, though, it’s the age structure of the population that matters, the number of young dependents receiving education and (mostly) not working, and the number of older individuals drawing on savings and pensions, and increasingly using healthcare resources. That’s for a much later post, and the relative size and growth rate of the labor force. I don’t detail the implications here, instead I focus on the demographic basics of Japan’s ongoing population decline.
Hand-wringing about how to "solve" the crisis is pointless (or more often, deliberate political grandstanding). That's for the same reason: policies to address the reasons women choose to have fewer children would take a decade to implement, and so even if effective – a big "if" – they wouldn't start to change the number of future mothers until 35 years have passed. Demographic trends exhibit huge momentum, as seen in the population pyramid graph. After the initial drop in fertility in Japan, fertility rates changed only slowly (and the surprises were downward). If the government had really wanted to "solve" the population problem, it should have started 50 years ago. By the 1990s, a change in the policy environment was too late to matter. Now it diverts attention from addressing the consequences of population change, the topic of a later post.
Demographers have found no simple answer for why women choose small families, but here are some factors.
As previous generations left the farm, the need for child labor vanished. In Japan, that occurred by the end of the 1950s. Mandatory education reinforced that. So for the past 70 years children incurred costs but brought no financial benefits. By 1952 fertility returned to pre-WWII levels, and by 1960 was about 2.0, meaning each women had on average one daughter, meaning that population growth would eventually end. Then from the mid-1970s fertility dropped below the replacement level and continued falling. That Japan's population would decline was then locked into place.3
By now late marriage and small families are deeply ingrained. Urbanization and education also changed women's employment options. Self-employment and small family businesses are no longer central to the Japanese economy, so the ability those provide for women to shift back and forth between childcare and work are no longer relevant. Instead, formal work ("labor force participation") becomes central. Almost all Japanese women graduate from high school, and 70% go on to a college or technical school. Having children thus bears the high opportunity cost of lost income, which is accentuated where women want to pursue a career where continuity of employment is important. Daycare can help when jobs are local and don't entail overtime – or where employers are understanding. Grandparents are even more helpful...spoken as one who cared for two granddaughters for 3 months in 2020, as their parents were both essential workers, and their mother was regularly exposed to COVID.
One way to indirectly observe these employment pressures is an "M-chart," named from its shape, because around the world women dropped out of the labor force to raise families and then reentered when children were in middle school.4 The shift in Japan over the past half-century, between 1973 and 2023, is huge, particularly for the age 25-29 bracket. (Note, too, the marked rise in older age brackets – the graph's scale starts at 10%, obscuring that in 2023 one in eight Japanese over the age of 70 were still working, though typically not full-time.
Data drawn from Japan eStat 長期時系列表３(2): 年齢階級(５歳階級)別労働力人口及び労働力人口比率 － 全国
Rearranging the underlying data to focus on the key childbearing ages highlights an additional dynamic. Starting in the mid-1970s, women's labor force participation age 25-29 increased steadily. That didn't happen for women ages 30-34 and 35-39. Then in the 1990s the 30-34 bracket started moving in parallel, with about a decade gap between them. In other words, on average the 20-something working in the late 1980s continued working.5 OK, keep looking: from the mid-2000s, the graphs for women in their 30s overlap. Women who returned to the labor force stayed in the labor force, and the share working rose from 60% to 80%. Work became the norm. But that's not central to demographics, because most Japanese women still complete their childbearing by age 35. Still, it's indicative of a change in mindset. A huge one – a 45 percentage point increase for women ages 24-29 and 40 percentage points for women ages 30-34.6
Finally, some miscellaneous data on marriage. For both women and men, the age of first marriage rose. Japan been a late marriage society for centuries, compared to northern Europe and the US, where in several states marriage at age 13 is still legal. Already in 1900, women on average didn't get married until age 24. It's now age 31, and the age at which women had their first child rose from 24.8 in 1950 to 30.6 in 2022. In addition, only 5% of women born in the early 1950s never married. In 2020, however, 19% of women in their early 40s were still unmarried. Marriage and childbearing are no longer universal.7
Most women who have a child have a second child. Indeed, surveys show that most married women’s desire for two children hasn’t changed over the past 5 decades (as per 2021 [16th iteration of the] Basic Survey on Childbearing 出生動向基本調査). However, with later marriage, the number who ultimately have only 1 child has risen from 1 in 10 to 1 in 5. Since only 1 in 6 women now go on to have a 3rd child, "higher parity" births fail to make up for the women who only have one child, willingly or not. In addition, today 9.9% of women who are past childbearing age never had a child, up from 3.1% of women in the early postwar years. (There are very few teen mothers in Japan, and very few unwed mothers of any age.)
Conclusion
Population decline is not something sudden, it's not a “crisis”. In Japan’s case it was set in motion a half century ago. Despite lots of hand-wringing, it's not amenable to policy, except in small countries that can encourage immigration in meaningful numbers. Even so, immigration has a smaller impact on the age distribution as immigrants tend to retire in their new home. While in Japan there are now 3 million immigrants and temporary residents, the language barrier impedes skilled working-age immigrants in healthcare, engineering and other critical sectors, because such jobs require an ability to read and write Japanese, and virtually no one coming to Japan is already literate.
Let me circle back to the "ordinary" trope. In demographics, Japan has been in the vanguard of change, with fertility falling earlier, and the aggregate population in decline since 2008. But in most of the world outside of sub-Saharan Africa fertility is now below replacement levels, and future population decline is locked into place. Indeed, the overall populations of Korea, China, Russia and parts of Europe are already falling. Japan is typical, not exceptional.
In living memory Japan was a society of large families and a rapidly growing population. Now schools are closing for lack of students, while old age care is strained for resources. In this Japan is now ordinary. I will discuss the economic impact of demographics at some point, but my next couple posts will return to the automotive industry.
I don't include data on men, because they don’t really matter for demographic projections. There are more than enough men to go around, as for past half century marriage in Japan hasn’t been universal, so any woman who wanted a husband could find one.
In my book Competitive Ties (Columbia University Press, 1991), I provide an example of a factory that set its workday to correspond to the hours of the local schools to facilitate women returning to work when their youngest child started school.
This is an exaggeration, because women, so the women who had children early and returned to work offset those who married later and started families later – the marriage age rose, and in Japan single motherhood, and teen-age motherhood, are very unusual.
Most data are from the annual statistical compendium of the National Institute of Population and Social Security Research, https://www.ipss.go.jp/syoushika/tohkei/Popular/Popular2024.asp. Few or none of the web pages have English headings. For example, see 表６-４ 性，年齢（５歳階級）別初婚率 on age at first marriage and 表６-２３ 性別50歳時の未婚割合，有配偶割合，死別割合および離別割合 on the share never-married at age 50.
Duh, I read Japanese, including fiction and history. The US has lawyer novels, but (except in translation) that's not a genre in Japan. Instead there's a genre of business(man) novels, almost entirely lacking in English. One series I have read, on the growth of supermarkets, was even been turned into a movie.
Share this post
Japan: Falling Population is a Challenge, Not a Crisis
Share this post
Introduction
OK, my first Japan post isn't my intended introduction to "ordinary" Japan. Instead, it’s from the bookend, because it is one of several areas in which change in Japan presaged rather than followed that in the rest of the developed world.
To start: a "crisis" is something suddenly salient. Absent famine, war and pestilence, that's not the case with demographics.
That's because we know today how many (potential) mothers there will be a quarter century hence, because they have already been born. We also know roughly how many children they will have, because reproductive behavior changes but slowly. So we can look to 2070 with considerable accuracy at the younger age end of the Japanese population.1 While most of those looking at the postwar Japanese growth miracle didn't realize it, the same was true of 1970: an eventual decline in population was already locked into place. Hence the headlines about a “crisis” with another fall in population in 2023.
Japan’s Population Decline
Looking back, Japan still had a young population in 1970, the blue data in the chart below. The impact of the baby boom of the immediate postwar period is clearly visible, and the "echo" from when those early post-WWII women began to have children. Fifty years later (the 2020 data in pink) and the "echo" babies of 1970 are already a decade past their childbearing years. When we look at the 2070 data, in gray, we see that despite the comparatively large size of the echo cohorts, continued low fertility meant the echo of the echo is barely visible. It's also delayed, because by the turn of the millennium women were having children later in life. In another fifty years the overall population will be smaller, the number of potential future mothers will be in rapid decline, and the population will be much older, with the largest single cohort age 70. That demographic momentum is irreversible.
Source: Author's graph, based on data from ipss.or.jp2
Now the focus of the news stories is on the total population. I generate the graph below using the same data. As an economist, though, it’s the age structure of the population that matters, the number of young dependents receiving education and (mostly) not working, and the number of older individuals drawing on savings and pensions, and increasingly using healthcare resources. That’s for a much later post, and the relative size and growth rate of the labor force. I don’t detail the implications here, instead I focus on the demographic basics of Japan’s ongoing population decline.
Hand-wringing about how to "solve" the crisis is pointless (or more often, deliberate political grandstanding). That's for the same reason: policies to address the reasons women choose to have fewer children would take a decade to implement, and so even if effective – a big "if" – they wouldn't start to change the number of future mothers until 35 years have passed. Demographic trends exhibit huge momentum, as seen in the population pyramid graph. After the initial drop in fertility in Japan, fertility rates changed only slowly (and the surprises were downward). If the government had really wanted to "solve" the population problem, it should have started 50 years ago. By the 1990s, a change in the policy environment was too late to matter. Now it diverts attention from addressing the consequences of population change, the topic of a later post.
Source: https://en.wikipedia.org/wiki/Marriage_in_Japan
Why Fewer Children
Demographers have found no simple answer for why women choose small families, but here are some factors.
As previous generations left the farm, the need for child labor vanished. In Japan, that occurred by the end of the 1950s. Mandatory education reinforced that. So for the past 70 years children incurred costs but brought no financial benefits. By 1952 fertility returned to pre-WWII levels, and by 1960 was about 2.0, meaning each women had on average one daughter, meaning that population growth would eventually end. Then from the mid-1970s fertility dropped below the replacement level and continued falling. That Japan's population would decline was then locked into place.3
By now late marriage and small families are deeply ingrained. Urbanization and education also changed women's employment options. Self-employment and small family businesses are no longer central to the Japanese economy, so the ability those provide for women to shift back and forth between childcare and work are no longer relevant. Instead, formal work ("labor force participation") becomes central. Almost all Japanese women graduate from high school, and 70% go on to a college or technical school. Having children thus bears the high opportunity cost of lost income, which is accentuated where women want to pursue a career where continuity of employment is important. Daycare can help when jobs are local and don't entail overtime – or where employers are understanding. Grandparents are even more helpful...spoken as one who cared for two granddaughters for 3 months in 2020, as their parents were both essential workers, and their mother was regularly exposed to COVID.
One way to indirectly observe these employment pressures is an "M-chart," named from its shape, because around the world women dropped out of the labor force to raise families and then reentered when children were in middle school.4 The shift in Japan over the past half-century, between 1973 and 2023, is huge, particularly for the age 25-29 bracket. (Note, too, the marked rise in older age brackets – the graph's scale starts at 10%, obscuring that in 2023 one in eight Japanese over the age of 70 were still working, though typically not full-time.
Rearranging the underlying data to focus on the key childbearing ages highlights an additional dynamic. Starting in the mid-1970s, women's labor force participation age 25-29 increased steadily. That didn't happen for women ages 30-34 and 35-39. Then in the 1990s the 30-34 bracket started moving in parallel, with about a decade gap between them. In other words, on average the 20-something working in the late 1980s continued working.5 OK, keep looking: from the mid-2000s, the graphs for women in their 30s overlap. Women who returned to the labor force stayed in the labor force, and the share working rose from 60% to 80%. Work became the norm. But that's not central to demographics, because most Japanese women still complete their childbearing by age 35. Still, it's indicative of a change in mindset. A huge one – a 45 percentage point increase for women ages 24-29 and 40 percentage points for women ages 30-34.6
Finally, some miscellaneous data on marriage. For both women and men, the age of first marriage rose. Japan been a late marriage society for centuries, compared to northern Europe and the US, where in several states marriage at age 13 is still legal. Already in 1900, women on average didn't get married until age 24. It's now age 31, and the age at which women had their first child rose from 24.8 in 1950 to 30.6 in 2022. In addition, only 5% of women born in the early 1950s never married. In 2020, however, 19% of women in their early 40s were still unmarried. Marriage and childbearing are no longer universal.7
Most women who have a child have a second child. Indeed, surveys show that most married women’s desire for two children hasn’t changed over the past 5 decades (as per 2021 [16th iteration of the] Basic Survey on Childbearing 出生動向基本調査). However, with later marriage, the number who ultimately have only 1 child has risen from 1 in 10 to 1 in 5. Since only 1 in 6 women now go on to have a 3rd child, "higher parity" births fail to make up for the women who only have one child, willingly or not. In addition, today 9.9% of women who are past childbearing age never had a child, up from 3.1% of women in the early postwar years. (There are very few teen mothers in Japan, and very few unwed mothers of any age.)
Conclusion
Population decline is not something sudden, it's not a “crisis”. In Japan’s case it was set in motion a half century ago. Despite lots of hand-wringing, it's not amenable to policy, except in small countries that can encourage immigration in meaningful numbers. Even so, immigration has a smaller impact on the age distribution as immigrants tend to retire in their new home. While in Japan there are now 3 million immigrants and temporary residents, the language barrier impedes skilled working-age immigrants in healthcare, engineering and other critical sectors, because such jobs require an ability to read and write Japanese, and virtually no one coming to Japan is already literate.
Let me circle back to the "ordinary" trope. In demographics, Japan has been in the vanguard of change, with fertility falling earlier, and the aggregate population in decline since 2008. But in most of the world outside of sub-Saharan Africa fertility is now below replacement levels, and future population decline is locked into place. Indeed, the overall populations of Korea, China, Russia and parts of Europe are already falling. Japan is typical, not exceptional.
In living memory Japan was a society of large families and a rapidly growing population. Now schools are closing for lack of students, while old age care is strained for resources. In this Japan is now ordinary. I will discuss the economic impact of demographics at some point, but my next couple posts will return to the automotive industry.
Mortality is the biggest source of uncertainty for the population in 2070, not births. The second big source will be immigration.
I don't include data on men, because they don’t really matter for demographic projections. There are more than enough men to go around, as for past half century marriage in Japan hasn’t been universal, so any woman who wanted a husband could find one.
The most commonly used measure is the Total Fertility Rate, a synthetic measure. Wikipedia details TFR nicely.
In my book Competitive Ties (Columbia University Press, 1991), I provide an example of a factory that set its workday to correspond to the hours of the local schools to facilitate women returning to work when their youngest child started school.
This is an exaggeration, because women, so the women who had children early and returned to work offset those who married later and started families later – the marriage age rose, and in Japan single motherhood, and teen-age motherhood, are very unusual.
Note that the labor force includes those engaged in part-time work.
Most data are from the annual statistical compendium of the National Institute of Population and Social Security Research, https://www.ipss.go.jp/syoushika/tohkei/Popular/Popular2024.asp. Few or none of the web pages have English headings. For example, see 表６-４ 性，年齢（５歳階級）別初婚率 on age at first marriage and 表６-２３ 性別50歳時の未婚割合，有配偶割合，死別割合および離別割合 on the share never-married at age 50.
Duh, I read Japanese, including fiction and history. The US has lawyer novels, but (except in translation) that's not a genre in Japan. Instead there's a genre of business(man) novels, almost entirely lacking in English. One series I have read, on the growth of supermarkets, was even been turned into a movie.