Japan's Slow Disappearance

In the rare event you haven't heard, a curious phenomenon arises as a people becomes more affluent and perceives no serious external threats to itself. The birthrate drops. There are numerous reasons for this, some that apply to some peoples and not others, some that apply to a portion of a people but not all of its constituents, etc. Japan is no exception. In fact, Japan has one of the lowest total fertility rates (TFRs) among the most afluent countries. At the time of this writing, it is 1.14 births per woman over the woman's childbearing years. As a reminder, 2.1 births is called the replacement level, the number of births required to replace just the two parents (the ".1" accounts for child mortality).

For reference, here are the official statistics for 2025 from the Ministry of Health, Labour, and Welfare, and here are detailed projections, explanations, and other related statistics presented in English. Some key takeaways besides the TFR are the number of births (approximately 670,000 people), the number of deaths (approximately 1,590,000), and the natural population change (approximately -918,000). Japan currently has a population of about 123,000,000, meaning that, ignoring immigration, Japan lost about 0.75% of its total population in roughly the last year.

Detailed mathematical models of the situation and thorough investigations into the causes are for another time. I do want to spend a moment with a toy model, though. I didn't realize how dire the situation was until I understood that even if Japan's TFR returned to replacement level in the next 5 years, there would still be large social consequences.

For the toy model, let's assume we have a population of 100 people. Twenty newborns, twenty 20-year-olds, twenty 40-year-olds, twenty 60-year-olds, and twenty 80-year-olds. Let's assume that everyone over 80 years old dies before or when they reach 100 years old, that only women 20 to 40 years old bear children, that if women bear children they do so exactly when they reach 20 years old, that the population is split 50/50 male-female, and finally that no one dies until they reach 81 years or older, meaning the replacement level is a simple 2.0. For simplicity, everyone is exactly the same age in their age group.

A generation is 20 years, the approximate amount of time required for newborns to grow into productive members of society and enter their childbearing window. If all women who are 20 years-old magically immediately had 2 kids and then stopped having kids, the replacement level is reached, and the population is at equilibrium.

Japan's current TFR is about half of the replacement level, so for our toy model, let's say that's 1.0. Over 1 generation, 20 people die and only 10 are born (one for every 20-year-old woman), meaning our original population of 100 is now 90 people. In the next generation, we have only 10 candidates for childbearing of which only 5 are women. 20 more people die, and the ten 20-year-olds have 5 children. The population is now only 75 people. Play the trend out with this TFR, and the toy society ends up extinct.

Let's look at what one approach to population recovery requires in the case a single generation in the toy model has a TFR of 1.0. Generation 0 is the base case, and generation 1 is the one resulting from a TFR of 1.0. In generation 1, the 5 women need to have a TFR of 4.0 to produce 20 children, to match the expected death rate. From that point on, we set the TFR of each generation to the replacement level.

One key takeaway from this table is that a single generation with low TFR impacts the population of a society across many generations assuming a radically-high (for first-world countries) TFR (in this example, 4) is not maintained. A more natural recovery would have a TFR of just above replacement level for many generations and would thus take more time to play out.

Similar to in the toy model, if Japan wanted to stop population decline immediately, this generation needs a TFR far above replacement level and then consecutive generations need to at least maintain replacement level indefinitely. For reasons I promised are for investigating another time, it is extremely unlikely Japan would return to even replacement-level TFR in a single generation, especially if it sustains its current level of affluence and national security. So if Japan were to reach replacement-level TFR, it will do so over generations, meaning that the population will decline until then and also for some generations after.

Japan's total population is much larger than our toy model, people are distributed across the age spectrum, and immigration reduces overall population decline, so the effects of Japan's low TFR are not as immediately obvious in society across just one generation.

What does accelerating population decline feel like for people living in these societies? It feels like nothing at first, until you're in need of something and can't figure out why it's not as readily available as it used to be. It feels like a toothache without a dentist, a leaky water main without a plumber, a broken train with no mechanic, a farm with no farmer, a malady without a doctor, a police force with no police officers, an army with no soldiers, a school with no teachers, a child with no daycare, an elder with no children or caretaker. And this is just staffing. Stocking of goods will be affected, too.

Most of these negative experiences are somewhat alleviated by technology, but we still need to build most of that technology and then operate it.

Japan's population pyramid is becoming inverted as older people increasingly outnumber younger people. After years of population decline, the positive effects of a birthrate trend reversal would not be felt until at least one generation later, when the generation has entered the workforce to help society. However, the strain on the workforce across generations increases as more responsibility is placed on fewer people. Discounting technological advancement, productivity is negatively affected, and the economy in turn suffers, making it more economically infeasible to raise families.

Population decline is cited as a major problem facing Japan, and it's true, but population decline itself is a symptom of other causes. Addressing population decline without, for example, paving over it with immigrants requires addressing the underlying causes. Breaking those down is for another time, though. In the meantime, here are a couple facts that surprised me.

The first is the average age of agricultural workers in Japan. Currently it is 69 years old. You can find statistics on the matter by searching for the Agricultural Census in official government reports. Farming nowadays is highly automated and shifting from family farms to industrial farming, but you still need humans in the loop. This year, the number of agricultural workers in Japan dropped below 1 million for the first time since comparable tracking began in 1985 as fewer people are interested in the general profession. The percentage of land used for farming in Japan also decreases as the number of farmers decreases, although it is not exclusively caused by a shrinking workforce.

The second, on the topic of TFR, is that of Tokyo. Tokyo is amazing in that it is a societal black hole of sorts. People move here from all across Japan for more opportunity, but Tokyo's TFR is less than 1.0. In fact, it is the lowest across all the Japanese prefectures. So the irony of moving to Tokyo to pursue more opportunity is that doing so accelerates Japan's erasure.

What it Means to be Japanese

That Japan is slowly disappearing doesn't sit well with me. "Why do I, as an American, care so much about Japan," is probably the most sensible question to ask first. I have a reason, but that too is a topic for another post.

After looking into the symptoms, some of which I alluded to above, I started investigating the causes. While I was doing so, a question popped into my mind: what are we saving, here? Which lead to more questions. What happens if Japan as a society disappears? Specifically what goes away? I certainly have ideas, and I can list general concepts from Japanese culture and traditions, but I realized that I couldn't explain how an arbitrary group of people from anywhere else could follow some recipe and recreate the Japanese experience.

It's not about the Obon festival, what it means, and how to participate. It's not about the diet. It's not about the language, the history, the procedural business practices, the in group and out group, the arts. There is something else there that is emergent from the combination of all of these things.

It's Japanese people. That's trite to say. But what does it mean to be Japanese? That turned out to be a hard question for me to answer.

Yoshio Sugimoto-san, in the book An Introduction to Japanese Society, conveniently outlines seven aspects of Japanese identity for us. The aspects are nuanced, so I'll just quote him.

One may consider at least seven aspects of Japaneseness -- nationality, ethnic lineage, language competence, place of birth, current residence, subjective identity, and level of cultural literacy [...]. While each dimension is in some respect problematic, the main objective of the [list] is to show that there can be a number of measures of Japaneseness and multiple kinds of Japanese. In combining the presence or absence of their attributes, one can analytically classify the Japanese into numerous types.

(Chapter 7, "'Japaneseness', Ethnicity, and Minority Groups," Part 2, "Deconstructing the 'Japanese'"

In the context of the book, Sugimoto-san is describing how Japanese minorities (the Ainu, Korean Japanese, etc.) are perceived by "fully Japanese" people and fit into the modern Japanese identity. Sugimoto-san uses the aspects to establish a spectrum of Japaneseness, but I'm going to pull an apples-to-oranges here and repurpose those aspects for the sake of my own line of reasoning.

I developed a thought experiment that is in some ways unsatisfying to me but in other ways the best criteria I've come up with to understand what it means to be Japanese. It goes as follows.

Imagine that all people who identify as Japanese magically disappeared from the face of the planet. Immediately, you are given the task to restore Japan to what it was just a moment ago. You are shown the group of people who identified as Japanese and asked to pick 1,000,000 of them to respawn into Japan. You are assured that 1,000,000 people are magically sufficient to carry out all previous, unique duties done in society. And somehow no human will be aware of the rapture-like experience they just had, nor of anyone they used to know now being entirely gone. Finally, you can somehow pick all 1,000,000 people very speedily.

Who do you pick?

Perhaps before answering we should consider how we'd measure success. What I settled on was also somewhat unsatisfying to me, but perhaps that's inevitable. If I could build a criteria that allows us to score an individual's Japaneseness with no relation to anyone else, there would be no point to this thought experiment.

The first smoke test would be to watch and see if any of the selected 1,000,000 go about their business after reappearing and turn to someone around them and wonder, "how did a coworker like this get hired here?" Or vice versa, the individual thinking, "what on Earth am I doing in this place?" Less telling but perhaps still useful is having tourists fly in and seeing whether on their way out they're saying, "yup, just like I remembered," or, "huh, I don't recall Tibetan chants and grilled cheese being so popular in Hakone..."

I encourage you to mull over the thought experiment for a bit. I'll cut to the chase regarding my conclusions.

The outcome of my thought experiment is that even if it is possible to be "fully Japanese" as the last "true" Japanese person alive in the universe, there would be no one else left to confirm this. This is the part I'm most unsatisfied with but can't reason my way out of and therefore (ultimately gladly) accept in lieu of any better explanation, that an additional aspect of Japanese identity is your fit in Japanese society. I can't say it's your fitness as judged by another individual, because as Sugimoto-san established, there is variance among individuals' judgment of whether someone is "fully Japanese." So we have to take an average. And that average is called Japanese society. You can probaly smell that my conclusion is going to be somewhat circular, a chicken-and-egg problem, suggesting that in order to build a Japanese society you must select people that can build a Japanese society.

By the way, I'm just going to ignore the "subjective identity" aspect. As someone who has lived and worked in Japan for about 7 years now and has practiced the language and studied the culture seriously for just as long, no amount of imagining myself as Japanese ever resulted in me actually being perceived as Japanese. Also, "current residence" is literally your instantaneous residence. So if you were born and raised in Japan for 40 years but just moved to New Zealand, your Japaneseness is slowly decreasing. If the requirement is that you're part of the zeitgeist, then we can't call Tokugawa Ieyasu-san Japanese. We have to call him "Japanese 400 years ago," which even the Japanese don't do. Also, could you imagine looking most of these guys in the face and telling them they aren't fully Japanese?

Back to who we'd pick. Actually, let's talk about who we won't pick, counterexamples. There are people who are ethnically Japanese, who speak Japanese, who were born in Japan, who have Japanese citizenship, who identify as Japanese, and who were raised their entire life outside of Japan. If you find and take 1,000,000 of these people and drop them into Japan in our thought experiment, the Japan you get does not operate like the Japan that exists today. We can permute the details of these people this way and that, and I hold that you still would not get the Japan as you know it today.

On the other hand, there are people who were born in Japan, speak Japanese, have citizenship, identify as Japanese, were raised their entire life in Japan, and who are ethnically not Japanese. If you found and picked 1,000,000 of these people, I think you'd be closer in reproducing Japan as it was, but you'd be missing the very obvious details of phenotype and more subtle commonalities in genotype. This is weird (unexpected) to us in the same way that José López, the manager of the local cervecería, being 100% Han Chinese would be weird. Not forbidden by any laws of physics, but not a story we'd ever expect to hear. By the way, by current Japanese law, the people in this counterexample group became citizens through naturalization, were not considered citizens at birth. Japan is jus sanguinis (right of blood) whereas countries such as the US are jus soli (right of the soil).

Japanese as an ethnicity is a quantifiable things. Japanese as an identity is a political and social construct. So people from both groups are certainly Japanese, if we conflate the term, which I have mostly been doing here. In America, we call members of the first counterexample group Japanese Americans. For the second group, there are plenty of documentaries and literature chronicling their journies of reconciling being Japanese in every way but ethnically. In some ways, the first counterexample group is more Japanese than the second. In other ways, the second is more Japanese than the first. But curiously there is some missing dynamic in both groups that prevents us from recreating modern Japanese society from just those people.

Finally getting to who I would pick. As noted, I'd drop "subjective identity." I'd also swap "current residence" with "have lived abroad." Then, boringly, I'd pick 930,000 (93% of the 1 million) people who score perfectly on the (now) 6 aspects of identity, where scoring perfectly on "have lived abroad" means answering "no." Although I have the chance to pick a perfectly cylindrical age distribution (meaning even across all ages, not a pyramid or any other shape), I'd probably replicate the current age distribution--more on that later. I'd then pick 10,000 (1% of the 1 million) who score similarly but have lived abroad. I couldn't find sources I consider authoritative, but a web search suggests that just over 1% of current citizens of Japan live abroad. Next, I'd permute the scores for nationality and ethnic lineage to accommodate the Japanese minorities (Ainu, Korean Japanese, etc.) and then select 30,000 (3% of the 1 million) people. Finally, I'd permute the scores more radically to accommodate the foreigner population living in Japan and would select the remaining 30,000 people.

"Okay, so did you just run a whole thought experiment to ultimately describe the statistical distribution across the existing Japanese population?"

It looks that way, but it took me some iterations. I first just ignored those living abroad and foreigners. But then I realized the one confounding thing that we cannot test for, the one thing that stops us from being able to perfectly quantify what it means to be Japanese. The perhaps unanswerable question. Would this Japanese society formed entirely by people who met the criteria of the 930,000 and 30,000 minorities I originally picked be as close as possible to Japanese society today? I can't prove it, but I think not. Which implies that part of what it means to be Japanese exists in the interplay between the varieties of Japanese identity and even foreigners living in Japan.

Japan has already had to reconcile something like this in the past. The Yamato line, of which 97% of Japanese are descended, was originally said to have divine origins. Meaning the old criteria to be Japanese was simple: possess citizenship, speak native Japanese, be born to two ethnically-Yamato parents. With modern archaeological and genetic evidence, we now understand that the Yamato themselves were a genetic combination of the island-native Jomon people and waves of migrants from the mainland arriving in the Yayoi and Kofun periods.

It is possible that the Japanese-living-abroad and foreigner groups I built do not contribute to being Japanese but instead are just character foils that expose my grossly-generalized idea of the collective Japanese psyche and unconscious. It is also possible, very likely I think, that current population demographics affect the Japanese psyche which in turn affects what it means to be Japanese. For example, if most Japanese families across, say, 300 years had 2.5 kids, there would likely be some social stigma if you're that one family without 2.5 kids. (Instead, in the past, you were just called a Christmas cake if you were a woman and weren't married by 25.) This (demographics, not the cake) is why I would align the ages in my selected 1,000,000 people to current Japanese demographics.

Whatever it may be, part of what it means to be Japanese is to participate in the collective psyche and unconscious and to perpetuate it. It will evolve over time, obviously will not be perpetuated as-is indefinitely. So one doesn't need to accept it, just understand it and one's relationship to it, instinctively, from having been raised in it.

So what are we preserving here, by addressing population collapse? We're preserving the ability to perpetuate what it means to be Japanese, from which comes Japanese society. Japanese and foreigners alike look back through the window of society to understand what it means to be Japanese. Some see parts of themselves, maybe even their whole self, in the reflection.