Japan’s life expentancy actually LONGER than we thought

Many of you are familiar with the scandal that has rocked Japan where hundreds of octogenarian believed to be living are in fact missing or dead. Adam posted on the event that launched the nationwide investigation several weeks ago, and since then it has mushroomed, with the problem being particularly pronounced in the Kyoto-Osaka-Kobe area. It appears that either because of mere neglect or deliberate pension fraud, relatives never filed a shibou todokede (notification of death) or shissou todokede (notification of a missing person), which Japan’s 19th century family laws rely on to track their elderly.

The joke among friends for the weeks after the scandal broke was that this would lead to the collapse of Japan’s reputation as the kingdom of longevity. Presently, the average life expectancy for women is 86.44 years and the average life expectancy for men is 79.59 years. Surely that number will now have to be readjusted, no?

Apparently not! The Ministry of Labor and Welfare has said that it expects the current events will have very little impact on Japan’s average life expectancy. Why? It turns out that women above age 103 and men above age 98 are excluded from longevity statistics, on grounds that they are statistical outliers. Furthermore, the census conducted every five years is conducted by visiting households, not by consulting the family registry records, so the missing elderly would not be included in that regardless.

Without glossing over the seriousness of the missing elderly, the irony of this fiasco is that it is publicizing the fact that Japan’s average longevity is actually LONGER than we thought.

33 thoughts on “Japan’s life expentancy actually LONGER than we thought”

  1. Surely a few hundred missing centenarians is hardly going to make a perceptible dent on the statistics of 120 million people anyway?

  2. chp wrote: “Surely a few hundred missing centenarians is hardly going to make a perceptible dent on the statistics of 120 million people anyway?”

    True, but with the average age of women at 86.44 and men at 79.59 then it would make more of a difference if people older than that were already dead but still registered since that population much larger. We don’t know because the initial investigation has concentrated on centenarians rather than the very old in general.

    One report noted that, while only 1 in 7 centenarians are men, roughly half those unaccounted for were men. It suggested that it’s much easier for them to drop off the radar screen than women.

    I’ve no reason to think that there would in fact be a significant number of 80-100 year old men and 86-100 year old women who are really already dead but there are probably more than we might have thought a few weeks ago.

  3. I’m with cbp on this. The small number of people who are excluded would be statistically insignificant. At best, I’m guessing the average lifespan would be expanded by days if all were accurately averaged into the totals.

    My main feeling is that the issue is quality of life. I can’t say that I think the quality of life for *anyone* (worldwide) over the age of 75 is stellar, but I can say that the situation in Japan is not encouraging. Most of the older folks end up on very small incomes (the government pension being a mere 60,000 yen a year – and many older women only get that money) and live in nursing homes or as “burdens” on their families. I realize Japanese people generally enjoy good health at older ages relative to other people, but they don’t really have great freedom or range of experiences. Old age communities in the U.S. (as opposed to nursing homes) tend to be much nicer places than the plight of the elderly in Japan.

    Living longer isn’t necessarily the best thing. I’ve had plenty of Japanese folks tell me they’d rather not live past 80 because they fear for their quality of life after a certain point.

  4. Japan is home to over 40,000 people over the age of 100. If several hundreds of those (and perhaps more?) do not exist, and in fact died many decades before, that would have a noticeable impact the average — it could certainly knock Japan out off its perch of world record holder.

  5. Curzon, think about what you’re saying. By the same logic, a jumbo jet full of AIDS patients or crack babies landing at Narita would cause a shift statistically significant enough to knock Japan off course.

    There are over a hundred million people in this country. Rounding down to that number, and using the male/female average life expectancy, you’re talking about 8.25 billion person-years. To bring down the average life expectancy by one year, you would need to knock off 100 million person-years from the whole population, which would require the government to be overcounting something like two or three million people (assuming they were believed to be 100+ but actually lived to be something like 60 or 70).

  6. The Awaji-Kobe earthquake, which killed some 6,000 people, put a very slight dent in Japan’s longevity record for that year. The current tempest in a teapot will not affect it in the slightest. The foreign media are falling over themselves with glee, but without the faintest clue about statistics. I also wonder how may seniors fall through the cracks in the US and other countries, and would be amazed if it is any fewer than in Japan.

  7. Come to think of it, the US may have as many as 20,000,000 undocumented immigrants, the vast majority of whom are no doubt living underclass lives with everything that says about access to quality health care, nutritious food, stress, etc. That’s a significant enough group to tank life expectancy numbers by a bit if they were counted and they very well could be if the amnesty debate ramps up.

  8. Joe, the calculation of national longevity is not a simple tally of all living people. There are a few factors weighed in such as age of currently living people, average death age, etc. If hundreds of people thought to be alive in their 110s have actually been dead since their 70s or 80s, and they now had to be included in the averages and be excluded from the living octogenerians, it could indeed impact the national longevity.

    Incidentally, there are many critics who say that Japan uses warped calculations of its longevity outside the norms of those used in the US and Europe resulting in an unnaturally long longevity — similar criticisms are made about statistics such as Japan’s unemployment rate.

    Wataru: “The foreign media are falling over themselves with glee, but without the faintest clue about statistics. ”

    I read the same line in articles on this topic — 海外メディアの中には「日本の長寿ブランドの崩壊」と報じたところもあったが — but it looks groundless. If you do a search for the news stories on news.google.com, I cannot find any prominent articles that pushes on the theme, gleefully or otherwise.

  9. The Japanese media sources that I have seen pushing the foreign gloating line have focused on Korea and one or two other countries that are within reach of “the record”. There apparently really are some news sources talking about those sneaky Japanese faking their way to the top.

    I’ve seen the unemployment rate criticisms as well (Japan’s rate should apparently double to around 11%) but I’ve also seen suggests that the USAs rate should be doubled to around 22%. Pretty confusing.

  10. Curzon, I do not know how many standard deviations 111 is or 125 is compared to the mean life expectancy, but it would seem that you’d need much more than a few hundred people to impact the average any more than the existing margin of error.

    The interesting thing to me in the article was the whole あいまいさがぬぐえない reason. Is that because they don’t know if the reported persons are actually still alive, or they don’t know that the ages of folks that old are reliable, or what?

  11. We really can’t draw any conclusions at this early date. I assume, perhaps wrongly, that “accounted for” means the local authorities have confirmed their registered centenarians are dead as well as alive (some of the former are foreign residents who left the country) so the “unaccounted for” are not the only category with the potential to reduce the totals.

    I doubt whether any of us who have lived in Japan, and regularly seen so many hale and hearty elderly, expect Japan’s reputation for longevity to be seriously called into question. However, it seems premature to draw a line under it while the Yomiuri writes in it’s most recently translated article:

    “With the authorities of many local governments still continuing confirmation of the centenarians’ whereabouts, indications are that the number of “missing” elderly people will increase further.”

    Virtually any story linked with the elderly in any way is receiving special attention at the moment. There was an account only yesterday of a 76 year old man who died of heatstroke. His apartment had the electricity cut off 10 years ago because the bills went unpaid. It turned out he was sharing the place with his 48 year old unemployed son and their only income was the old man’s pension.

    The son might have had some strong incentives to conceal his father’s death to keep pension payments coming and yet he reported it immediately. It’s only one case but it does raise questions about how prevalent pension fraud might be as a motive.

    The western press doesn’t seem to have been any better or any worse than Japanese media outlets in its coverage of the story. All of the issues I’ve seen in overseas reports were raised first by Japanese reporters. It’s still a major ongoing topic so expecting foreign journalists to put it to bed reminds of the scene in “Naked Gun” when Frank Drebin tells a crowd “Move on, nothing to see here,” while a fireworks factory explodes behind him.

  12. “reminds of the scene in “Naked Gun” when Frank Drebin”

    Am I the only one who thought “Weekend at Bernie’s” when the story first hit?

  13. Nope. I think I was joking that the sequel would now have to be “1,500 Weekends At Bernie’s”

  14. “Now that “The A-Team” has been remade, can “Bernie’s” be far behind?”

    There’s a pretty morbid joke about viral marketing somewhere in there.

  15. 5,000 people in Osaka city (not prefecture) alone aged 120 or older are still registered on their koseki as alive, with the oldest being 152. Mie Prefecture has a person born in 1847, 6 years before the first black ships arrived, who would be 163. If these people had been counted on the longevity statistics, we would expect to see a noticeable impact on the numbers.

  16. I’ll see your 152 year old and raise you the 186 year old man in Yamaguchi prefecture.

    One small positive from this affair it gives people an a opportunity to reflect on history. I’m sure I’m not the only one who started thinking “What was happening in 1824 Japan?”

  17. BTW, does anyone think the problems of missing records, or people missing from the records, or deaths not getting properly reported/recorded, are unique to Japan or more common in Japan than elsewhere?

  18. I think Japan is probably a first-world leader in its inability (unwillingness?) to computerize and network all these various records, which is why you end up with the silliness of city employees going out “to check whether the 150-year-old is dead yet.” But there are certain to be individuals disappearing from the grid all over the world.

  19. They’re going to discover that at least one of these 150 plus year old people is a vampire.

  20. I see Nagasaki has upped the ante with a 200 year old man. I wonder if there is a theoretical upper limit on the oldest person not recorded as dead based on when the current system of records was instituted.

  21. Putting aside the fact that many of the cases that have been unearthed in the past few weeks weren’t actually people thought to be alive by anybody (remember the fraudulent 111-year old who started this whole thing off was officially the “second-oldest man in Japan” at the time — I don’t think any of these 152/186/200-year olds were in the system anywhere)
    I’d like to see someone with some understanding of statistics provide an informed explanation of how (in)significantly these cases could actually affect life span stats. I’m not such a person, but my sense is that some people are wildly underestimating just how huge this problem would have to be to make a dent in national life span figures. My math might be naive, and I know this isn’t quite how life expectancy is calculated, but the way I figure it, if every single person in Japan were to be erroneously recorded as living one year longer, that would boost the life expectancy figure by exactly one year. That’s roughly 127,000,000 person-years of error. A single 127,000,080 year-old man would have the same impact. Or a million 207-year old men or 213-year old women. If you only wanted to raise the life expectancy by 0.01 years, you’d still need roughly 10,000 200ish-year-olds, not 2 or 3 or 100.

  22. @bingobangoboy: Just read the post, and it makes clear why a bunch of improperly-registered (and that is what the large majority are, they are not “missing”) 100-plus-year-olds is not going to have an negative effect on Japan’s longevity statistics: because they were not included in the statistics to begin with!

    “The Ministry of Labor and Welfare has said that it expects the current events will have very little impact on Japan’s average life expectancy. Why? It turns out that women above age 103 and men above age 98 are excluded from longevity statistics, on grounds that they are statistical outliers. Furthermore, the census conducted every five years is conducted by visiting households, not by consulting the family registry records, so the missing elderly would not be included in that regardless.”

  23. LB basically said what I wanted to say, but once more, with feeling, to address BBB and Joe:

    Joe: “Curzon, think about what you’re saying. By the same logic, a jumbo jet full of AIDS patients or crack babies landing at Narita would cause a shift statistically significant enough to knock Japan off course.”

    BBB: “If you only wanted to raise the life expectancy by 0.01 years, you’d still need roughly 10,000 200ish-year-olds, not 2 or 3 or 100.”

    You are both making broad assumptions about how the statistics are calculated, which both sound barmy to me. I also think that by Joe’s logic, the decline in the birth rate by itself would result in an increase in the average longevity.

    I don’t know how a country’s life expectancy is calculated, but I don’t see how anyone reading this, whether you be 15 or 65 years old, should impact the Japan’s life expectancy statistics . In five years, those two readers will be 20 and 70. Should Japan’s life expectancy rates adjust upwards accordingly?

    Or is it a look at the death rate? (About a million people die every year). Or just a look at the age of the elderly population and how many of them survived to live that long?

    I find it further interesting that the death rate of 80 for men and 87 for women is apparently, if you read the fine print, the age to which those people BORN TODAY can expect to live. The rest of us should look at the stats of the year we were born, apparently. Although how that statistic makes any sense I also don’t know.

    It would seem to make sense that the average life expectancy should be those elderly people and see how many of them born the year they were born survived to this day. If people who were believed to be living 110, 120 or 160 year olds, when they in fact died in their 60s, 70s or 80s, and since it now appears that there are indeed thousands of these people, and perhaps many thousands still to be discovered, depending on the statistics used, that could impact Japan’s longevity rate.

  24. The life expectancy stats are indeed calculated by using census data, which is compiled through household surveys, and not through the household registries that seem to be the source of this problem. There is a chance that the elderly found as corpses would have been counted, but not the vast, vast majority of these cases of “vanished” super elderly, many of whom probably died in the war. Basically, relatives would have to lie about a person’s death for them to be included, and so far it looks like we only have two or three cases of that. Curzon presents a fair meta-level stats critique, and it is an interesting thought experiment, but it seems that, if anything, this survey might actually EXTEND Japan’s life expectancy by making people aware that elderly being shucked around between relatives’ houses need to be included in the census. No doubt that a few more corpses will be found, but I have a feeling that there are far more cases where Ichiro doesn’t mention Mama is living with him for 6 months in the census because he can’t wait for Jiro to take her off his hands.

  25. Here’s the English Yomiuri translation which covers some of the points others have already made:


    “…According to an official of the health ministry’s Vital and Health Statistics Division, which compiles the nation’s life expectancy statistics, ‘This issue won’t affect the average life expectancy,’ the official said.

    “To begin with, citizens older than a certain age are excluded from the data used to calculate average life expectancy. Because there are so few people in the oldest age bracket, the life or death of each person within the group would disproportionately affect the overall statistics, a division official said. The cut-off age for inclusion in the data set differs every year, the official said. In 2009, men aged 98 or older and women aged 103 or older were excluded.

    “‘Because the population aged younger than the group [of people excluded from the 2009 data] is so large, there will be little impact on the overall statistics even if it turns out that quite a large number of people in the [excluded] group were dead,’ said another official of the division.

    “Another reason why the nation’s average life expectancy rates won’t be greatly affected is that national census results are the primary data used for calculating the rates. In the national census, which is conducted by the Internal Affairs and Communications Ministry every five years, census takers should, in principle, directly contact and interview a member of every household they are assigned to.

    “On the other hand, basic resident registers–as recent discoveries have highlighted–do not always reflect citizens’ actual status, including even whether they are alive. Local municipalities update the registers according to information submitted by residents, including notifications of change of address and registrations of births and deaths.

    “The health ministry believes many people do not provide the resident register with the most up-to-date information, and therefore regards national census data obtained by interviews as being of higher integrity.

    “Yet the completion rate of the national census has been declining lately, causing concern that the accuracy of its results may be affected. While the non-completion rate was 0.5 percent in 1995, it rose to 1.7 percent in 2000 and then to 4.4 percent in 2005. While keeping intact the present requirement that census takers personally deliver census forms to households, beginning this year the internal affairs ministry plans to allow households to return the forms via mail, in the hope of boosting the completion rate.”

    I’d make a couple of points. Firstly, no-one has yet investigated whether there is a high unrecorded death rate in the age groups which are included in life expectancy calculations. I appreciate that any significant level there may not directly impact conclusions from the census data, for the reasons the Ministry gives, but it may raise questions about the accuracy of that data.

    Japan doesn’t have a stellar reputation for national statistics. There’s an awful lot of data but it isn’t always well compiled or easy to verify. I’ve no doubt that Japanese have one of the longest national life expectancies in the world and may well continue to top the rankings by a handy margin. It is, however, part of the country’s self-image so there’s a risk of false outliers going unquestioned. I’m not suggesting that anyone seriously imagined there was a 200 year old in Nagasaki but there could still be a lack of rigour in the verification of, say, the numbers of people in their 90s.

  26. Go go gadget Wikipedia!

    The short summary of what’s gathered there is as follows. Life expectancy is an aggregated mean of how long people of any given age will live. The usual figure used for cross-country comparison is life expectancy at birth, but it is possible to calculate life expectancy at any age (here is a full matrix for the US).

    Life expectancy at birth is a bit thorny because the most important variable is usually infant mortality. Wikipedia notes the example of ancient Rome, where life expectancy was 25 at birth but 48 at age 5, simply because a huge proportion of deaths occured before age 5. (Same applies to much of the developing world today.) Wikipedia goes on to note that Japan may be overcounting its life expectancy because of the way its statistics treat still births, though the article doesn’t go into specifics.

    The statistics are intended to predict life expectancy for people born at the point in time applicable to the calculation, but in reality they are based on past results with some extrapolating correction for the overall trend of life expectancy rising over time. The decline in the birth rate should not directly affect longevity because the statistics only care about people who are actually born, though there might be indirect effects on the life expectancy for women if they are experiencing fewer childbirths.

    Anyway, it sounds as if the handful of missing centenarians are one of the least relevant populations in the calculation, even if they were counted before. There was such a tiny chance of a Japanese person surviving all the way from 1910 to 2010 to begin with…

  27. Is 230,000 the number of centenarians registered as alive with the welfare office? As we know, census records are entirely independent of other offices.

  28. I’m still a little confused. There are several types of data under discussion here so far. Some media reports say the 230,000 number refers to family register records. There are also local resident register records and these seem to be the basis upon which welfare is claimed and distributed (recipients are a third dataset). Then there’s the census data which is a compilation of general population statistics based on (supposedly) anonymous household surveys.

    By and large, when the government talks about general population statistics, it uses the national census as the basis. When a local administrative department talks about the population under its control it looks at the resident register. Foreign residents are separate matter again. Immigration records are under the jurisdiction of the Ministry of Justice while local alien registration comes under the same administration as resident registration. (The proposed changes will synchronize all this under the MoJ).

    This is where I get a little lost. Some of the earliest names to surface in the hunt for missing centenarians were foreign residents. Where were these names recorded? I suppose they may have been in a family register somewhere but the impression given at the time is that they were on the resident register. More than that, they were being cross-referenced with MoJ data to see whether they had left the country.

    In fact, most of the initial discussion appeared to be about resident registry information. The media talked about how local authorities hadn’t kept on eye on the centenarians in their jurisdiction despite being responsible for welfare payments as well distributing the silver sake cup to those who have just turned 100. That’s surely all referring to resident register data.

    In 1963, the number of new centenarians was 153; in 1986 it topped 1,000 for the first time while in 2002 there were over 10,000 new centenarians. Fiscal year 2008 recorded 19,768 and it was expected that the number last fiscal year would be over 20,000. That data was reported in a Yomiuri article in March 2009 which said the silver sake cup budget was rising so fast, the Ministry would have to reduce the size of the cups.

    Those numbers don’t bear any relation to the 230,000 or so total announced recently. The discussion appears to have moved from a focus on inaccurate resident registrations to inaccurate family registrations. The latter is not an insignificant issue but it’s the former which has fiscal implications.

    The lack of systematic coordination between resident registrations and family registrations is no bad thing for anyone who values their privacy but if there are flaws in both sets of data that would seem to be a breeding ground for identity theft, not just benefit fraud.

    Currently, the missing 230,000 is being described as a family register administrative cock-up. The English Yomiuri called it “a mystery for the ages”. However, I’d like to know what proportion of individuals in that total were also still in basic resident registries and, in principle, entitled to welfare payments. After all, the case of the 111 year old man, which started this affair, is an example of just that. Some cross-referencing must already have taken place so I’m not clear why we are so far only hearing about family register data.

Comments are closed.