The provocative title of this post is intended to parallel the provocative title of Robin Lustig’s 2009-09-04 post to his blog on the BBC website, in which he asks, “Is Japan a dying nation?”
Mr Lustig introduces his post by explaining that he’s “on the plane back from Tokyo” after a week of reporting on the Japanese election. This could excuse quite a lot. Going to Japan for the first time is quite a mind-blowing experience for a Brit, especially for someone with interest in the Japanese nation: so many things are the same yet somehow different! or different but somehow familiar! You come back imagining that you have acquired a sophisticated understanding of Japanese society and you find yourself unguardedly retailing cultural generalizations. I’ve done all this myself, so I have a certain amount of sympathy. Nonetheless, were I to type up my naivety for everyone to see on one of the most-visited websites in the UK, then I think I’d deserve every bit of criticism that I’m about to heap on Mr Lustig’s article.
I have five criticisms of Lustig: (1) promoting offensive cultural stereotypes; (2) no links or citations to the evidence he relies on; (3) failing to give the context needed to understand his statistics; (4) promoting a bogus theory for Japan’s low fertility; (5) his general attitude to ageing and fertility.
Here’s Lustig presenting an explanation for Japan’s low fertility rate:
I’m no social psychologist, so I wouldn’t dare to come up with an explanation for why Japanese couples aren’t having enough babies. But one theory is that Japanese women are increasingly reluctant to marry, because they think Japanese men have shown themselves unable to adapt to the needs of a new, more flexible society - and have retreated into a fantasy world of comics, video games and animated pornography where they feel less threatened. [… Japan] is the nation of manga comics and young women who dress up as French maids to pander to the fantasies of lonely men.
(For readers unfamiliar with modern Japanese culture: Lustig is referring here to otaku (obsessive fans, typically male), manga (comics), hentai (pornographic comics and animation), and cosplay restaurants.)
Note Lustig’s introductory caveats: it seems to me that he knows that what he’s going to say is offensive, and he doesn’t want to take the heat for it, so he takes care to distance himself from the explanation, palming it off as “one theory”. Whose theory is it, then?
Maybe if there were compelling evidence for this theory, then it would be fair for Lustig to mention it. But even were that the case (which it isn’t, as we’ll see later), he should have taken more care than this, just on the basic principle of politeness. This isn’t you and your mates at the Radio 4 studios, Mr Lustig. Japan is reading you and is not all that keen on what you have to say (see translations of the comments by Danny Choo).
It’s my opinion that Lustig picked this theory to write about because he finds it plausible and congenial to his prejudices, and because he knows that if he doesn’t give any of his sources, then it’ll probably be too much work for his readers to track them down and refute him.
This isn’t a particular fault of Robin Lustig, it’s a general failing of the whole profession of journalism. But it annoys me more each time I find myself spending time trying to track down the original sources behind some absurd claim in the media.
Mr Lustig, this is 2009. Close to 40% of 18-year-olds go to university, and many of them probably can’t remember the world without the web. Your readers are among the most highly educated people in the country. We know how to follow links and are as able as you are to read and understand original sources. So why persist in this absurd pretence that you’re the sole intermediary between scholars and a mass audience? You’re not fooling anyone any more. And maybe if you were conscientious about providing sources and references, then you’d have to think twice before parading your prejudices in public, which would be a win for everyone.
Let’s take a few of Lustig’s statistics and see if I can find some sources for them and add some context.
A country that knows it is ageing more rapidly than any other major industrialised nation on earth.
The caveats “major” and “industrialised” suggest that Lustig suspects that this statistic isn’t as clear-cut as all that. And what does it mean to “age rapidly” anyway? Let’s take “speed of ageing” to refer to the change in the fraction of the population over 65 between 1999 and 2009. I don’t know for sure what Lustig means by a “major” nation or an “industrialised” one, but I’m going to assume that “major” means “population of 10 million or more” and “industrialised” means “GDP per capita (purchasing parity power) of $10,000 or more”.
This table gives the top ten major industrialised countries by speed of ageing (plus the UK), in roman text. Non-major or non-industrialised countries that would otherwise appear in the top ten are included but given in italics.
|% population 65 and over|
Data from the CIA World Factbook for 2009 and 1999.
So no qualifiers were in fact necessary, by the measure I chose. Maybe Lustig had a different measure in mind? If only he had said, or given his source, we might be able to tell.
Which has the highest proportion in the world of people over the age of 65.
I think that when Lustig says “over the age of 65” he means “aged 65 and over” since the latter is a standard demographic category. The phrasing he uses is ambiguous between “65 and over” and “66 and over” which is why demographers avoid it. But this is an easy mistake for a non-specialist to make.
Using the data from the CIA World Factbook 2009, here are the top twelve countries by proportion of population 65 and over (plus the UK):
|Country||% population 65 and over (est. 2009)|
Also from the CIA World Factbook, the bottom ten countries by total fertility rate (plus the UK).
|Country||total fertility rate (est. 2009)|
|Northern Mariana Islands||1.15|
|Bosnia and Herzegovina||1.25|
on current trends, the population of Japan will have halved by the end of the century.
I’m pretty sure that Lustig is referring to the population projections made by the National Institute of Population and Social Security Research, which publishes 100-year projections on a five-yearly basis. The most recent is Kaneko et al. (2008), “Population Projections for Japan: 2006–2055”, Japanese Journal of Population 6:1. The title of the paper is a bit misleading because they include a long range (100-year) projection as well as a medium-range (50-year) projection.
Here’s figure A-1 from Kaneko et al., page 114, showing 50- and 100-year projections for nine scenarios using different predictions for fertility and mortality.
To get a bit more context for these projections, let’s compare with similar projections for Germany and the UK.
For Germany, I’ve chosen Eisenmenger et al., “Germany’s population by 2050: Results of the 11th coordinated population projection”, Federal Statistical Office.
This set of projections only goes to 2050, but the pattern up to that date is similar to Japan (a fall to 1950s levels by 2050). I didn’t find a longer projection, but I imagine that with suitable choice of fertility scenario, we could easily say, “on current trends, the population of Germany will have halved by the end of the century”.
For the UK, I’ve chosen Helen Bray (ed.) “National Population Projections: 2006–based”, Office for National Statistics, PP2:26, which projects the population to 2081. This is figure 9.6 from page 56, “Population of the United Kingdom according to principal and variant 2006-based projections, 1981–2081”.
Using this figure, we could write the opposite headline, “on current trends, the population of the UK will double by the end of the century!” But how much credence can we really put in these kinds of long-range projections? The problem is that population projection as it’s currently done is a process that iteratively applies fertility, mortality and migration rates to a population. This is a process that exponentially enlarges the differences between the scenario and reality. So you have to interpret long-range forecasts as saying, “if things go on like this,” while remembering that things won’t go on like that after all.
A salutary reminder of the difficulty of long-range population projection comes from Chris Shaw (2007), “Fifty years of United Kingdom national population projections: how accurate have they been?”, Population Trends 128, Office for National Statistics. His figure 1 on page 10 compares the actual population with six projections made between 1955 and 2004.
So, Lustig is accurate in his description of Japan as an ageing society. It’s a country that’s at the extreme on several demographic measures.
But it’s not a total outlier: Germany, Italy, and Greece are not far behind on proportion aged 65 and over. Germany in particular is (on current trends) looking at a similar population trajectory to Japan. Several countries have even lower fertility, and many European countries have total fertility rates in the range 1.2–1.4. This ought to give a careful writer pause: if several countries have low fertility rates and high (and increasing) proportions of people aged 65 and over, then we need to look to common explanatory factors before looking at cultural specifics.
Demographers refer to the change in population structure experienced by Japan as the “demographic transition”. It’s caused by falls in the rates of death and birth. The fall in death rates results from improvements in public health, hygiene, nutrition, and vaccination for childhood diseases. The fall in birth rates is due to improvements in the rights and social status of women, and the availability of cheap and effective contraception. Because the fall in birth rates happens later than the fall in death rates, there’s a period of rapid growth in population, which leaves a “bulge” in the population pyramid, an ageing society as this bulge reaches old age, and a declining population as members of the bulge eventually die.
However, since Japan is at the extreme end of the ageing continuum, there must be specific causes at work which serve to exaggerate the ageing trend. Maybe anime, otaku, and maid cafés do have a role to play here?
Here’s a plot of crude (per 1,000) birth and death rates for Japan from 1872 to 2008 (excepting 1944–1946, for which period there is no data). This could be a textbook figure for a demography primer.
(The sudden drop in birth rates in 1966, and its immediate recovery in 1967, is striking. According to Robert William Hodge and Naohiro Ogawa, Fertility change in contemporary Japan, “In Japanese culture there is a traditional belief about the Year of the Fire Horse [丙午], which occurs once every 60 years. According to this long-standing superstition, a female born in a Year of the Fire Horse is destined to both an unhappy life and killing her husband if she marries … many couples time their births so that they will not occur in such a year.” I’ll note in passing that unlike Lustig’s theory, this is an example of a specific cultural explanation for a change in fertility for which there is good evidence.)
The plot of birth rates shows two periods of sharp falls in fertility: the first from 1945–1960, the second from 1975–1990. It seems to me that if we’re going to look for specific cultural factors, we ought to find them in these periods, not in more recent fads like cosplay restaurants (2000s).
Hodge and Ogawa, in Fertility change in contemporary Japan, consider the distinctive features of Japan’s demographic transition from the perspective of 1991. From their foreword:
The present [i.e. 1991] Japanese situation is very similar to that of the West, where fertility rates in the range of 1.4 to 1.8 are common. All these countries, like Japan, share the experience of rapid population aging—an inevitable by-product of very low fertility. […] Immigration, which offsets to some extent the negative natural increase in West Germany and many other Western countries, does not enter the picture in Japan, which has strong barriers against it.
Japan, like the Western countries, has achieved its very low fertility levels partly through later-age marriages but mainly through the increasing use of birth control measures—both contraception and abortion. […]
While Japan resembles the West in its general demographic situation, there have been and are some important differences. Legal abortion has played a distinctively important part in Japan’s fertility decline, especially during the decades of the 1950s and 1960s. […]
Japan is also unique in terms of the use of contraceptives. The condom is the most prevalent method, with rhythm in second place. […] It is remarkable that the Japanese have maintained such low fertility for so many years despite the use of contraceptive methods that require exceptional care and discipline. […]
Hodge and Ogawa go on to analyze other social factors: the “patriarchal extended-family system”, arranged marriages, the obligation to take care of parents in old age, education, urbanization, migration, and family size. None of the factors that Lustig mentions appear in the book.
For a slightly more recent look at this question, see Makoto Atoh (2001), “Very Low Fertility in Japan and Value Change Hypotheses”, Review of Population and Social Policy 10, pp. 1–21. Atoh considers the question of whether “secular individuation” (that is, lack of conformation to traditional religious values) is responsible for Japan’s low fertility. I suppose that by a very long stretch of concepts, you might consider this to be a proxy for the factors in Lustig’s theory. But this would be no help to Lustig in any case, because Atoh concludes:
According to various nationally representative time-series and comparable attitudinal surveys which have been undertaken in the post-war period by various institutes, there has hardly been any dramatic change in attitude toward religion and only a moderate change from social conformism toward individualistic attitude over the last 40 years. In contrast, there has been a tremendous attitudinal change related to women’s social and family roles, in such areas as premarital sex, divorce, gender-role division, and the care of elderly parents, especially since the middle of the 1980s. All these survey results suggest that the rapid rise in the proportion never married in Japan in this latest decade can be related to the change in the value system regarding women’s social and familial role and status, a change toward the valuation of a gender equal society, rather than to secular individuation or the end of a child-centered society.
An even more recent study is Naohiro Ogawa, Robert D. Retherford, Rikiya Matsukura (2006), “The Emergence of Very Low Fertility in Japan: Changing Mechanisms and Policy Responses”.
Most of the decline in the [total fertility rate] between 1973 and the present occurred because of later marriage and less marriage. (In this regard it should be noted that only about 2% of births occur out of wedlock in Japan.) […] the main reasons for later marriage and less marriage after 1973 in Japan are the following:
- Remarkable educational gains by women. The proportion of women of the relevant age enrolled into tertiary education increased from 5% in 1955 to 50% in 2005.
- Massive increases in the proportion of women who work for pay outside the home. At present, about 99% of women work before marriage, and almost all of them are in paid employment, so that they have no financial reason for getting married.
- A huge decline in the proportion of marriages that are arranged, i.e., from 63% in 1955 to 2% in 2002. Now, people have to find their own spouses, which is not so easy in Japan because the marriage market is not well-developed.
- A major decline in the proportion of young couples who coreside with parents when they get married, i.e., from 64% in 1955 to 29% in 2002. Young couples increasingly do not want to live with parents, and the decline in coresidence makes it financially more difficult to get married and set up a household.
- A major increase in premarital sex, thus implying that one does not have to get married to be in a sexual relationship. Between 1990 and 2004 the proportion of single women aged 20 and over who reported that they were currently using contraception rose from 39 to 57%.
To fully understand the baby bust since 1973, we should examine not only changes in the marriage component but also in the marital fertility component. To a considerable extent, marital fertility declined after 1973. […] The main reasons for the decline […] are:
- the rise in direct costs of children (higher education is a major component of rising direct costs, which involve a substitution of quality for quantity of children),
- the rise in opportunity costs of children for women (in terms of lost income as a result of temporarily dropping out of the labor force),
- the shift in preferences away from children toward “other goods” (in other words, the “consumption utility” of children has fallen),
- the fact that families are less secure,
- the reality that women want more help from husbands in childrearing and housework
(I renumbered the reasons so I can refer to them uniquely.) The only one of these that seems to have any bearing on Lustig’s theory is reason 3: it might just about be the case that one of the obstacles to marriage in Japan is that young men are more interested in manga, anime, video games, etc. than they are in dating. But reason 4 seems to directly contradict this possibility.
In summary, I can find no evidence whatsoever for Lustig’s theory, and what evidence I can find seems to me to amount to a pretty compelling refutation.
In this section, I’m going to switch from facts to opinions, and consider some of Lustig’s implied attitudes.
First, the implication that ageing is primarily a problem:
According to one United Nations estimate, it’ll need to import 17 million foreign workers over the next 40 years, just to keep its economy afloat and provide enough carers to look after the elderly. (By 2050, there will be more than a million Japanese over the age of 100.)
I wish that we had such problems in the UK: I’d love to have as good a chance of surviving to 100 as an average Japanese man of my age.
Second, the lack of any kind of appreciation that there are limits on growth, and ageing and population decline is something that every country in the world is going to have to learn to cope with if it wants to avoid environmental disaster. Japan’s extreme position on the continuum of ageing makes it a society that we could learn from, if we weren’t blinded by our prejudices.
Thirdly, the implication that couples somehow owe it to society to have “enough” babies. I disagree completely with this: if an increased birth rate is important to society, then society is going to have to pay for it. (Something that the Japanese government does in fact recognize, though £170 per month children’s allowance seems pretty feeble when set against factors 6–10 from Ogawa et al. that I quoted above.)
Lustig mentions the science-fiction film Children of Men as a dramatization of the effects of infertility. I’ll quote from Lois Bujold’s novel Ethan of Athos instead. On the remote planet of Athos, artificial uteruses permit the existence of an all-male society. Visiting the rest of the galaxy, Ethan learns about the mores of other worlds from Elli Quinn:
“Maybe they meant to raise battalions of mutant super-soldiers in vats like you Athosians and take over the universe or something”
“Not likely,” remarked Ethan. “Not battalions, anyway.”
“Why not? Why not clone as many as you want, once you’ve made the mold?”
“Oh, certainly, you could produce enormous quantities of infants—although it would take enormous resources to do so. Highly trained techs, as well as equipment and supplies. But don’t you see, that’s just the beginning. It’s nothing compared to what it takes to raise a child. Why, on Athos it absorbs most of the planet’s economic resources. Food of course—housing—education, clothing, medical care—it takes nearly all our efforts just to maintain population replacement, let alone to increase. No government could possibly afford to raise such a specialized, non-productive army.”
Elli Quinn quirked an eyebrow. “How odd. On other worlds, people seem to come in floods, and they’re not necessarily impoverished, either.”
Ethan, diverted, said, “Really? I don’t see how that can be. Why, the labor costs alone of bringing a child to maturity are astronomical. There must be something wrong with your accounting.”
Her eyes screwed up in an expression of sudden ironic insight. “Ah, but on other worlds the labor costs aren’t added in. They’re counted as free.”
Thanks to Alex Selby for badgering me into refining this argument, and for pointing me at Ogawa et al. (2006).