Japan's Low Fertility

Earlier this month the Japanese Health Ministry revealed that the Total Fertility Rate (TFR) recorded in Japan rose slightly - to 1.32 babies per woman - in 2006, a figure which is up from the record low ever of 1.26 recorded in 2005.

Japan's fertility rate rose last year for the first time in six years in 2006, while the number of suicides fell below the 30,000-case mark for the first time in four, the government said Wednesday.

``The latest figure alone doesn't indicate whether there is a turnaround in the country's recent trend of falling number of births,'' said Emi Sato of the vital statistic division with the Health Ministry.

Japan's fertility rate was 1.33 in 2001, 1.32 in 2002 and 1.29 in both 2003 and 2004 - the lowest figure since the government began releasing fertility rate data in 1947, according to the ministry.

Last year, the number of births in Japan totaled 1,092,662, exceeding the number of deaths by just 8,174, the report showed. Marriages in Japan totaled 730,973 last year, up 16,708, while divorces totaled 257,484 people, down 4,433.

But before we all start jumping up and down with joy about this, it should be noted that the increase is very small and that the Health Ministry itself is warning that the reading may well be in part a statistical construct, with the TFR possibly once more dropping slightly in 2008.

For those interested in a fuller assessment of the longer term fertility situation in Japan, this article by Toru Suzuki in the March 2006 edition of the Japanese Journal of Population makes a good read:

Fertility Decline and Policy Development in Japan

As Suzuki notes:

The Japanese government was shocked with the TFR of 1.57 in 1989 and launched a variety ofpronatal policy measures.

As he also goes on to suggest the Completed Cohort Fertility Rate (CFR) is a more accurate measure of fertility than the TFR, because the latter suffers from tempo distortion and the so-called "parity composition" effect (for an explanation of some of this we have this post on my Demography Matters site).

Suzuki is none too optimistic that Japanese CFRs are going to rebound all that significantly (ie to anywhere anywhere near replacement level) even in the medium term:

Although the 1955 cohort was behind its predecessor in the early twenties, it succeeded in catch up and will fulfill a near replacement level. However, a significant decline in the CFR for cohorts born after 1960 seems to be inevitable. The cumulative fertility of the 1960 cohort is 1.84 at age 43 and will not reach 1.9 eventually. Though it is difficult to predict the CFR for cohorts born after 1965, the postponement in the early twenties seems too serious to be compensated later. Thus, the CFR of younger cohorts in Japan can be as low as 1.6, which is predicted for Italian cohorts.


Now I think it is important to try and understand the Japanese fertility situation in a more general international context. With this in mind it is worth noting that the Economist has an informative and interesting (if somewhat complacent) article on European fertility this week, and I have a pretty lengthy post on A Fistful of Euros which goes through the main issues raised.

There seem to be two central points here is

a) that there is no such thing as "European Fertility", but rather there are various patterns (the Economist divides Europe in two, I try to be a bit more subtle, and distinguish between four groups of countries)

b) that low European fertility cannot be looked at in isolation, since it forms part of what is effectively a global phenomenon, as birthrates steadily drop in country after country across the planet. The situation in China after many years on a one child policy is by now well known, as is the situation which prevails in Japan. Less well known perhaps is that several states in Southern Indian are now with below replacement fertility (and apparently heading down to the lowest-low bracket) as is, for example, Thailand.

Before going any further I would like to stress yet one more time that I am NOT a demographer, but a macroeconomist, a macroeconomist who has simply become interested in demographic processes due to their evident interface with economics.

I would now like to go on to address what Japan has in common with what is happening elsewhere, knowing full well that in each and every case there are also "country specific" features.

Various explanations for the generalised below-replacement phenomenon have been offered, some of them social and some economic. Undoubtedly there are a number of factors at work, but for present purposes I want to highlight two points:

i) In some form or another the fertility decline is associated with the process of economic development (one immediately might think here of Becker's child investment thesis, and the substitution of quality for quantity), and in particular what are termed the lowest-low levels of fertility (TFRs below 1.3) seem to be associated with very rapid rates of economic development (the Asian tigers and southern Europe might be considered good examples here). Since the big economic news since the late 90s has been the very rapid economic growth which is taking place in a large wedge of emerging economies, there would seem to be prima facie grounds for fearing that this lowest low fertility level may arrive in an increasing number of countries, and comparatively soon.

ii) The growing disconnect between those countries who fall below replacement level, but then steadily recover ground - typical cases here would be the US, France, the UK, Ireland, and Scandinavia generally (these could be increasingly considered the "outlier", special-case countries, although of course they themselves are a pretty heterogeneous bunch) - and the "other group" (ie the increasingly "path normal" countries) where fertility falls below the (apparently) critical TFR 1.5 level, and then subsequently fails to break upwards again (and here of course Japan would seem to be a good example of the phenomenon).

Why this latter process of trawling the bottom is the case is in fact the big headache we all face. The persistence of this below 1.5 TFR phenomenon has lead the Austrian demographer Wolfgang Lutz to formulate a low fertility trap hypothesis (LFTH). The first point to make would be that this trap idea is simply a hypothesis awaiting confirmation at this point. Lutz identifies three mechanisms which might be operating in perpetuating the trap:

a) Negative population momentum
b) Ideational factors
c) Economic processes

On the negative momentum situation, one Japanese newspaper recently made the following point:

".... In the future, there is expected to be a phenomenon in which the number of births declines but the birthrate rises. Because of this, some experts say the number of births is a more appropriate measure than the birthrate for evaluating government policies and for setting targets".

Well, this is negative momentum at work. After decades of below replacement fertility the cohort base is continuously reduced and as a result fewer and fewer children arrive (as shown in this recent post on Germany) regardless of anything other than very sizable (and pretty much inconceivable) movements in the fertility rate. It is hard to overstate the corrosive impact of this process on the population pyramid in the longer term.

This momentum factor on its own does not, however, generate a fertility trap, since it does not affect "fertility" directly. However, it may be a factor which has an indirect impact via the processes characterised as (b) and (c).

(b) Ideational mechanisms: the idea here is simply that in an environment where very few children are actually being born, the woman's idea of "ideal family size" may change. Thus the negative population momentum which produces comparatively few children may impact in this way. The phenomenon being referred to in some Japanese media as "parasite single women who adamantly refuse to marry and bear children" may indeed be one example of this process at work (remember, the two child family is in many countries a very recent phenomenon, and with rather few deep social roots), but if indeed something like this does exist in Japan it is only a culturally specific example of a more general process. Around 25% of German women now remain permanently childless.

Economic mechanisms: Well one example of this would be the impact of structural reforms on seniority bonuses in Japan:

"As age-wage curves flattened in Japan, women can no longer marry to become full-time housewives, as in general their husbands' salaries aren't going to rise enough to make for a comfortable living."

I think a long and pretty obvious list of similar economic determinants which influence fertility decisions could easily be drawn up. For the LFTH to work, however, these need to contain an element of circularity (ie they need to be economic processes which are in part driven by ageing and low fertility in the first place), and I think it is not sufficient to draw attention to a phenomenon with a general impact like globalisation.


My own work as a macroeconomist in part relates to this, and I have been pretty focused of late on internal imbalances in the eurozone, and what (if any) relation these may have with differential rates of population ageing as between countries (and, thus, by implication, differential fertility, since those who are ageing most rapidly are those whose fertility continues stubbornly to remain below the TFR 1.5 level).


I think, if we look at the three most "elderly" societies in terms of median age - Japan, Germany and Italy - we can now begin to discern certain "stylised facts":

a) Ongoing weaknesses in domestic consumer demand
b) absence of housing booms (since 1995 in the European cases)
c) Comparatively high rates of personal saving, which then begin to decline
b) Increasing structural dependence on exports for growth
c) Lack of attractiveness as a destination for migrants
d) An increasingly flat wages curve (across time) despite demographically driven labour market tightening
e) Growing government deficit issues and dilemma's about how to fund
health and pension systems, with a tendency to try and load the cost onto the tax system rather than reducing provision.


Now it is evident that there is some variance between countries in the level of "fit" here. Italy for example, has experienced very low economic growth (even during the recent global spurt) precisely because it has (for political gridlock reasons) been unable to make the kind of reforms we have seen in Germany and Japan, and thus has been able to achieve export lead growth. Unfortunately this is bad, rather than good news for Italy. Likewise Italy has been receiving rather more migrants of late than in the past, but at the same time has been having a large human capital deficit on the migrant flows, since the Italian economy simply is not able at this point to generate the kinds of employment which many educated young Italians need, and hence as unskilled migrants enter educated Italians leave. (This has also been happening to some extent in Germany, and I would be very interested if anyone had information on this vis-a-vis Japan).

The main point about the above list - were future data to confirm these trends - is that (as a complex) they all tend to reinforce birth postponement decisions via their impact on the economic well-being of young people. This is especially true of (e), where young people are, via the tax and contributory systems having a growing burden placed on their shoulders. This tendency could become even more pronounced as a majority of voters come to be over 50.