Explanations for fertility trends in Japan since 1973

Fertility in Japan has been declining more or less steadily since 1973, with minor upward movement. The total fertility rate started in 1973 at 2.135, just slightly above replacement. After gradual decline, it reached an all time low of 1.254 in 2005. It has since rebounded somewhat to 1.361 in 2009, but it is unclear if the rebound will continue, or whether it has reached a relatively stable low. A similar decline can be seen in completed cohort fertility data starting around the 1947 birth cohort and going till the 1969 birth cohort (CCF is available only till 1959, but the approximate measure CCF40 is available for birth cohorts till 1969): CCF40 declined from 2.01 in 1947 to 1.444 in 1969.

For more on curve-fitting the total fertility rate, see fertility in Japan: curve-fitting the total fertility rate.

This page discusses potential explanations for both the absolute low magnitude of fertility and its decline, evaluating them in light of the evidence specific to Japan.

Abortion

See relevant raw data on abortion and fertility at abortion in Japan

The availability and cost of abortion are insufficient to explain fertility decline since 1973. It is possible that tightening abortion somewhat will increase fertility by around 10-20% (at most), but that would still not bring it anywhere near replacement. The following observations are relevant:

Abortion has been legal in Japan since 1948. While it seems to have had a huge effect on fertility at the time of its introduction (though we don't have reliable data since pre-1948 fertility data is not high-quality), it is inconsistent both with the data and with common sense that it would exert an impact on the trend since 1973.
Both the abortion ratio (abortions per 1000 live births) and abortion rate (abortions per 1000 women of childbearing age) peaked in the late 1950s (around 1955-1958) and have since been in steady decline. A quick sense of magnitude: the peak abortion ratio was 716.4 (1957), the ratio in 1973 was 334.9, and the ratio in 2010 was 198.5. The peak abortion rate was 55.5 (1955), the value in 1973 was 25.6, and the value in 2003 was 13 (data for other years hasn't been computed by the data source, though it could be computed in principle). In other words, pregnancies aren't being substituted by abortions.
If anything, abortion laws have become marginally tighter (via the 1996 amendment to the Eugenic Protection Law). These changes to the laws had no discernible effects on abortion or fertility trends.
If we assume that every abortion were substituted by a birth, and that the age distribution of abortions is same as that of pregnancies, the fertility rate adjusting for abortions declined from about 2.85 in 1973 to about 1.65 in 2009. This is an even sharper decline than what actually occurred.

Contraception

Contraception fails as a convincing explanation either for the level or the trend in fertility:

Condoms, the main contraceptive technology used in Japan, have been available since well before 1973.
Birth control pills, which have been assigned the credit/blame for reducing fertility significantly in the West, became properly legal only in 1999. The legal availability of these did not have an impact on either abortion or fertility trends. This NBC article documents the reluctance of the Japanese to use birth control pills.

Marriage and other arrangements conducive to childrearing

Marriage

The conjunction of these three reasons can provide a partial explanation of declining fertility. Note that some of the trends may have reversed in recent years:

Fewer people are getting married: The age of first marriage is increasing, and the proportion of ever-married people is declining.
Marriages are less stable: A larger fraction of marriages end in divorce, and the duration of stable marriage is declining.
Marriage is a de facto precondition for having children: People are highly reluctant to bearing and rearing children outside marriage.

Causality is unclear: it is likely that a large part of the decline in marriage arises from the fact that having children is considered one goal of marriage, and people's reduced desire for children makes marriage correspondingly less appealing. Some of the causation probably runs in the direction from the decline of marriage to the decline in fertility. In particular, the following hypotheses are consistent with this direction of causation:

A decline in the traditional custom of arranged marriages.
The absence of any corresponding rise in love marriages, i.e., people are not finding romantic partners themselves at a rate that would make up for the decline in arranged marriages.
A declining overall interest in romance, with sex toys and video games substituting for real-world romance for many males, and travel-based lifestyles substituting for romance for females.

Inflexible family structure: women in the workplace

The following mechanisms have been proposed:

Females in Japan are highly educated and the workplace has many openings for young single females. This is a change relative to 30-40 years ago.
In comparison, the career options for married women, particularly married women with children, are poor. This is due to a combination of explicit discrimination against hiring such women, and the absence of flexible work policies (such as work-from-home, flexible hours, leave for childcare). Workers are expected to work the standard "male" 50-hour week. In addition, men do not help around the house enough, but have high expectations of good housework, requiring women to work more in the house.

Note that the absence of options for mothers in the workforce isn't new -- what's new is the presence of options for single females.

In particular, this means that:

Career-oriented females find it unattractive to marry and have children.
Men face a greater financial burden from marriage and childrearing than if women could also earn: they have to support two adults and however many children they have. This makes marriage and childrearing unattractive for males as well. (Note that this isn't a factor that has changed much, so it can explain levels but not trends).

High and increasing cost of childrearing

The Japanese education system features a great deal of cutthroat competition, including private cram schools that impose significant out-of-pocket expenses, plus a lot of stress regarding whether one's child will get coveted university spots. It has been claimed that these costs have been increasing over time, making it less attractive to have children.