D ATA A N D P E R S P E C T I V E S East Asian Childbearing Patterns and Policy Developments TOMAS FREJKA GAVIN W. JONES JEAN-PAUL SARDON CHILDBEARING BEHAVIOR in East Asian countries has changed rapidly and fundamentally during the past 50 to 60 years. In the middle of the twentieth century large families with an average of five to seven children were the rule. In Japan almost immediately thereafter, and in the other countries 20 to 30 years later, replacement level was reached and since then fertility has continued its decline to the lowest levels in the world. Fifty years ago almost all women in East Asia married and had children, whereas at present up to one-quarter of women may remain single at age 50 and one-fifth to onethird may remain childless. Analyses of the institutional and socioeconomic environment affecting childbearing point to the likelihood of protracted very low fertility (Jones et al. 2009). Concern is growing about the future decline of these populations and their aging. Governments in East Asia increasingly perceive that the societal costs of rapid population aging and shrinkage are likely to outweigh possible benefits (ibid.). Therefore, at the end of the first decade of the twenty-first century most East Asian governments are pursuing pronatalist policies, whereas 50 years ago they were pioneers in implementing antinatalist policies. Thus far, the pronatalist policies have not generated the desired results. The analysis in this article reinforces the conclusion that fertility is likely to remain very low for years if not decades and that trends will not be reversed unless forceful and innovative policies are implemented.1 The first section analyzes fertility trends in Hong Kong, Japan, Singapore, South Korea, and Taiwan primarily using cohort fertility data and methods.2 This is followed by an examination of the social and economic causes of the childbearing trends. The next section provides an overview POPULATION AND DEVELOPMENT REVIEW 36(3): 579–606 (SEPTEMBER 2010) 579 580 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S and evaluation of policies that have been pursued to reverse the direction of fertility trends. A summary and conclusions constitute the last section. Data and methodological considerations Data to prepare long series of cohort fertility (total and by parity) for Hong Kong, Singapore, South Korea, and Taiwan and to update the series for Japan have been assembled with the help of colleagues in these countries. The basic requirement is a minimum of 30 years of single-year data on births by age of mother (preferably by parity) and single-year data on the female population by single year of age. For Hong Kong, Japan, and Taiwan the cohort fertility data were calculated from registration data, and for South Korea from a combination of registration and census data. For Singapore we did not have the needed 30 years of single-year data, but only five-year cohort age-specific fertility rates. These were prepared and discussed by Koh (2010). Given the nature of the Singapore data, the formal presentation in the figures differs somewhat from that of the other countries; however, they are sufficient for the present analysis. The tail ends of the cohort fertility series in Figures 2, 3, 4, and 5 are robust estimates. For the most part 99 percent or more of each measure is based on registered data. At most less than 5 percent of any measure is estimated, that is, at least 95 percent is based on registration. A major portion of the estimated part of any measure is also real, because it is based on values of previous cohorts, and only marginal parts of the measure can turn out to be different in the future. For instance, in the 1967 total cohort fertility rate for South Korea, 99.3 percent is based strictly on registered data and 0.7 percent is estimated. A major part of the estimated 0.7 percent, however, is rooted in real developments, because the real values of the age-specific fertility rates (ASFRs) from previous cohorts for ages 41–49 are used for this estimate. The real ASFRs might be marginally different in the future. In the estimate of the 1971 total cohort fertility rate for South Korea, 95.1 percent is based on registration and 4.9 percent is estimated as described above—that is, the values of the ASFRs for ages 37–49 are real values of recent cohorts, and the real future values can only be marginally different. Most of the units of measurement—the total period and cohort fertility rates, cohort age-specific fertility rates, cumulated cohort age-specific fertility rates—are commonly used and do not require any elaboration. One measure requires some explanation, namely the “recuperation index.”3 The process of childbearing postponement can be described by comparing the birth cohort of interest—the index birth cohort—with a benchmark birth cohort. Women in the index birth cohort, for instance 1970, when they are young, usually in their teens and early to mid-20s, have fewer births than women of the benchmark birth cohort—for instance, the 1960 cohort. When the women of the 1970 birth cohort become older, that is, in their late 20s T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N 581 and 30s, they may then bear some or all of the children they did not have when they were young. The recuperation index measures the proportion of the notional deficit in births that the women of the index cohort had, or made up, when they became older. The process of childbearing postponement and recuperation manifests itself as a decline in births among young women of a particular birth cohort compared to the childbearing of a previous cohort up to a certain age, usually in the mid or late 20s (a trough), and when these women become older their fertility tends to be higher than at the trough age. The extent to which women of a cohort recuperate births from a trough age to an older age, the recuperation index, can be measured by dividing (1) the difference between the deviation of the cumulated cohort agespecific fertility rate from a benchmark cohort (in numbers of births per woman) at a higher age and the value of the deviation at the trough age by (2) the absolute value of the deviation at the trough age. To illustrate, if the number of births at the trough differs from the base cumulated cohort ASFR by –0.5 and the number of births at age 404 deviates from the base cohort by –0.2, then the amount of recuperation is 0.3 births per woman (1). This value divided by 0.5 gives a recuperation index of 0.6 (2). Multiplied by 100 it is the extent of recuperation in percent. In our example 60 percent of the maximum number of births that were delayed by women when young were recuperated when these women were older. A recuperation index equal to 1.0 implies that all postponed births were recuperated.5 This index is calculated in Figure 9 based on data in Figure 8. Long-term fertility trends At the time when Japan had reached replacement fertility in the mid-1950s, the other countries of East Asia still had total period fertility rates (TPFRs) on the order of five to seven children per woman (Figure 1). During the following decades fertility declined rapidly in Hong Kong, Singapore, South Korea, and Taiwan. Singapore’s population reached replacement fertility in the mid-1970s, the other three in the 1980s. The fertility decline in these countries did not pause at that level but continued during the 1980s, 1990s, and into the 2000s. Japan’s TPFR fell below replacement in the mid-1950s, retained that level through the mid-1970s, and thereafter declined moderately and continuously into the 2000s. By the mid to late 2000s these countries had the lowest period fertility rates in the world, termed “ultra-low fertility” by Jones, Straughan, and Chan (2009). In 2005 the TPFR was 0.97 in Hong Kong, 1.11 in Taiwan, 1.12 in South Korea, 1.23 in Japan, and 1.25 in Singapore. The long completed cohort fertility series for South Korea demonstrates the fertility transition from a stable total cohort fertility rate (TCFR) of over four children per woman among women born in the 1920s and early 1930s to 582 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S FIGURE 1 Total period fertility rates, East Asian countries, 1947–2008 7 I 6 X Singapore I X F TPFR 5 I 4 F I Japan Taiwan I FF 3 X FI F I F X FF 2 1 X South Korea IX IX I II FF X F I I I II FX XIIIII II II X FX XIXX XXXXXX FX XX IX II FX FFFFFFFFF X IX FF XX I IX F F F F F F FX FF Hong Kong II I FF 0 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 Year SOURCES: Koh 2010; Observatoire Démographique Européen 2010; Saw 2007. replacement-level fertility among the late 1950s birth cohorts (Figure 2). Fertility remained around that level among the 1960s cohorts and then resumed its decline among the cohorts born around 1970. In the other four countries trends of completed fertility rates of the 1950s and 1960s birth cohorts illustrate the rapid changes in childbearing behavior taking place. Women born around 1950 in Hong Kong, Japan, and Singapore had an average of two children. Those born only 20 years later had an average of less than 1.5 children FIGURE 2 Total cohort fertility rates, East Asian countries, 1917–73 4.5 South Korea 4.0 TCFR 3.5 H 3.0 2.5 Taiwan Singapore H Japan H 2.0 H H HHH 1.5 HH HH Hong Kong 1.0 1915 1920 1925 1930 1935 1940 1945 1950 Birth cohort SOURCES: Koh 2010; Observatoire Démographique Européen 2010. 1955 1960 1965 1970 1975 T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N 583 in Singapore, 1.4 in Japan, and 1.2 in Hong Kong. In Taiwan childbearing descents were similarly fast but with a delay of about ten birth cohorts. Among the cohorts of the 1950s and 1960s not only were declines of higher-order births large and ongoing, but also declines of first and second births were considerable (Figure 3). The decline in first-order births in Japan and Hong Kong was so significant that about a third of all women of the cohorts born around 1970 remained without any children (Figure 4). These rates of childlessness are higher than anywhere else in the world. In Singapore over 10 percent of women born in the early 1940s were childless; among the cohorts born 20 years later, over 20 percent were childless, the trend implying a likely further increase in childlessness in subsequent cohorts. In Taiwan 20 percent of women born in the early 1970s were childless, and this proportion is also likely to continue to increase. The level of first-order births was the highest in South Korea. Nonetheless, here too these births were steadily declining: among the cohorts of the early 1970s almost 10 percent of South Korean women remained childless (Figure 4). Parity distributions Changes in parity distributions are another expression of the rapid transformation of childbearing behavior in East Asia (Figure 5). In Japan the overriding trends have been a decline in the share of two-child families from 55 percent among the cohorts born in the late 1940s to almost 20 percentage points lower in the late 1960s birth cohorts, and an increase from 7 percent to 32 percent in the share of childless women between the same cohorts (Figure 5, panel B). Further, between the cohorts of the mid-1950s and those of the late 1960s the share of three-child families declined considerably and onechild families increased moderately. Even though the available cohort parity distributions for Hong Kong and Taiwan are relatively few, the trends are unmistakable. The pattern of parity distribution changes in Hong Kong is similar to that of Japan (Figure 5, panel A). The share of two-child families in Hong Kong declined from a high of 45 percent in the mid-1950s cohorts to 35 percent among the birth cohorts of the mid to late 1960s. The increase in the share of childless women was even more pronounced than in Japan. Within the space of only ten birth cohorts, from those of the mid-1950s to those of the mid and late 1960s, the share of childless women increased in Hong Kong from 17 to 35 percent. The shares of large families were declining rapidly. In the cohorts born around 1950 some 33 percent of women had three or more children. This share declined to 15 percent in the cohorts of the early 1960s and has probably continued to decline. Among women born in the early 1960s the parity distribution in Singapore was still relatively balanced (Figure 5, panel C). Nonetheless, shares of childless women and of women with only one child were clearly on the rise, whereas shares of three- and four-child families were declining. Even the FIGURE 3 Total cohort fertility rates, selected East Asian countries, by birth order, birth cohorts 1945–72 1.0 Birth order one South Korea Japan 0.9 0.8 Taiwan 0.7 TCFR 0.6 Hong Kong 0.5 0.4 0.3 0.2 0.1 0 1940 1.0 1945 1950 1955 1960 Birth cohort 1965 1970 1975 Birth order two 0.9 Japan South Korea 0.8 0.7 Taiwan TCFR 0.6 Hong Kong 0.5 0.4 0.3 0.2 0.1 0 1940 1.0 1945 1950 1955 1960 Birth cohort 1965 1970 1975 Birth order three 0.9 0.8 0.7 TCFR 0.6 0.5 Taiwan 0.4 Hong Kong 0.3 0.2 Japan South Korea 0.1 0 1940 1945 1950 1955 1960 Birth cohort SOURCES: Observatoire Démographique Européen 2010; Saw 2007. 1965 1970 1975 585 T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N FIGURE 4 Proportions of childless women, East Asian countries, birth cohorts 1943–73 0.40 Hong Kong Proportion childless 0.35 0.30 Singapore 0.25 0.20 H Japan 0.15 0.10 H H H H Taiwan 0.05 0.00 1940 South Korea 1945 1950 1955 1960 Birth cohort 1965 1970 1975 SOURCES: Koh 2010; Observatoire Démographique Européen 2010; Saw 2007. proportions of two-child families were starting to decline among the cohorts of the late 1950s and early 1960s. A relatively long series of parity distributions for South Korea (based on census data) illustrates the fundamental changes that occurred from the birth cohorts of the mid-1930s through those of the mid-1960s (Figure 5, panel D). Over 70 percent of women in the older cohorts had four or more children whereas almost none of the women born 30 years later had families of this size. On the other hand, less than 10 percent of women had two-child families among the birth cohorts of the late 1930s, whereas among women born in the 1960s almost two-thirds had two children—a higher proportion than in any other country in the world. In contrast to the other East Asian countries, less than 10 percent of Korean women born in the mid-1960s were childless; nonetheless, their shares were rising. The parity distribution in Taiwan was also experiencing rapid changes among the 1950s and 1960s birth cohorts, but of a different nature than in Japan and Hong Kong. Over 50 percent of women were still having families of three or more children in the cohorts of the early 1950s (Figure 5, panel E). Their share declined swiftly to about 20 percent in the late 1960s cohorts. Between the cohorts of the mid-1950s and those of the early 1970s proportions of childless women and of one-child families were rising, each exceeding 20 percent and apparently poised to increase further. Proportions of two-child families rose to more than 40 percent in the early 1960s birth cohorts, and their shares appeared to start declining among the late 1960s A. Hong Kong 0.7 0.6 0.5 2 0.4 0 0.3 3 1 0.2 4+ 0.1 0 Proportion of women with X children 0.8 0.8 B. Japan 0.7 0.6 2 0.5 0.4 0.3 3 0.2 0.1 0 1 4+ 0 1935 1940 1945 1950 1955 1960 1965 1970 1975 1935 1940 1945 1950 1955 1960 1965 1970 1975 Birth cohort Birth cohort 0.8 C. Singapore Proportion of women with X children Proportion of women with X children Proportion of women with X children FIGURE 5 Parity distribution of completed fertility, East Asian countries, birth cohorts 1935–73 0.7 0.6 0.5 0.4 0.3 4+ 2 I 3 0.2 0.1 1 0 0 1941– 1946– 1951– 1956– 1961– 1966– 1971– 1945 1950 1955 1960 1965 1970 1975 0.8 0.7 D. South Korea 4+ 0.5 0.4 0.3 0.1 Proportion of women with X children 0.6 0.5 2 3 4+ 1 0.2 0.1 0 1935 1940 1945 1950 1955 1960 1965 1970 1975 0.7 0.3 1 0 E. Taiwan 0.4 3 0.2 Birth cohort 0.8 2 0.6 0 0 1935 1940 1945 1950 1955 1960 1965 1970 1975 Birth cohort SOURCES: Koh 2010; Observatoire Démographique Européen 2010; Saw 2007. Birth cohort 587 T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N cohorts. It is plausible that future trends in parity distribution in Taiwan could resemble those of the 1960s birth cohorts in Hong Kong and Japan, given that Taiwan’s cohort fertility levels and trends seem to be lagging by about 10 to 15 cohorts. Age patterns of fertility The average age of cohort childbearing has been increasing in the four countries with data (Figure 6). In Hong Kong and Japan the average age was below 28 in the early 1950s cohorts and rose to around 30 for women born in the late 1960s. In South Korea and Taiwan the increases in average ages at childbearing have been at similar rates but from lower bases, from age 26.5 to 29 in Korea and from 25.7 to 27.6 in Taiwan. The declines in completed cohort fertility (Figure 2) and in part the increases in average ages of cohort childbearing (Figure 6) have been due to fertility declines among women in their teens and 20s without sufficiently large compensating increases of fertility at later ages.6 In Figure 7 age-specific fertility rates of the birth cohorts 1960 through 1985 are compared for each country; in Singapore the comparison is for birth cohorts 1955–59 to cohorts 1975–79. Figure 8 illustrates how cumulated cohort fertility rates of women born in 1965, 1970, 1975, 1980, and 1985 compare with the 1960 benchmark; in Singapore the comparison is between the 1960–64, 1965–69, 1970–74, and 1975–79 cohorts and the 1955–59 benchmark. Finally, except for Singapore Figure 9 provides a concrete measure of the extent of childbearing recuperation among women born in the early to mid-1960s. FIGURE 6 Cohort average age of childbearing, selected East Asian countries, birth cohorts 1925–71 Mean age of cohort childbearing 30.0 29.5 29.0 Hong Kong 28.5 28.0 27.5 27.0 Japan South Korea 26.5 Taiwan 26.0 25.5 25.0 1920 1925 1930 1935 1940 1945 1950 Birth cohort SOURCE: Observatoire Démographique Européen 2010. 1955 1960 1965 1970 1975 588 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S Despite some differences in the age patterns of childbearing between the five countries, the decline of fertility among young women was common to all (Figure 7). This indicates that the “postponement” of childbearing is ongoing. Hong Kong displayed some peculiar irregularities in the childbearing age patterns with, for instance, the 1970 cohort having almost identical age-specific rates from age 24 to age 32 (panel A). It is notable that this cohort and that of 1975 are showing a propensity to recuperate some of the delayed births. The age pattens of cohort childbearing were quite regular in Japan, with clear declines of fertility among young women and apparently very weak recuperation in the 1965 cohort, but an indication of possible recuperation in the 1970 and 1975 cohorts (Figure 7, panel B). Singapore showed consistent fertility declines among women in their early 20s and evident declines in recent cohorts among women in their late 20s; there were practically no signs of childbearing recuperation among women in their 30s (panel C). The childbearing age patterns in South Korea are shifting into the higher ages with some notable increases of fertility among older women (panel D). In Taiwan some irregularities appeared in the childbearing age patterns; however, the decline of fertility among young women is obvious and the propensity to recuperate appeared to be weak (panel E). Childbearing postponement and recuperation Differences in cumulated cohort age-specific fertility rates between cohorts provide a reasonably good indication of the extent to which fertility is declining among young women and whether some of this decline is subsequently recuperated when women of the same cohort are older.7 For instance, women born in Taiwan in 1965 had 0.35 fewer births by the time they were 28 years old compared to women in the 1960 benchmark cohort at the same age (Figure 8, panel E). This was the largest age-specific difference between the two cohorts, the trough in the curve of the 1965 cumulated cohort age-specific fertility rates. After this age the curve followed a moderate upward slant until about age 35, illustrating that some of the deficit in births at young ages was being recuperated. Subsequently, after age 35 the curve remained virtually flat, because no additional births were being recuperated. In contrast, the curve of the 1970 birth cohort reached a trough of 0.55 births below the level of the 1960 cohort, also at age 28, but thereafter it remained flat. This indicates that practically none of the postponed births were recuperated. The graphs for Hong Kong and Japan in Figure 8 illustrate notable declines of childbearing among young women in the birth cohorts of the 1960s and 1970s; however, very few of the postponed births were recuperated (panels A and B). In other words, most of the declines in childbearing of the young women turned out to be real declines in completed fertility, not postponed FIGURE 7 Age-specific fertility rates, East Asian countries, selected birth cohorts 1960–85 (Singapore, birth cohorts 1955–59 to 1975–79) B. Japan A. Hong Kong 0.25 0.25 0.20 0.20 0.15 ASFR ASFR 1960 1960 0.10 0.15 1965 0.10 1975 1970 1980 1965 1975 1980 1985 0.05 0 15 20 25 0.05 1970 1985 30 35 Age 40 45 0 15 50 C. Singapore 20 25 30 35 Age 40 45 50 45 50 D. South Korea 0.25 0.25 0.20 0.20 0.15 ASFR ASFR 1965 1960–64 0.10 0.15 1975 0.10 1970–74 1955–59 1960 0.05 0.05 1975–79 1980 1965–69 0 15 0 15–19 20–24 25–29 30–34 35–39 40–44 20 Age E. Taiwan 0.25 0.20 ASFR 1960 0.15 1970 0.10 1965 1975 1980 0.05 1985 0 15 20 25 30 35 Age 40 45 1970 1985 50 SOURCES: Koh 2010; Observatoire Démographique Européen 2010; Saw 2007. 25 30 35 Age 40 FIGURE 8 Differences in cumulated cohort age-specific fertility rates between 1960 benchmark cohort and selected subsequent cohorts, 1965–85, East Asian countries (Singapore, birth cohorts 1955–59 (benchmark) and 1960–64 to 1975–79) A. Hong Kong (1960 TCFR = 1.58) B. Japan (1960 TCFR = 1.81) 0 Difference (number of births) Difference (number of births) 0 1985 –0.2 1965 –0.4 1980 1970 1975 –0.6 –0.8 –1.0 –1.2 15 20 25 30 35 Age 40 45 –0.2 1965 1980 –0.4 1970 1975 –0.6 –0.8 –1.0 –1.2 15 50 C. Singapore (1955–59 TCFR = 1.92) 20 30 35 Age 40 45 50 D. South Korea (1960 TCFR = 2.08) Difference (number of births) 1960–64 –0.2 1965–69 –0.4 1970–74 1975–79 –0.6 –0.8 –1.0 1965 –0.2 15–19 20–24 25–29 30–34 35–39 40–44 –0.6 1975 –0.8 –1.0 1980 20 Age E. Taiwan (1960 TCFR = 2.22) 0 –0.2 1965 –0.4 1985 1970 –0.6 1975 –0.8 1980 –1.0 –1.2 15 20 25 30 35 Age 40 45 1970 1985 –0.4 –1.2 15 –1.2 Difference (number of births) 25 0 0 Difference (number of births) 1985 50 SOURCES: Koh 2010; Observatoire Démographique Européen 2010; Saw 2007. 25 30 35 Age 40 45 50 591 T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N 1.0 0.8 0.6 Deficit in trough 0.4 0.2 Recuperation by age 40 0 Recuperation index –0.2 1960 1965 1970 1975 Birth cohort C. South Korea 1980 1.2 1.0 0.8 Deficit in trough Recuperation index 0.6 0.4 0.2 Recuperation by age 40 0 –0.2 1960 1965 1970 Birth cohort 1975 1980 Births per woman and recuperation index A. Hong Kong 1.2 Births per woman and recuperation index Births per woman and recuperation index Births per woman and recuperation index FIGURE 9 Deficit in births per woman in trough (difference in births among young women in a given cohort and births in the 1960 benchmark cohort), recuperation in births per woman by age 40, and recuperation index, selected East Asian countries, cohorts 1961–78 B. Japan 1.2 1.0 0.8 0.6 Deficit in trough 0.4 0.2 Recuperation index 0 Recuperation by age 40 –0.2 1960 1965 1970 1975 Birth cohort 1980 D. Taiwan 1.2 1.0 0.8 Deficit in trough 0.6 0.4 0.2 Recuperation index Recuperation by age 40 0 –0.2 1960 1965 1970 1975 Birth cohort SOURCE: Observatoire Démographique Européen 2010. births. For instance, Japanese women born in 1970 had 0.48 fewer births by age 33 compared to the 1960 cohort; and in their mid to late 30s they had about the same number of children as women born in 1960—that is, none of the delayed births were recuperated (panel B). Among the cohorts of the 1960s and early 1970s in Singapore, women’s propensity to recuperate births when they were older was either weak or non-existent (Figure 8, panel C). For instance, women ages 25–29 in the 1970–74 birth cohort had 0.35 fewer births than the 1955–59 cohort. For a childbearing recuperation to take place, the difference at ages 30–34 between the 1970–74 cohort and the benchmark cohort would have to diminish, but 1980 592 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S in reality it increased to 0.44 births. A similar pattern was observed for the 1965–69 cohort, with differences in the rates between this cohort and the benchmark cohort increasing with age instead of declining. The propensity to recuperate was also very weak in Taiwan (Figure 8, panel E). In South Korea there was an appreciable recuperation among women born in the 1960s and 1970s (panel D). FIGURE 10 Differences in cumulated cohort age-specific fertility rates between 1950 benchmark cohort and subsequent cohorts 1955–85 (A), and recuperation parameters for birth cohorts 1951–74 (B), Japan A. Difference between 1950 benchmark and later cohorts Difference (births per woman) 0 1955 1985 –0.2 1960 –0.4 1965 1980 –0.6 1970 –0.8 1975 –1.0 15 20 25 30 35 40 45 50 Births per woman and recuperation index Age B. Recuperation parameters 1.0 0.8 Recuperation index Deficit in trough 0.6 0.4 0.2 Recuperation by age 40 0 1950 1955 1960 1965 Birth cohort SOURCE: Observatoire Démographique Européen 2010. 1970 1975 1980 T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N 593 Figure 9 supplements the general picture of the extent of recuperation given in Figure 8 with precise calculations for four of the countries, in each case with the 1960 birth cohort as the benchmark. The longest curve in each panel depicts the maximum cumulated deficit in births to young women (teens and early to mid-20s) in each consecutive cohort, that is, the troughs of the cumulated cohort age-specific fertility rates. (In Taiwan, for example, these ranged from 0.083 births per woman in the 1961 cohort to 0.808 in the 1975 cohort.) The curve labeled recuperation by age 40 depicts the relative increase in childbearing of older women—that is, the increment in childbearing between the trough and age 40. (In Taiwan these ranged from 0.038 in the 1961 cohort to 0.073 in the 1967 cohort.) The product of the two values for each cohort gives the recuperation index.8 Figure 9 demonstrates that the recuperation of postponed births in the 1960s cohorts has been weak. In Hong Kong about 20 percent of births in the 1961 and 1962 birth cohorts were recuperated, but among the cohorts of the mid-1960s no births were recuperated (panel A). In Japan between 7 and 20 percent of delayed births were recuperated among women born in the early to mid-1960s (panel B). In South Korea between 40 and 60 percent of births were recuperated in the 1960s cohorts (panel C), and in Taiwan there was a decline in recuperation from over 40 percent in the 1961 cohort to under 20 percent in the 1967 cohort (panel D). Thus far, only relatively short data series are available for Hong Kong, South Korea, and Taiwan. For Japan a longer cohort series is used in Figure 10, extending the picture already shown in Figures 8B and 9B. Panel A of Figure 10 shows that childbearing postponement has been underway since the 1950s birth cohorts, and it is obvious that none of the postponed childbearing of young women was ever fully recuperated. Panel B shows that around 70 percent of the delayed births in the early 1950s cohorts were recuperated. A considerable decline in the extent of recuperation occurred between women born in the mid to late 1950s and those born in the early 1960s. For the mid-1960s cohorts recuperation was around 20 percent. An international comparison Total period fertility rates in the five East Asian countries studied here were among the lowest in the world. Because period fertility rates do not reflect the fertility experience of actual cohorts, one can obtain a better picture by comparing a set of measures. Selected fertility indicators are listed in Table 1, which compares the East Asian countries with German-speaking countries of Central Europe—Austria, Germany, and Switzerland—whose fertility measures tend to be among the lowest in Europe (Frejka 2009). The TPFRs in all the East Asian countries are lower than in the Germanspeaking countries. The total cohort fertility rate of the 1968 birth cohort 594 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S TABLE 1 Selected fertility measures, East Asian and German-speaking European countries, recent years Measure East Asian countries German-speaking countries Hong Singa- South SwitzerKong Japan pore Korea Taiwan Austria Germany land Total period fertility rate, 2005 0.97 1.23 Total cohort fertility rate, 1968 cohort 1.16 1.46 Percent of women childless (late 1960s cohorts) 35 31 Percent of recuperated births in 1960s cohorts 0–20 5–20 1.25 1.12 1.11 1.41 1.34 1.42 1.72 1.96 1.81 1.61 1.48 1.64 21a 8 16 21 40–60 20–40 40–60 n.a. 28b 40 21 30–40 a Circa 1961–65. West Germany. SOURCES: Koh 2010; Observatoire Démographique Européen 2010; Frejka 2009; OECD 2010; authors’ calculations. b in Hong Kong (1.16) is by far the lowest. Japan’s TCFR is comparable to Germany’s and lower than in the other German-speaking countries. TCFRs in Singapore, South Korea, and Taiwan are higher; however, together with Japan these are on steeply declining trajectories and are likely to continue to decline. The shares of childless women in the late 1960s birth cohorts in Hong Kong and Japan were higher than in Austria and even higher than in West Germany. In East Germany the share of childless women was 16 percent, implying that the overall childless rate in Germany as a whole was around 24 percent (Frejka 2009). The degree to which births are being recuperated arguably is the broadest measure and provides an important insight with regard to fertility trends. Among the East Asian countries, the percentage of births that were being recuperated was comparable to the German-speaking countries only in South Korea; elsewhere the degree of recuperation was distinctly lower. This comparison confirms the conclusion of the previous analysis that childbearing recuperation in East Asia is weak. Causes of low East Asian fertility Theories of low fertility have been well summarized by McDonald (2002) under four headings: rational choice, risk aversion, post-materialist values, and gender equity. Elements of all four theories can be invoked here to explain the sinking fertility levels in these East Asian countries, and particularly the lack of evidence for any substantial recuperation of delayed births. It needs to be emphasized that we are dealing here with countries whose economic success over extended periods was largely unexpected: first, Japan from the 1950s up to the “lost decade” of the 1990s,9 and then the other T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N 595 “Asian tigers” (Hong Kong, South Korea, Singapore, and Taiwan) from the 1970s up to the Asian financial crisis of 1997–98. Consumerism has helped to ensure that, despite rapid gains in real incomes, families increasingly feel that they need two incomes to survive. Women want to be in the workforce, but they are placed under immense pressure if they also aim to raise a family in the way that East Asian societies expect: These are highly competitive economies and governments are determined to increase productivity and keep wages down. Employers remain relatively unforgiving of the divided loyalties inherent in the effort to combine child-raising with working, and society remains unsupportive of those who want to pursue non-material goals or who are not totally devoted to pressing their children to maximum performance. Moreover, in recent years greater uncertainty has entered the labor market, making many young people more apprehensive about taking on the commitments that starting a family entails. (Jones, Straughan, and Chan 2009: 210) The rapid economic growth in these countries has been paralleled by remarkable increases in education, especially for women. In South Korea, the proportion of young women graduating from high school who advanced to higher education increased from 20 percent in 1975 to 34 percent in 1985, 50 percent in 1995, and 81 percent in 2005. The rise in the proportion of women entering high school who subsequently proceeded to higher education was even more remarkable, with high school education becoming almost universal over the period. “The educational advancement of young Korean women during the last three decades is nothing but spectacular and, to our knowledge, unprecedented in the recent history of the world” (Tsuya, Choe, and Wang 2009: 16). The increases in women’s education in Hong Kong, Japan, and Taiwan, while not quite as spectacular as those in South Korea, are impressive nonetheless (Jones and Gubhaju 2009). The men these women will (or will not) marry have also experienced substantial improvements in levels of education, though not to the same extent as women. It is the aspirations of these young women and men that underlie the sharp declines in fertility in the region. The life options of young women have widened, and at the same time women have been exposed to values that compete with their domestic roles (Rindfuss et al. 2004; Atoh 2001, 2008). Although a decomposition analysis of the decline in Korean fertility between 1995 and 2005 shows that only about 7 percent can be attributed to changes in population composition by women’s education, Choe and Retherford (2009: 285–286) speculate that the effect of compositional change is actually much greater than this, acting through problems of employment for the surging number of college and university graduates; the near impossibility of temporarily shifting from full-time to part-time work in order 596 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S to care for young children, resulting in delayed marriage and falling fertility; and a tradeoff in favor of quality of children rather than quantity, owing to apprehension about children’s future job prospects. Women with low education were also under pressure in the job market because of competition from better-educated women who could not find jobs commensurate with their schooling. Such arguments can help explain why the upsurge in singlehood and the decline in fertility in Japan and South Korea since the early 1990s have affected women in all education and social groups, rather than simply being a consequence of more women entering the higher educational groups where singlehood and low fertility are more common (Jones and Gubhaju 2009). The removal altogether of an increasing proportion of women from the childbearing population as a result of non-marriage, and the late initiation of childbearing because of delayed marriage, have been key factors in fertility declines in the region. While cohabitation is becoming more widely accepted, at least in Japan (Raymo et al. 2008), childbearing within cohabiting relationships remains rare,10 and unless and until normative changes make inroads into this rarity, marriage will remain the gatekeeper into the possibility of childbearing in the region. Part of the delay in marriage in the region may be a direct response to a desire to avoid or delay childbearing. As Bumpass et al. (2009: 217) note with reference to Japan, numerous intra-familial roles are bundled together in a ‘marriage package.’ Marriage, childbearing, childrearing and frequently, care of the elderly, are linked. When women marry, they are expected to assume these roles, and marriage is expected to be a life-long commitment. Household tasks within marriage life are extremely gender-segregated, with the burden falling overwhelmingly on the wife. Further, the mother is chiefly responsible for the educational success of her children. She not only helps with the child’s homework, but must also attend to all the arrangement details if her child attends one of the very popular juku, i.e. after-school ‘cram schools’ that prepare children for entrance examinations to higher-level schools.11 And “[f]or many, the entire package of marital roles of the wife is what is being delayed, including children with their intensive care needs, a heavy household task load, and co-residence with parents-in-law, which is potentially included in the bargain” (ibid., p. 218). There are other barriers to marriage in the region. Hypergamy remains strong, and in general the notion of who should marry whom is less flexible than in Western countries (Jones, Straughan, and Chan 2009: 208). In much of East Asia, women are outpacing men in graduating from tertiary levels of education, while traditional age gaps between husbands and wives remain. This is a context tailor-made for rising levels of involuntary non-marriage, T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N 597 which may be playing a substantial role in the low fertility currently recorded in these countries (Jones 2007: 472). The trends in marriage and fertility are also influenced by unequal gender relations. In the patriarchal and patrilineal traditions of countries in the Chinese cultural sphere (which includes Korea), as well as in Japan, strong norms about separate gender roles have “served as the basis for the persistence of unequal gender relations at home” (Tsuya, Choe, and Wang 2009: 3). Although paid employment of women and its corollary—dualearner households—in these countries have increased considerably, the Western pattern in which men’s contribution to household chores has also increased substantially is not being replicated in Japan or South Korea, despite women’s changing expectations about gender roles. Thus it is becoming increasingly difficult for women to balance their economic and family responsibilities. Greater job uncertainty since the 1990s (especially in Japan) and the Asian economic crisis of 1997–98 have underlined the need for caution in marrying and beginning a family. The tendency in East Asian countries for young people to delay marriage because of the perceived need to become settled in a career, and the difficulties in finding a suitable spouse once the search begins in earnest, “pushes potential childbearing into an age range where the obstacles and opportunity costs are likely to loom even larger, and where decreasing fecundity of women in the late 30s and 40s becomes a more salient but under-recognized factor” (Jones 2007: 472). To summarize, rising education of women has played a key role in the very low fertility reached in East Asia. Yet it is not the high level of education per se that is important. High levels of female education in some northern and western European countries, the United States, Australia, and New Zealand have not prevented their levels of fertility being close to replacement at present. Rather, it is the extraordinarily rapid expansion of education—more so than in the West—in the East Asian institutional context that is the key. “There is considerable tension between the demands of an urban, postindustrial society and the traditional package associated with marriage” (Rindfuss et al. 2004: 841). Neither labor market conditions, employer and co-worker attitudes, nor husband’s or potential husband’s attitudes make it easy for women to combine full-time work with raising a family. This fact poses a stark dilemma for women who would like to combine childbearing with a career. Policy responses The history of policy responses in East Asia to the decline in fertility to ultralow levels has been one of delayed reaction in the first place,12 followed by modest policy and programmatic changes that can be characterized as “too little, too late,” and then most recently by a concerted attempt to address the 598 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S issue. The policy responses have been summarized in a number of recent publications (Jones, Straughan, and Chan 2009; Suzuki 2009; Lin and Yang 2009; Lee, Meilin 2009; Lee, Sam-Sik 2009; Eun 2006). Suzuki (2009) argues that the Confucian tradition in such countries as South Korea and Taiwan favors an authoritarian approach even after the establishment of democracy and a market economy; nevertheless, South Korea and Taiwan show very different policy approaches to influencing the value orientations of their populations, with the South Korean government interested in preserving conservative family values and the Taiwanese government, apparently bowing to women’s political influence, accepting feminist values. We summarize below the kinds of pronatalist policies some East Asian countries have been introducing.13 Japan Japan has gradually been expanding its pronatalist policies since 1990 as the level of concern with very low fertility has risen.14 Japanese policy has followed two main approaches: direct subsidies for childbearing and childrearing; and changes to the institutional framework to facilitate marriage and childrearing. As in many other countries, child allowances were introduced (first in 1972) as a family policy measure to assist low-income families, rather than as a pronatalist measure. But after 1990, pronatalist concerns led to large increases in the allowances, though an eligibility criterion remains. In 1991 unpaid leave for child care (except for part-time workers) was introduced. The “Angel Plan” introduced in 1994 called for major expansion of child care centers, with eligibility criteria varying by locality. Later, partial payment of salary during child care leave was introduced. In 1999 the New Angel Plan stressed the need to improve gender equity and working conditions. It called for further expansion of the heavily subsidized day care centers, after-school programs, and family support centers. The eligible age range for receiving a child allowance was lengthened (and lengthened further in 2006). In 2001, the proportion of salary received by an employee on child care leave (a leave made available to husbands as well as wives, for up to one year) was raised from 25 percent to 40 percent, and raised again to 50 percent in late 2007. In 2006, the monthly cash benefit of the child allowance was raised from 5,000 to 10,000 yen until a child’s third birthday, though large tax deductions for children are not yet entertained. Few men appear to take child care leave, and many women were not doing so either because of disapproval from fellow workers. Thus the government introduced measures aimed at creating an atmosphere within firms that would encourage parents to take the child care leave to which they were entitled, and also introduced measures to support mothers’ re-entry to the labor market. In 2005, the government extended the right to child care leave to part-time workers under certain circumstances. Workers on short-term T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N 599 (e.g., 3-month or 6-month) contracts whose employer does not renew the contract are not eligible for such leave, however. South Korea In 2006, by which time fertility had been below replacement for more than 20 years and the TPFR below 1.2 for the previous five years, South Korea’s government, after extensive discussions with representatives of employers, workers, activists, and feminists, announced the First Basic Plan for Low Fertility and Aged Society, covering the period 2006–10, to be followed by a second plan for 2011–15. This plan attempts to transfer some of the burden of childrearing from family to society. Subsidies for the costs of childrearing and education, previously available for low-income families, were provided for the middle class as well. Day care for children up to age four is subsidized according to the family’s income level. After-school programs were expanded, particularly for lower-grade primary school children, as an alternative to expensive private tutoring. Taxes were also lowered for households with young children or large families, and the tax system was altered to reduce the costs of health insurance for such families. (For details, see Lee, Sam-Sik 2009, Tables 7 and 8.) Maternity and child care leave was expanded, and a child allowance is being considered. The government is playing a central role in increasing the number and quality of child care facilities. This is part of an effort to create a familyfriendly and gender-equitable society, with work and family more compatible with each other. Companies providing maternity leave are being supported through measures such as subsidization of maternity leave grants for smalland medium-size establishments and official recognition of family-friendly establishments (Lee, Sam-Sik 2009). Child care leave that previously applied only to parents of children under the age of one will now be extended to children up to three years of age. Taiwan Taiwan, too, had no pronatalist measures until 2006, when the Mega Warmth Social Welfare Program was introduced, followed in 2008 by the White Book of Population Policy, with specific measures to tackle low fertility. Feminists’ arguments against “buying of feminine wombs for cash” prevailed against payment of birth allowances (Lee, Meilin 2009). The measures introduced include maternity and parental leave benefits, a child care subsidy system, and early childhood education and care (Lin and Yang 2009). A provision for eight weeks of paid maternity leave is covered by labor insurance, which means that employers shoulder only 70 percent of the cost. Parental leave of up to six months on half pay is now available to mothers. The government 600 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S plans to certify qualified child-minders and subsidize child care services for low-income working families. Hong Kong Hong Kong’s government, one of the most laissez faire in the world, has been reluctant to introduce maternity leave (let alone paternity leave) policies that raise the cost of labor for businesses, or to provide major subsidies or tax breaks for dependent children. Working hours in Hong Kong are very long, but subsidization of child care costs is restricted to families meeting a means test. Singapore Singapore started earlier and has gone further than the other East Asian countries in some areas of policy designed to influence fertility, particularly in the use of tax rebates to encourage early childbearing, the scale of child allowances, fairly generous maternity leave provisions, subsidization of child care, and programs to encourage marriage, especially among highly educated women (Saw 2005; Yap 2009; Straughan, Chan, and Jones 2009). It may well be that Singapore’s more comprehensive policies to support marriage and childbearing help to explain why fertility rates in Singapore, though disappointingly low from the government’s perspective, are about 25 percent higher than in Hong Kong, and considerably higher still than in some other cities in the region: Beijing, Busan, Seoul, Shanghai, and Taipei. In contrast to the other countries in this study, Singapore’s female singlehood rates remained roughly constant in the 15 years up to 2005, and, when controlled for education, they actually declined (Jones and Gubhaju 2009: 248–251). These idiosyncratic trends suggest that Singapore’s unique pro-marriage policies may be having some effect. On the other hand, the baby bonuses and tax concessions for children are not substantial enough to make much of a dent in the high monetary costs of raising children. An evaluation of policy developments What can we conclude about policy developments in these five East Asian countries? First, the policies in Japan and Singapore have become progressively more comprehensive over time, but they are not yet as comprehensive as those in the universalistic or liberal welfare states of Europe.15 Second, South Korea and Taiwan have introduced far-reaching pronatalist measures only recently. Third, consistent with the continued dominance of patriarchal family and societal structures in these countries, policy does not yet take men’s role in childrearing as seriously as it does in European countries. T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N 601 None of them has introduced paid paternity leave. Fourth, metropolitan cities such as Seoul, Taipei, and Tokyo have introduced their own programs to support multi-child families, in addition to those introduced by the central government. Seoul has 65 programs in the following areas: encouragement of childbirth, child care benefits, establishing a family-friendly environment, and human-resource development plans for women (Jun 2009). The fertility-depressing factors already noted continue to exert their influence. One or more of these factors would need to be substantially modified if there is to be much likelihood of a resurgence of fertility in East Asia. For example, the trend toward more irregular and part-time employment, while it provides conditions that may enable more women to combine work with childrearing, increases the unlikelihood of continued uninterrupted employment for single people and couples; the trend may have to be reversed if young couples are to have the confidence to marry and start raising a family. Although the policy armory is being gradually built up throughout the region, assessing the impact of pronatalist efforts is likely to be difficult, given that it is probably being counteracted, at least in the short term, by caution about marriage and childbearing stemming from the global financial recession. The typically very small contribution of husbands to housework and childrearing means that the burden borne by working women remains high. Much remains to be done from a policy standpoint to encourage greater male involvement in childrearing and ensure the spread of family-friendly workplaces. In all these countries, women are apprehensive about taking full advantage of the parental leave benefits available to them, for fear of provoking negative reactions from their employers and fellow employees and damaging their future career prospects; and governments are apprehensive about extending maternity and child care leave benefits too far, for fear of resistance from employers. Particularly in smaller enterprises, which have the most difficulty in adjusting their staff allocation to worker absences, employers may avoid hiring reproductive-age women or find spurious reasons to dismiss those who become pregnant. A final point is that compared with Europe, East Asian countries will have a harder time dismantling the cultural barriers to raising fertility. In these patrilineal and patriarchal societies, long-held attitudes and gender relationships cannot be changed rapidly. The focus of policy will have to extend beyond making it financially easier to raise a family and ameliorating the dilemmas of reconciling family and workforce roles. More broadly, policy will have to consider how social institutions can be changed in the interest of overcoming ingrained sentiments and attitudes within the family, the business sector, and institutions of governance. These are societies that expect and respect moral leadership from government (Caldwell 1993), hence governments in the region should promote public discussion of controversial 602 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S issues such as the need for income transfers to families who have children and changes to make workplaces more family friendly. One hopeful sign of change is the return of South Korea’s sex ratio at birth to near-normal levels after rising to levels around 115 males per 100 females in the early 1990s. Although part of the explanation for this development is the strong government commitment to deter and penalize sex-selective abortions, it appears that the rising status of women in Korean society as a result of educational and employment advances was also critical (Guilmoto 2009). This evidence of modification of traditions in the rapidly changing economic and social order in Korea could have implications for institutional changes favorable to increases in fertility rates. Summary and conclusions Completed cohort fertility rates in the five East Asian countries studied in this article are as low as, or lower than, the lowest rates in Europe. Births of all birth orders are declining. About one third of all women in Japan and Hong Kong who reached the end of their potential childbearing ages early in the twenty-first century remained childless. In successive cohorts, women in their teens and 20s are bearing fewer children than those in previous cohorts, and only small proportions of the presumably “postponed” births are actually being recuperated when women are older. Social and economic factors depressing fertility appear to be more dominant than they are in Europe. In the past several decades East Asian countries have experienced rapid economic growth with fast-rising living standards. Their economies generated a robust demand for labor, and both single and married young women were eager to satisfy this demand. Large proportions of women acquired a college education and joined the labor force. Social conditions, however, have not kept pace with changing economic realities. Women are expected not only to run the household and bear and raise children, but frequently are also expected to care for elderly parents. Employers, who are focused on production and profits, are not intent on providing family-friendly work conditions for women. Consequently, many women are delaying marriage and childbearing. East Asian policymakers are gradually realizing that the current ultralow fertility leads to aging and shrinking populations, which in turn generate serious problems in the health and social security systems. The realization that increased fertility would slow population aging is taking hold. Some policymakers are promoting and adopting family welfare and pronatalist policies, although many bureaucrats are not convinced of their necessity and often oppose them. The governments of Japan and Singapore have been implementing relatively comprehensive policies intended to raise fertility for close to two decades, governments in South Korea and Taiwan have started T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N 603 to do so more recently, and so far the government of Hong Kong is taking a laissez faire attitude. The policies consist mainly of financial and material incentives, together with support for parents to combine work and family, such as various child care allowances, parental leave, and child care facilities. Little attention is devoted to generating broad social change supportive of children and parenting. Patriarchal customs and attitudes in the family, the workplace, and the political domain are deeply ingrained. Changing the patriarchal social environment will require special focus on policies to increase male involvement in the household and in the upbringing of children and to change the attitudes of employers. Unless current conditions are radically changed and child- and family-friendly environments are fostered, it is difficult to believe that fertility patterns will change. Notes Figures in this article are available in color in the electronic edition of the journal. The authors are grateful to the following colleagues, who kindly assisted in assembling the data for this international comparative analysis: Tsang Tat-shing, Hong Kong Census and Statistics Department; Rikiya Matsukura and Naohiro Ogawa, Nihon University, Tokyo, Japan; Koh Eng Chuan, Asia Research Institute, Singapore; Robert D. Retherford and Minja Kim Choe, East-West Center, Honolulu, Hawaii; Peter McDonald and Rebecca Kippen, Australian National University, Canberra; Chen Yu-hua, National Taiwan University, Taipei, Taiwan. 1 For the most part the analysis in this article is limited to Hong Kong, Japan, Singapore, South Korea, and Taiwan, because detailed demographic data were available only for these populations. Although Singapore is located in Southeast Asia, it fits logically within this East Asian grouping in cultural terms and as one of the rapidly developing Asian tigers. 2 For simplicity, we refer to Hong Kong SAR as a country. 3 The recuperation index is included in work in progress by one of the authors of this article, Tomas Frejka, with Tomáš Sobotka, Kryštof Zeman, and Ron Lesthaeghe to improve the measurement and analysis of childbearing advancement, postponement, and recuperation. For earlier methods to measure recuperation see Lesthaeghe 2001 and Frejka and Sardon 2004, pp. 326–327 and 348–350. 4 Theoretically age 50 years is accepted as the end of the reproductive period; however, in developed countries in recent decades women bear very few births in their 40s. In the interest of calculating childbearing recuperation for relatively recent cohorts, age 40 was selected as appropriate for this calculation. In the countries dealt with in this article the percentage of births that occurred after woman’s age 40 in the 1960 birth cohorts was 1.09, 0.89, 0.50, and 0.49 in Hong Kong, Japan, South Korea, and Taiwan. 5 Childbearing postponement takes the form of a fertility decline among young women that is a combination of actual postponement and a quantum decline of fertility. Some young women have fewer children than women in earlier cohorts or have none. The proportion of the fewer births that was actually postponed becomes apparent only when the generation reaches the end of its childbearing age and it is seen how many of the postponed births were born in these women’s later years. For discussion of this issue see section 2.6, “The postponement of childbearing,” in Frejka and Sardon 2004, pp. 17–18. 6 Data availability permits the analysis of childbearing age patterns only beginning with the 1960 birth cohort in Hong Kong, South Korea, and Taiwan. In Japan the decline of fertility among younger women and the 604 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S relatively weak propensities of older women to recuperate delayed births have been evident starting with the birth cohorts of the early 1950s (cf. Frejka and Sardon 2004, pp. 279–282 and Figure 10 below). 7 A decline in childbearing of young women is often referred to as fertility postponement. For discussion of this issue see section 2.6, “The postponement of childbearing,” in Frejka and Sardon 2004, pp. 17–18. 8 These calculations can be performed only for cohorts that have (almost) completed their childbearing, in this case for women born in the early to mid-1960s, given the data available for the East Asian countries (and only for countries with single-year data, which excludes Singapore). 9 Although, for Japan, the 1990s were the “lost decade,” Japan had already reached very high levels of productivity and of family and personal income before then. The lost decade therefore represented a failure to advance further in a context of rising aspirations, and this no doubt contributed to the unease about making a commitment to family building. A similar point can be made about South Korea following the 1997–98 financial crisis. Rising unemployment rates shook self-confidence among Koreans (Eun 2006) and coincided with a sharp upturn in singlehood rates. 10 For example, in South Korea and Singapore, 98.5 percent and about 99 percent of births in 2007 were to legally married couples (Lee, Sam-Sik 2009: 57; Koh 2010: 79), and in Japan this figure has been around 98–99 percent for at least the past 40 years (Rindfuss et al. 2004: 841). 11 For a study of the dilemmas facing Japanese mothers in this respect, see Hirao 2007. 12 Delays in reversing antinatalist policies in the region have been remarkably long: typically between one and two decades after fertility first fell to replacement level (Jones, Straughan, and Chan 2009: Table 1.2). One reason is that each of these countries except Japan had developed a strong family planning program, oriented to reducing fertility from what had been perceived as unacceptably high levels in these densely populated coun- tries; and the mindsets of planners and the bureaucratic interests involved in continuing this program proved difficult to adjust to the radically altered demographic situation. Another reason, of course, was that population momentum was still increasing total population size, even though fertility levels were well below replacement. 13 We have not included China in our analysis because of the lack of appropriate data. It should be obvious, however, that the demographic circumstances in China are analogous to those in the countries covered, with a period TFR around 1.4 to 1.6 and well below 1.0 in the large cities of Beijing and Shanghai. Demographers in China agree that antinatalist policies need to be reversed, but this step is resisted by the central government (Wang 2010; Gu 2010). Fears that a repeal of the one-child policy would lead to a fertility rebound seem to be misplaced, because fertility intentions for most of the population appear to be below replacement as a result of changing economic and social conditions (Morgan et al. 2009; Gu 2009: 80–86; Zheng 2010). 14 For detailed discussion, see Ogawa, Retherford, and Matsukura 2009; Suzuki 2009. 15 Common classifications of welfare state regimes in Europe distinguish between universalistic welfare states (the Nordic countries), conservative welfare states (continental European countries), liberal welfare states (Anglo-Saxon countries), and Southern European welfare states (see, for example, Gauthier 2002: Table 1). Southern European countries display a high degree of familialism. In this respect, they clearly share an important distinction with the East Asian countries currently confronting the issue of how to raise birth rates, namely that of being among the countries with the lowest levels of fertility in the world. McDonald (2000a, 2000b) has argued persuasively that the sharing of these two common elements is not accidental; it is precisely these countries’ familialism, in the context of greater educational and employment opportunities for women, that poses strong conflicts of interest for women and leads to delayed marriage and low levels of childbearing. T O M A S F R E J K A / G AV I N W. J O N E S / J E A N - P A U L S A R D O N 605 References Atoh, Makoto. 2001. “Very low fertility in Japan and value change hypotheses,” Review of Population and Social Policy 10: 1–21. ———. 2008. “Family changes in the context of lowest-low fertility: the case of Japan,” International Journal of Japanese Sociology 17: 14–29. Bumpass, Larry L., Ronald R. Rindfuss, Minja Kim Choe, and Noriko O. Tsuya. 2009. “The institutional context of low fertility: The case of Japan,” Asian Population Studies 5(3): 215–235. Caldwell, John C. 1993. “The Asian fertility revolution: Its implications for transition theories,” in Richard Leete and Iqbal Alam (eds.), The Revolution in Asian Fertility: Dimensions, Causes and Implications. Oxford: Clarendon Press. Choe, Minja Kim and Robert D. Retherford. 2009. “The contribution of education to South Korea’s fertility decline to ‘lowest-low’ level,” Asian Population Studies 5(3): 267–287. Eun Ki-Soo. 2006. “Lowest-low fertility in the Republic of Korea: Causes, consequences and policy responses,” paper presented at UN-ESCAP Seminar on Fertility Transitions in Asia: Opportunities and Challenges, Bangkok, 18–20 December. Frejka, Tomas. 2009. “Russia and Germany: Some similarities and many differences in fertility patterns and trajectories,” Conference on Challenges in Family and Fertility Development in Russia and Germany, 7–8 December «http://www.dia-dem.de/». Frejka, Tomas and Jean-Paul Sardon. 2004. Childbearing Trends and Prospects in Low-Fertility Countries: A Cohort Analysis. Dorbrecht: Kluwer Academic Publishers. Gauthier, A. 2002. “Family policies in industrialized countries: Is there convergence?,” Population (English Edition) 57(3): 447–474. Gu, Baochang. 2009. “The arrival of low fertility in China,” in Gavin Jones, Paulin Tay Straughan, and Angelique Chan (eds.), Ultra-low Fertility in Pacific Asia: Trends, Causes and Policy Issues. London: Routledge. ———. 2010. “A journey of eight years: A study of China’s fertility policy,” paper presented at the Annual Meeting of the Population Association of America, Dallas. Guilmoto, Christophe Z. 2009. “The sex ratio transition in Asia,” Population and Development Review 35(3): 519–549. Hirao, Keiko. 2007. “Contradictions in maternal roles in contemporary Japan,” in T. Devasahayam and Brenda S. A. Yeoh (eds.), Working and Mothering in Asia: Ideologies and Identities. Singapore: NUS Press. Jones, Gavin W. 2007. “Delayed marriage and very low fertility in Pacific Asia,” Population and Development Review 33(3): 453–478. Jones, Gavin W. and Bina Gubhaju. 2009. “Factors influencing changes in mean age at first marriage and proportions never marrying in the low-fertility countries of East and Southeast Asia,” Asian Population Studies 5(3): 237–265. Jones, Gavin, Paulin Tay Straughan, and Angelique Chan (eds.). 2009. Ultra-low Fertility in Pacific Asia: Trends, Causes and Policy Issue. London: Routledge. Jun, Hey-Jung. 2009. “An analysis of Seoul City’s support system for multi-child families,” paper presented at the Seoul International Forum on Low Fertility Policy, Seoul, 4 September. Koh Eng Chuan. 2010. “Will Singapore’s fertility rise in the near future? A cohort analysis of fertility,” Asian Population Studies 6(1): 69–82. Lee, Meilin. 2009. “Transition to below replacement fertility and policy response in Taiwan,” The Japanese Journal of Population 7(1): 71–86. Lee, Sam-Sik. 2009. “Low fertility and policy responses in Korea,” The Japanese Journal of Population 7(1): 57–70. Lesthaeghe, Ron. 2001. “Postponement and recuperation: recent fertility trends and forecasts in six Western European countries,” Interuniversity Papers in Demography 2001-1, Vrije Universiteit Brussel. 606 E A S T A S I A N C H I L D B E A R I N G P AT T E R N S A N D P O L I C Y D E V E L O P M E N T S Lin, Wan-I and Shin-Yi Yang. 2009. “From successful family planning to the lowest of low fertility levels: Taiwan’s dilemma,” Asian Social Work and Policy Review 3: 95–112. McDonald, Peter. 2000a. “Gender equity in theories of fertility transition,” Population and Development Review 26(3): 427-439. ———. 2000b. “Gender equity, social institutions and the future of fertility,” Journal of Population Research 17(1): 1–16. ———. 2002. “Sustaining fertility through public policy: the range of options,” Population (English Edition) 57(3): 417–446. Morgan, S. Philip, Zhigang Guo, and Sarah H. Hayford. 2009. “China’s below-replacement fertility: Recent trends and future prospects,” Population and Development Review 35(3): 605–629. Observatoire Démographique Européen. 2010. Data Bank. St. Germain-en-Laye, France. OECD 2010. OECD Family Database «http://www.oecd.org/els/social/family/database», OECD Social Policy Division, Directorate of Employment, Labour and Social Affairs. Ogawa, Naohiro, Robert D. Retherford, and Rikiya Matsukura. 2009. “Japan’s declining fertility and policy responses,” in Gavin Jones, Paulin Tay Straughan, and Angelique Chan (eds.), Ultra-low Fertility in Pacific Asia: Trends, Causes and Policy Issues. London: Routledge. Raymo, James, M. Iwasawa, and L. Bumpass. 2008. “Cohabitation and family formation in Japan,” CDE Working Paper 2007-14, Center for Demography and Ecology, University of Wisconsin-Madison.. Rindfuss, Ronald R., Minja Kim Choe, Larry L, Bumpass, and Noriko O. Tsuya. 2004. “Social networks and family change in Japan,” American Sociological Review 69(6): 838–861. Saw Swee-Hock. 2005. Population Policies and Programmes in Singapore. Singapore: Institute of Southeast Asian Studies. ———. 2007. The Population of Singapore, 2nd edition. Singapore: Institute of Southeast Asian Studies. Straughan, Paulin Tay, Angelique Chan, and Gavin Jones. 2009. “From population control to fertility promotion: A case study of family policies and fertility trends in Singapore,” in Gavin Jones, Paulin Tay Straughan and Angelique Chan (eds.), Ultra-low Fertility in Pacific Asia: Trends, Causes and Policy Issues. London: Routledge. Suzuki, Toru. 2009. “Population policy in Eastern Asian low fertility countries,” paper presented at the IUSSP 26th International Population Conference, Marrakech, Morocco, 2 October. Tsuya, Noriko, Larry L. Bumpass, Minja Kim Choe, and Ronald R. Rindfuss. 2005. “Is the gender division of labour changing in Japan?,” Asian Population Studies 1(1): 47–68. Tsuya, Noriko, Minja Kim Choe, and Wang Feng. 2009. “Below-replacement fertility in East Asia: Patterns, factors and policy implications,” paper presented at the 26th IUSSP International Population Conference, Marrakech, Morocco, 2 October. Wang, Feng. 2010. “China’s looming demographic crisis: Causes, prospects, and policy options,” paper presented at the Annual Meeting of the Population Association of America, Dallas. Yap Mui Teng. 2009. “Ultra-low fertility in Singapore: Some observations,” in Gavin Jones, Paulin Tay Straughan, and Angelique Chan (eds.), Ultra-low Fertility in Pacific Asia: Trends, Causes and Policy Issues. London: Routledge. Zheng, Zhenzhen. 2010. “Below replacement fertility in China: How important is the one child policy?,” paper presented at the Annual Meeting of the Population Association of America, Dallas.