As with most settings in Asia, mid-twentieth-century Hong Kong was characterized by high fertility rates, with TFRs of above 4.0 births per woman and crude birth rates of around 35–39 per 1,000 until the late-1950s (Tu et al. 2007). As Fig. 4.2 shows, Hong Kong has seen a dramatic decline in its total fertility rate (TFR) since the 1960s, reaching a low point of 0.9 children per woman in 2003. Hong Kong currently vies with the Republic of Korea and Taiwan for having the lowest TFR in East Asia (UNPD 2013).

As can be seen in the inset in Fig. 4.2, there has been a notable increase in TFR since 2003, most likely related to an increase in fertility among older women. Like many settings across East Asia and Europe, there has been a significant rise in the postponement of childbearing in Hong Kong. The mean age of women at first birth increased from 25.5 in 1976 to 28.7 in 2001, and at second birth from 27.8 to 31.4 over the same period (Tu et al. 2007). Crucially, the marital fertility rates for women aged 30–49 increased from 28.3 per 1,000 in 2001 to 43.8 in 2011 (HKCSD 2012d). This has clearly played an important role in shaping the recent upturn in period total fertility rates.

At the same time, there has been a clear shift away from the absolute primacy of the two-child family in Hong Kong. From the cohorts born in the mid-1950s through to those born in the mid- to late-1960s, the share of two-child families has declined from 45 to 35 % (Frejka et al. 2010). As Fig. 4.3 demonstrates, a concomitant result of this has been the rise of childlessness and of the one-child family. High-parity births have all but disappeared; of the 865,224 births registered in Hong Kong between 1995 and 2009, 7.7 % were parity 3, 1.4 % were parity 4, and 0.4 % were parity 5 or above.

A direct consequence of the decline in the first-parity fertility rate is a significant increase in the number of people age 50 or above without any children at all. For women born in 1946, the childless rate was just 2.6 %. This rose to 9.3 % and 16.6 % for the 1951 and 1956 cohort respectively. Tu and colleagues (2007) estimate that the childless rate for the 1961 birth cohort may reach 20 %, while Frejka et al. (2010) estimate childlessness among the cohorts born in the mid- to late-1960s to be as high as 35 %. As Fig. 4.4 demonstrates, these numbers are substantially higher than elsewhere in East Asia (Frejka et al. 2010). Within a familial reductionist social-policy context, where filial obligation still underpins notions of care for the elderly, unless these childless cohorts are able to set aside considerable personal savings, they could run into difficulties in securing support in later life. Furthermore, a number of recent studies have identified a series of negative outcomes for childless Hong Kong populations, such as an increased tendency to depression and loneliness (Cheng et al. 2014; Chou and Chi 2004).

We should, however, pause to consider the implications of a shift toward a growing number of (voluntarily chosen) one-child families. Firstly, it is possible to learn from the Chinese so-called “4-2-1 issue,” where a couple made up of only children have to bear the responsibility for both of their sets of parents (in the absence of any kin support) as well as their child. Of course, the burden of care would likely fall disproportionately on the woman. This has been cited as a reason for possible long-term low levels of fertility in mainland China (Basten and Gu 2013). Given the rise in the number of only children in the current marriage market, as well as an increasing preference for only having one child among current young people in Hong Kong (Basten 2013c), this burden of care has the potential to be a drag on future fertility—especially in the absence of significant improvements in the availability of (affordable) public private assistance for the elderly.

The relationship between Hong Kong and mainland China is instrumental in shaping the population dynamics of the territory. During the period 1981–2001—an era of very low fertility rates in Hong Kong—the territory’s population increased from 5.18 million to 6.27 million, primarily driven by the high number of entrants from mainland China rather than by natural increase (HKCSD 2001). Indeed, it is this dynamic which has largely shaped the recent trend in live births in the territory (see Figs. 4.5 and 4.6). As Basten and Verropoulou (2012, 2013) have observed, this increase in live births has been largely caused by an increase in what they term “maternity migration.” In July 2001, the Court of Final Appeal ruled that babies born in Hong Kong to Chinese nationals had the right of abode in Hong Kong. In 2001, just 620 babies were born to mainland women whose spouses were not Hong Kong permanent residents. By 2011, this figure had mushroomed to 35,736 (HKCSD 2012d).

Figure 4.6, which depicts the proportion of births by mother’s residency, shows the increasing proportion of births ascribed to what the registers term “transient” mothers from the mainland. Basten and Verropoulou (2012, 2013) observe that there are a number of “push and pull” factors associated with this increase. For one thing, the right of abode brings certain key privileges, such as enhanced access to the territory’s educational system and visa-free travel to many more countries than granted for mainland PRC citizens. Basten and Verropoulou (2013) have also demonstrated that Hong Kong has served as an outlet allowing mainland couples to evade the PRC’s family-planning restrictions. Upon close inspection of the Hong Kong microdata, the vast majority of “transient” mothers, were giving birth to their second child in the territory. Furthermore, these second (and third) children were disproportionately male. This resulted in a highly skewed sex ratio at birth—something which, historically, was generally absent in Hong Kong. Finally, the average educational attainment level of such “transient” mothers was generally high, indicating a higher level of income.

Only births to mainland women whose partners are Hong Kong permanent residents (so-called “Type I” babies) are included in the calculation of the TFR by the Hong Kong Census and Statistics Department (HKCSD). Compared to “Type II” babies (born to mainland women whose partners are not Hong Kong permanent residents), such births increased only from 7,190 in 2001 to 9,879 in 2005 before falling back to 6,110 in 2011 (HKCSD 2012d). While the growth of maternity migration has been critical in shaping the number of births in Hong Kong, it has had relatively little effect on the recent increase in the officially reported total fertility rate because “Type II” babies are not included in the calculation of TFR. As noted above, changes in the TFR have predominantly been driven by the tempo effect of postponement of births. That is not to say that maternity migration is irrelevant to a policy discussion of population in Hong Kong: As Basten and Verropoulou (2013) observe, while almost all “Type II” babies return to the mainland with their mothers immediately after birth, it is almost impossible to identify the point at which they may return to Hong Kong—if at all. This means that planning for future provision of public services and housing and, indeed, forecasts of the size of the labor force, must take into account the temporal uncertainty of “return” migration. This situation forms part of the broader difficulty in forecasting population dynamics for Hong Kong due to its border with the mainland.

In order to understand the reasons for fertility decline in Hong Kong, it is perhaps useful to differentiate between what might be termed the proximate, or demographic reasons, and the cultural-social reasons.

According to various studies, the underlying reasons for this “retreat from marriage and motherhood” are two-fold: incompatibility between women’s economic and domestic roles, and a squeeze in the marriage market (see, among many examples, Basten 2013c; Chang 2003; Frejka et al. 2010; Song et al. 2013; Jones et al. 2009; Sun 2012). The public roles of women have been revolutionized in advanced Asian economies to the extent that in many areas younger women are better educated than men, thus increasing—and enhancing the quality of—female labor-force participation (Esping-Andersen 2009; MacDonald 2000). Developments in the private sphere, meanwhile, have been much slower and more resistant to change. Entry into a married relationship where the onus is still frequently firmly placed upon the woman to care for children, parents, and parents-in-law, as well as look after the household, is often unappetizing given the extraordinarily high opportunity costs.⁵ This so-called “Incomplete Gender Revolution” is held to be a key element in the avoidance of the “marriage package” of home and children (Esping-Andersen 2009).

Related to this is a squeeze in the marriage market. Although female public roles have changed dramatically, systems characterized by women marrying older and/or better-educated men (possibly of a higher social status) are arguably breaking down much more slowly (Jones and Gubhaju 2009). As such, increasingly educated women find themselves squeezed out of a marriage market privileged males “marrying downwards.” The corollary of this, of course, is that many less-educated males also struggle to find partners. This pattern has been cited as a major cause of the increase in cross-border marriages elsewhere in East Asia (Jones and Shen 2008; So 2003).

Housing in Hong Kong is extremely expensive and has, in a recent economic analysis, been suggested to have a causal effect on Hong Kong’s low fertility (Yi and Zhang 2010). A recent survey of 360 cities found that, with average home prices at 14.9 times the gross annual median household income, Hong Kong was the most unaffordable city in the world to buy property (Holliday 2014). The same survey found that Singapore’s house prices were 5.1 times the gross annual median household income. A commercial survey of 30- to 40-year-old middle-class individuals found that the high cost of property had an effect on personal decisions, with 53 % of the single respondents saying they would delay their marriage to have more time for saving. Nearly half of those who were married also chose to delay having children, or to have no children, so they could boost their savings. Tellingly, the survey found that “property purchase remained the top reason for savings, followed by travel, having children, and further education” (reported in Chen 2013a). This is despite the fact that almost one-half of all housing in the SAR is either “Public Rental Housing” or “Subsidized Home Ownership Housing” (HKHA 2013).

Finally, on a macro-level, it is worth remembering that Hong Kong has one of the highest population densities in the world, at more than 6,500 people/km². After taking into account that more than one-half of the SAR is extremely sparsely inhabited because of its mountainous terrain, the figures for population density in the other districts are even more striking. For example, in the Kowloon area, with more than two million people, the population density is 44,917 people/km², while for Hong Kong Island as a whole the total is 15,924 people/km² (even though, again, much of the southern part is very sparsely populated due to its topography) (HKCSD 2012c). For comparison, the population density of Manhattan is 10,194 people/km², while in Tokyo Prefecture it is 6,029 people/km² (Tokyo Metropolitan Government 2014).

Long working hours and the negative impact upon work-life balance have been widely cited as a drag on fertility in East Asia (McDonald 2006). The average contracted working hours in Hong Kong are 45.2 h per week, ranging from 40.6 h for the lower quartile (LQ) to 49.7 h for the upper quartile (UQ). Clearly, however, overtime forms a critical part of the structural labor system in Hong Kong. On top of contracted working hours, lower-skilled workers tend to work an average of 7.0 h (LQ: 2.0; UQ: 9.7) paid overtime, while in more professional sectors between one-fifth and one-third of workers work an average of 8.0 h (LQ: 5.0, UQ: 10.0) unpaid overtime. Overall, a survey conducted in 2011 by the Hong Kong Labour Department found that the average number of hours worked by full-time employees was 49.0 (HKLD 2012, p. 129)—compared to a range of 39.1–43.7 h among member countries in the European Union. Around 10.3 % of people work 60–70 h per week, while around 3.0 %—mostly double-shift security guards—work more than 70 h per week (HKLD 2012, p. 130). It should not, however, be assumed that this only relates to employees working in corporate offices until late in the evening, as the East Asian stereotype might suggest (see Fig. 4.7). Closely related is the notion of a “24-hour city” with very long opening hours for shops, restaurants, and services. In this context, the rapid growth since 2001 in consumption-related employment among highly skilled workers (+28 %) and service and shop sales workers (+24 %) is also important (HKLD 2012, p. 110). The average contracted hours for those employed in hotels and restaurants is above 50 h per week, while in elderly homes it is 54.8 h (LQ: 45.0; UQ: 66.0), and in estate management and security it is 57.5 (LQ: 48.0; UQ: 69.2).

Finally, at just 7 %, Hong Kong has a very small proportion of the population in part-time employment (compared to an average of 14 % for developed countries that are members of the Organisation for Economic Cooperation and Development—OECD) because of the demand for full-time workers in a liberal labor market without working time regulations. While part-time employment has very different meanings in different parts of the world, even within low-fertility settings in East Asia such employment is more prevalent in Taiwan (8 %), Singapore (10 %), Korea (11 %), and Japan (20 %) than in Hong Kong (HKLD 2012, pp. 113–14). Such an emphasis on full-time employment and long working hours can create tremendous pressure, particularly on women, and works strongly against combining work and childrearing.

Even the Hong Kong government is pessimistic about a medium-term increase in period total fertility rates in the territory.⁸ In their latest round of population projections, the HKCSD assumes that TFR will decrease from 1.204 in 2011 to 1.177 by 2016, then fluctuate at around 1.190 until 2031, before falling to 1.164 by 2036 and 1.151 by 2041 (HKCSD 2012b). Such tiny changes in fertility are, fundamentally, irrelevant, but the key message is one of assumed stagnation.⁹ Unlike many other local statistical offices, the process of formulating the Hong Kong Census and Statistics Department’s fertility assumptions “is not strictly a mechanical one that follows the extrapolated trends” (HKCSD 2012b). The overall pattern of marital fertility is projected to stay relatively constant to the end of the projection period (2041).¹⁰ Most significantly, however, the department assumes a continuing pattern of later and fewer marriages. The Hong Kong government projects that by 2041 around one-third of all women will remain never married at the end of their childbearing period, compared with around one-eighth in 2012. If the link between marriage and childbearing stays strong, this will have a negative impact upon both the tempo and quantum of childbearing.

A second way to consider short- to medium- term prospects for fertility in Hong Kong is to examine reported fertility preferences—both as a micro-indicator of family formation and as a macro-indicator, or “barometer,” of general attitudes towards childbearing at the societal level. Using evidence from a number of rounds of the Hong Kong Knowledge, Attitude, and Practice (KAP) Surveys, we can examine lifetime fertility ideals through the question “If you could start all over again, how many children would you like to have in your lifetime?” As Table 4.1 demonstrates, a majority of women questioned in 2007 desired two or fewer children (88.4 %), with around one-half (49.7 %) stating a preference for two children (HKFPA 2007). Regression analysis performed by the Hong Kong Family Planning Association on data from six rounds of the survey over 25 years demonstrates that an ever-increasing number of women state a preference for zero or only one child, while the proportion desiring three or four children is significantly decreasing. This evidence of aggregate fertility ideals well below two represents a paradigm shift compared with the results of similar surveys in Europe.