This is the core question pertaining to the fertility transition and its mechanisms, and, in a broader perspective, to the (first) demographic transition. To date, there is considerable disagreement about the extent to which the fertility transition is driven by structural and socio-economic forces, or by technological and ideational changes (Bryant 2007). Michel Poulain, Dany Chambre, Pino Ledda, and Anne Herm address the main premise of the demographic transition theory: i.e., that a decrease in infant and child mortality precedes the onset of fertility decline, and that mortality decline is an important driving force of fertility change. Poulain et al. provide an in-depth investigation of marital fertility transitions in two villages in Sardinia. Their study, covering a period from the late 19^th century to the present time, concludes that the close association between the decreases in mortality and fertility indicates that the decline in stillbirths and infant mortality in the studied populations “might have been at least partly responsible for the corresponding decrease in fertility” (p. 231). At the level of individual families, the “number of surviving children, rather than a specific number of births, serves as an important predictor of the likelihood of having another child” (p. 232). However, the causal link between mortality and fertility appears to run in both directions: longer birth intervals and a lower number of children born – two manifestations of the progressing fertility transition – are also found to be associated with better chances of children surviving, and, thus, with lower early mortality. In addition, broader socio-economic and cultural developments – including improvements in health care, well-being, transport links and education – are shown to be associated with both declining mortality and family size limitation.

The contemporary debate on the fertility transition often focuses on higher-fertility countries, especially in Sub-Saharan Africa, and the drivers of fertility change in the region. Two important contributions, by Bongaarts and Casterline (2013) and by Bongaarts (2017), examine the unique features of the African fertility transition. They document that fertility decline in Sub-Saharan Africa, once initiated, has been slower than in other world regions (see also Shapiro and Hinde 2017), and that the ideal family size remains higher there than it is elsewhere. In addition, women in many parts of the continent have poor access to modern contraception and high rates of unmet need for contraceptive use. While the correlations between the main indicators of socio-economic development (GDP per capita, education, life expectancy and urbanisation) and fertility in the region have the expected negative sign, Bongaarts (2017: 55) highlights what he calls the Africa effect: at ”a given level of development, Africa’s fertility is higher, contraceptive use is lower, and desired family size is higher” than in the less developed countries outside the continent. Many societies in Sub-Saharan Africa are characterised by a strong pro-natalist culture that sustains higher family size preferences.

David Shapiro and Andrew Hinde contribute to the debate on the factors driving the fertility transition in the region. They examine fertility change and broader socio-economic and demographic factors affecting fertility in seven countries of Sub-Saharan Africa that have the highest fertility rates globally, and are the “laggards” in the global fertility transition: Chad, the Democratic Republic of Congo (DRC), Equatorial Guinea, The Gambia, Mali, Niger and Somalia. While all of these countries have experienced the onset of fertility decline, their fertility levels have declined slowly. Indeed, by 2010–2015, five of these countries had not even met the criterion of a 10 per cent decline from the peak level that is often used as a benchmark for fertility transition. Shapiro and Hinde document that the analysed countries are experiencing a combination of social, economic and cultural factors that are sustaining their higher fertility levels and slowing their fertility declines. In comparison with other countries in the region, they display much lower levels of modern contraception use – especially among women in a union – earlier marriage, higher ideal family sizes (reaching a staggering 9.2 children per woman in Niger; see Table 2 in Shapiro and Hinde, this volume), fewer years of schooling and higher shares of the population with no education (except in the DRC); and, in most countries, lower GDP per capita and higher infant and child mortality. All of the analysed countries except the DRC have a high proportion of Muslims in the population, ranging from 65% in Chad to 100% in Somalia, and thus reflect the stronger preference for children among Muslim women. Taken together, these factors form a “multi-dimensional space” associated with “fertility promoting values” as well as “resistance to fertility decline”. The current high fertility rates in these countries build up momentum for continued rapid population growth in the coming decades, posing challenges to improvements in well-being, and contributing substantially to future population growth in Sub-Saharan Africa.

Future fertility trends in the “laggard countries” and in the region as a whole will also be affected by violence and governance failures. Alicia Adsera’s debate contribution brings up the “lack of leadership” as one of the key forces hampering broader socio-economic development, and thus slowing down the pace of fertility decline. It is no coincidence that a number of the countries analysed by Hinde and Shapiro are experiencing violent ethnic and religious conflicts (Chad, DRC, Mali, Niger, Somalia), and their governments do not have full control over the whole territory. Most of these countries have dysfunctional and weak governments that are not able to deliver even basic infrastructure, education, health care and security to their populations. In the countries that are managing to expand education, especially for girls, Wolfgang Lutz (this volume) highlights the dual effect of education on lowering fertility: it “reduces the ideal family size, presumably through changes in values and in the associated quantity/quality considerations; and it empowers women to actually realize this smaller family size by averting unwanted births” (p. 64).

Once the fertility transition is under way, is it likely to proceed without interruptions until fertility reaches low levels? Or does fertility decline often follow a less straightforward trajectory marked by stalls and reversals? Which factors sustain the transition to low fertility? Studies published since the early 21^st century have examined the rising number of countries in Africa, Asia and Latin America where the shift towards lower fertility has at least temporarily stalled, and have discussed the factors contributing to these stalled transitions. Looking at seven countries across these three continents, Bongaarts (2006) concluded that these stalls in the fertility transition were linked to stagnating contraceptive use, and to a stabilisation in the demand for contraception and in the number of wanted births, with no evidence of rising unwanted or unplanned births. A number of studies have focused on the evidence of stagnation in the fertility transition in Africa in the 1990s and early 2000s. Bongaarts (2003) and Shapiro and Gebreselassie (2008) highlighted the role of educational improvements in the course of the fertility transition, observing that less educated women tend to have both a higher (although declining) desired family size and higher levels of unwanted fertility. They suggested that low or stagnating levels of schooling can cause the fertility transition to stall. In line with these expectations, Goujon et al. (2015) and Kebede et al. (2019) linked the stalling of the fertility transition in some countries of Sub-Saharan Africa to the setbacks in education in the 1980s and 1990s, when the share of women without formal education stagnated or declined only slightly in Kenya, Nigeria, Côte d’Ivoire and some other countries. At the same time, fertility rates among women with no education remained high or even increased, fuelling the overall stagnation or slow-down of the decline in period fertility.

The current evidence suggests that the experience of fertility stalls or reversals is relatively common in the countries undergoing the fertility transition, with some societies experiencing a prolonged period of stalling fertility at higher levels lasting two or more decades. Al Zalak and Goujon’s (2017) analysis of increasing fertility in Egypt between 2005 and 2015 also showed that changes in the timing of marriage and childbearing could explain some of the fertility stalls. During that period, more women married at an early age (before age 25), and the previous gradual shift towards later marriage and childbearing had come to an end.

Although the fertility transition and the increase in educational attainment often lead to a shift towards later marriage and parenthood (Bongaarts et al. 2017), this trend is often uneven, and is far from universal among the transitional societies. Three of the Central Asian countries that were formerly part of the Soviet Union – Kazakhstan, Kyrgyzstan and Turkmenistan – experienced an upturn in their period fertility rates, and a shift towards earlier childbearing or a stabilisation in their patterns of early births in the 2000s and 2010s (Spoorenberg 2015). The study by Vladimir Kazenin and Konstantin Kozlov puts a spotlight on the absence of first birth postponement in Kyrgyzstan. The authors look at the socio-economic and cultural factors that have been sustaining the pattern of early union formation and childbearing in these countries. Their research underscores the role of the “expected” factors (especially lower education and rural residence), but also the importance of cultural and technological factors, such as access to information (limited access to the internet); acceptance of domestic violence as a measure of women’s lack of power; and ethnicity, which is closely linked to religiosity and “traditionalism” in family relations. Women from the Uzbek ethnic minority, who have the lowest levels of educational attainment and internet access and the highest levels of acceptance of domestic violence among the three groups analysed, also display a pattern of early union formation and of a faster transition to the first birth after union formation. This analysis underscores that culture, religion and women’s autonomy continue to affect fertility change and variation.

In the last three decades, many countries in Asia and Latin America – including Bangladesh, Brazil, China, Colombia, Iran, Malaysia, Turkey and Viet Nam – experienced a decline in their period TFR to sub-replacement levels, thus marking the completion of their fertility transition. A locus of low fertility has been increasingly moving towards middle-income countries. As ever more countries are set to join the “low-fertility club” in the near future, it is worth looking at the experiences of the countries that completed their fertility transition in the 1960s–1990s.

What are the key emerging features of post-transitional fertility? Arguably, the most important finding is the absence of a stabilisation of fertility around the replacement level. As Wolfgang Lutz (this volume) puts it, “we are groping in the dark and only feel confident in saying that the long-held view that all countries will ultimately converge to the so-called replacement level of 2.1 is untenable and without scientific basis” (p. 65). In most countries, period fertility continues to decline once this threshold has been hit. Over the last decade, all post-transitional countries with relatively higher fertility have experienced a decline in their period TFR, with most of them seeing their TFR drop from around 1.9-2.1 to around 1.7 (Figure 1). However, some of these countries saw a sharper fall in fertility: in 2019, the TFR plunged to 1.53 in Norway, and to a record low of 1.35 in Finland (Rotkirch, this volume).

These sharp declines are most unexpected in the Nordic countries, which are often seen as forerunners in implementing work-life balance policies (Lappegård this volume; Lutz this volume) that enabled them to achieve a “virtuous combination” of high levels of women’s employment, high gender equality and relatively high fertility. While the recent fertility declines have been driven in part by a renewed trend towards delayed family formation, the spread, the magnitude and the duration of these fertility downturns indicate that they are likely to be more entrenched, and to lead to declines in cohort fertility as well (Hellstrand et al. 2020b for the Nordic countries).

With the possible exception of China (Basten and Gu 2013), the very low fertility observed in many post-transitional countries today does not reflect the reproductive intentions, desires and ideals of individual men and women. In Europe, the United States, Japan, Canada and other low-fertility countries, the indicators of family size preferences typically reach around or slightly above two births on average, and have remained relatively stable over time (Sobotka and Beaujouan 2014; Beaujouan and Berghammer 2019; Edmonston et al. 2010; Fukuda and Saotome 2018; see also Esteve et al. this volume). Among younger women across educational groups, including women with a university degree, the intended or desired family size is generally around two children (Beaujouan and Berghammer 2019; Berrington and Pattaro 2014). In addition, the share of younger respondents who express an ideal or an intention to be permanently childless remains low in most countries (e.g., Rybińska (2020) for the United States; Esteve et al. this volume). This evidence suggests that one of the key explanations for low fertility rates is the inability of women and men to realise their fertility plans, with the intentions-behaviour gap typically being larger in countries with very low fertility (e.g., in Italy, Greece and Spain) and among women with higher education (Beaujouan and Berghammer 2019). Albert Esteve et al. (this volume) argue that the gap is built over the life course, and is partly explained by less frequent and later partnership formation (see also Section 3 below).

In many countries, post-transitional fertility has been unstable, characterised by distinct waves of fertility declines and upturns over time (Sobotka 2017). These ups and downs of “rollercoaster fertility” are largely driven by changes in the timing of parenthood in response to economic shocks, changes in family policies and other factors; and do not affect the completed fertility rate, which generally follows a much more stable trend. As a result, period TFRs often send the wrong signal about the direction and the pace of fertility change, as well as about the differences in fertility levels between countries and populations (Sobotka and Lutz 2011). This fallacy of overinterpreting short-term period fertility shifts is also found in some contributions studying the changing links between selected determinants of fertility and fertility rates (e.g., Myrskylä et al. 2009).

Another key feature of post-transitional fertility is a shift towards entering parenthood at later reproductive ages that has been progressing in many countries without interruption since the 1970s (Sobotka and Beaujouan 2018). Anna Rotkirch (this volume) highlights the complex nature of fertility decisions today, which involve a much higher degree of ambivalence about when “it is safe and desirable to have children” and a longer list of “preconditions” for parenthood that have to be fulfilled before men and women are ready to have a child. Among these preconditions, finding the right partner and being able to live independently score especially high. Esteve et al. (this volume) argue that the timing of partnership formation is one of the key determinants of contemporary fertility differences. This link is most apparent in East Asian countries, where marriage remains a precondition for parenthood (Fukuda this volume). The rapid expansion of tertiary education, especially among women, and the increasing involvement of women in the labour force, have contributed to a sharp rise in the mean age at first marriage across the region (Cheng this volume, Fukuda this volume).

The ever-later timing of entry into parenthood, in combination with a rise in lifetime childlessness, indicate that women and men are spending increasing shares of their reproductive lives without children. For instance, in Japan and Spain, close to 40% of women currently remain childless at age 35 (Human Fertility Database 2020), a threshold at which infertility starts to increase (Leridon 2008). To account for both the changes in childlessness and the shift towards later parenthood, Ryohei Mogi and Michael Dominic del Mundo propose a new index of the average length of life women spend without children during their reproductive period (15–49): the Expected Years Without Children (EYWC). Focusing on women in eight countries born in the 1940s to the early 1960s, they show that first birth postponement was the main driver behind the rise in the EYWC in Canada, the Netherlands, Norway, Sweden and the United States.

The increase in childbearing after age 35 has been fastest among childless women and women with one child. This implies that these women are at higher risk of being unable to realise their fertility plans due to infertility (e.g., Beaujouan et al. 2019). As Esteve et al. (this volume) highlight, fertility decisions are increasingly constrained by the ages of the prospective parents. Postponed parenthood is also linked to a rising reliance on assisted reproductive technology (ART), which is becoming an important means for achieving pregnancy among many women and couples past age 35. For instance, in Japan, the share of children born to women who had ART treatment jumped from 1.0% in 2000 to 2.7% in 2010, and to 6.0% in 2017 (own computations based on Ishihara et al. 2019). In Spain, 8.6% of all births in 2018 were to women who underwent ART treatment (computations based on Table 3.17 in Registro SEF 2018).

Within the current variation in post-transitional fertility trends, demographers are especially concerned about the countries and regions with “ultra-low” or “lowest-low” fertility rates. The question of “how low can fertility fall” is both fascinating and relevant from the standpoint of governments, societies and policy-makers dealing with the likely long-term consequences of depopulation and population ageing. While some studies have tried to empirically determine the lower bound of period and cohort fertility (Golini 1998), a more plausible view suggests that there is no obvious limit to low fertility. At least from a period perspective, zero is the ultimate limit once reproduction is culturally determined (Lutz this volume). Equally important, also from the perspective of population projections, is the question of how long extreme low fertility is likely to last: Is it a transient phenomenon, merely driven by a temporary shift towards later parenthood, or is it likely to become a permanent feature of many societies? Lutz et al.’s (2006) low fertility trap hypothesis argues that in societies with a long-standing experience of very low fertility, the family size ideals of younger generations socialised in small families and in an environment with few children around could change, leading to a spiral of declining numbers of births, low family size ideals and very low fertility.

Currently, two broader regions, Eastern Asia and Southern Europe, are the “hotspots” of global low fertility, with the Republic of Korea (South Korea) achieving new lows in its period TFR almost every year: the TFR in the country dropped to 0.92 in 2019, and reached the even lower level of 0.72 in the capital city of Seoul (KOSIS 2020).¹ What are the characteristics of countries with very low fertility? Table 1 summarises the selected features of period and cohort fertility in seven countries and territories in Southern Europe and East Asia: Greece, Italy, Spain, Hong Kong, Japan, South Korea and Taiwan. In each of these countries, the tempo-adjusted indicator of period fertility in 2016 and the completed fertility of women born in 1978 were above the recent conventional period TFR levels, signalling that the “ultra-low” fertility levels are partly fuelled by the ongoing postponement of parenthood. At the same time, the tempo-adjusted period fertility rates and the completed fertility rates in these countries generally reached levels of around 1.4–1.5. Thus, regardless of which indicator is used, fertility in these countries is low, with the completed fertility rate reaching the lowest level globally in Hong Kong (1.25), followed by Spain (1.36). The low mean family size in these societies is due to a combination of a high share of women remaining childless (reaching 28% in Japan and 30% in Hong Kong) and a high share of women having only one child (see also Zeman et al. 2018). Southern European and East Asian societies are also characterised by a pattern of late parenthood, with the mean age at first birth among women surpassing 30 years, and reaching close to 32 years in South Korea.

What are the key factors contributing to “ultra-low fertility”? Alice Cheng (this volume) provides a review of the institutional factors that underlie very low fertility in East Asia, with a particular focus on Japan, Hong Kong, South Korea and Taiwan, but also incorporating data and literature on China (see also Raymo et al. 2015). Cheng highlights the historical and cultural roots of ultra-low fertility in the region, with traditional Confucian ideology placing a strong emphasis on formal academic attainment. Moreover, Confucianism promoted a patriarchal view of male and elder dominance, and stressed women’s responsibilities to provide care as part of their roles as wives, mothers, and daughters. Female empowerment in education and employment, together with long work hours and rigid labour market structures (Brinton and Oh 2019), clash with these traditional norms and expectations: the “patriarchal cultural and family systems of East Asia are maladapted to women’s rising status and economic independence” (Cheng, this volume: 84).

An important feature of East Asian cultures is their low acceptance of unmarried cohabitation, of non-traditional family forms and, even more so, of childbearing outside marriage (Fukuda this volume; Cheng this volume; Raymo et al. 2015). The prevailing gender inequalities, the rising economic uncertainty, the high direct and opportunity costs of childbearing – especially for women – and the notion that marriage and childbearing are “too stressful” and involve “too much personal sacrifice” (Cheng this volume), have led to both the erosion and the postponement of marriage. Cheng argues that “the social mentality and the prevailing norms about marriage and childbearing are incompatible with the rising agency of women” (p. 110). Stuart Gietel-Basten (this volume: 40) refers to the “cultural and institutional grip that older generations have on the young”, which sustains gender inequality. As childbearing is exclusively linked to marriage, less frequent marriage, later marriage and the rising instability of marriage have negatively affected fertility rates in East Asia (Fukuda this volume). In addition, the strong focus on the academic achievements of children has increased the investments of time and money parents are expected to make in their children’s education, including in extended extracurricular activities. This “educational arms race” results in lower fertility intentions and reduced fertility (Cheng this volume).

By contrast, in Southern Europe, the initially close link between marriage and childbearing has eroded fast. However, Southern European and East Asian societies share the trend towards a pattern of late independent living and partnership formation among young adults (e.g., Esteve et al. this volume). The late transition to residential independence is fuelled in part by extended education, but even more so by the high and rising costs of housing and the high levels of economic uncertainty among young adults, with many facing unemployment, or employment that is poorly paid or with a non-standard contract. While economic uncertainty has affected young adults and their fertility decisions across all highly developed countries, it has had the most “scarring” effects in South-Eastern, Eastern and Southern Europe (Adserà 2011), especially in the wake of the financial crisis of 2007-2012 (Comolli 2017; Matysiak et al. 2020). Southern Europe is also characterised by relatively underdeveloped family policies, with more government resources being spent on the elderly, and a large share of the burden of family care responsibilities being shouldered by kinship and family, and especially by women (Cooke 2009; Pailhé et al. 2019).

Over the past decade, numerous studies have discussed and evaluated the factors and institutional conditions that affect post-transitional fertility levels, fertility trends and their variation, while often proposing new conceptual frameworks. Five broader, partly interrelated factors – i.e., economic uncertainty, gender egalitarianism, changing educational and social stratification in fertility, the impact of migration and institutional support provided to families – feature prominently in these debates. Adverse economic and labour market conditions have been repeatedly shown to have a negative impact on fertility intentions (Fahlén and Oláh 2018; Busetta et al. 2019; Vignoli et al. 2020a) and fertility (Adserà 2011; Sobotka et al. 2011; Comolli 2017). Over time, the conceptualisation of uncertainty has widened beyond the impact of economic shocks. One strand of research has documented the impact of delayed economic independence, more precarious jobs, declining relative income and the general increase in pressures and uncertainties among the younger generations (Mills and Blossfeld 2005; Sanderson et al. 2013). Recently, Vignoli et al. (2020b) extended the concept of uncertainty to incorporate the “imagined futures”: i.e., perceptions and narratives of the future that may be disconnected from an individual’s current situation.

From a different perspective, Nicolò Cavalli (this volume) argues that demographic behaviour is strongly future-oriented, and that considering the role of “time preferences can add another theoretical micro-foundation to the mechanisms that link individual fertility decisions to institutional factors that operate at the macro level” (p. 257). His main argument, grounded in the concepts of hyperbolic (time) discounting, is built upon two key insights on decision-making: first, that individuals have a preference for earlier rather than later gratification, opting for rewards that are smaller, but occur earlier; and, second, that a conflict between current and future preferences results in people consuming more at present than initially planned (thus “discounting”, or penalising, future rewards). Cavalli hypothesises that societies that are more future-oriented are likely to show a lower “bias towards the present”, with families tending to invest more in the quality of their children, resulting in lower fertility rates. Empirically, this hypothesis is tested by constructing a Future Orientation Index (FOI) based on worldwide Google queries. The results of this analysis confirm that there is a negative correlation between future orientation and the total fertility rate, which flattens out at high levels of future orientation. It would be worth expanding this analysis to fertility timing: i.e., future orientation is likely to lead to more careful planning of births among couples, which may contribute to the fertility postponement observed in many countries.

The structural/economic perspectives on fertility have often neglected the impact of massive changes in women’s educational levels and economic achievements over the last half a century. Mills (2010) highlights the complexity of arguments linking different micro- and macro-level perspectives on gender equality and fertility, which focus on issues such as the division of household work and childrearing among couples, but also on institutional factors that lead to unequal opportunities and participation in education, employment and social and political life. Esping-Andersen and Billari (2015) and Goldscheider et al. (2015) link the broader trend towards gender equality both within households and in public institutions to the transformations in fertility patterns in the past decades. The gender revolution framework proposed by Goldscheider et al. (2015) suggests that an initial trend towards institutional gender equality, marked by a rapid rise in higher educational attainment among women and a sharp increase in women’s participation in employment, resulted in “less family” and low fertility, as women’s labour market activities remained incompatible with their family roles. This incongruence between rising opportunities and aspirations in the public sphere and rigid norms and expectations in the domestic sphere is especially sharp in East Asia today (Cheng this volume). According to Goldscheider et al. (2015), the “second half of the gender revolution”, which is driven by the spread of gender egalitarian attitudes and practices within the household, will eventually result in a return to “more family”, and, thus, to a rise in fertility and marriage rates. Furthermore, changes in the structure of economic opportunities and improvements in the compatibility of family and career for highly educated women, supported by family policies, could lead to a reversal of the negative association between education and fertility among women. Trude Lappegård (this volume) discusses these arguments with reference to the unexpected recent decline in fertility in the Nordic countries, which are characterised by high levels of both institutional and domestic gender equality. She argues that gender equality must be considered within a wider context of a “changing economy, globalization, and increasing economic uncertainty” (p. 45), and linked to systems of social stratification. Whether gender equality positively affects fertility will depend on the responses of different socio-economic groups (see also Section 3 below).

Across the low-fertility countries, the evidence on the changing social stratification of fertility is mixed. For the Nordic countries, Jalovaara et al. (2019) document that, with the exception of Finland, completed fertility has converged among women across different educational groups born in the late 1960s and early 1970s. In Denmark and Sweden, childlessness has risen sharply among women with low education, reaching much higher levels than among women with medium or high education. Past data for countries in Southern, Central and Eastern Europe, as well as for the United States, Australia and the United Kingdom, show the persistence of a negative education gradient in fertility among women born in the 1960s and 1970s (e.g., Berrington et al. 2015; Sobotka et al. 2017; Zang 2019; Gray and Evans 2019). One factor that continues to contribute to lower fertility among highly educated women is their late family formation. Women and men tend to postpone the first birth until after they finish their studies, and educational expansion is the main driver of the trend towards late parenthood across the highly developed countries (Ní Bhrolcháin and Beaujouan 2012).

In the context of continuing education expansion, Michaela Potančoková and Guillaume Marois evaluate the importance of accounting for women in education in their microsimulation projection model for 28 European Union countries (including the United Kingdom): when educational enrolment is included, the projected EU-wide TFR is lower than it is in the scenario that accounts for educational attainment, but not enrolment. Their analysis reveals, however, that migration will be the major source of heterogeneity, as it will have strong effects on fertility, the number of births, population size and population distribution across the European Union. The impact of future migration is most clearly illustrated in the high immigration scenario, which assumes that there will be a doubling of the inflow of immigrants from outside of the European Union, roughly corresponding to current levels of immigration to Canada. When contrasted with the scenario in which recent (2010–14) levels of immigration continue, the high immigration scenario would result in a larger female population of reproductive ages, which would, in turn, lead to higher fertility and a steadily increasing number of births, reaching 34.9 million in 2055–59 (vs. 24.8 million in the baseline scenario).

From a different perspective, Alicia Adsera (this volume) argues that return migrants and transnational migrants may contribute to fertility declines in their countries of origin. The flow of financial resources via remittances should ease the pressure on local families to put their children to work, and encourage them to invest in their children’s education. At the same time, cheap and accessible communication technologies, including social media, should speed up the transmission of information, and foster new norms about family planning, women’s empowerment and smaller family size ideals.

The impact of uncertainty, gender equality, migration and social stratification on fertility is altered by family-related policies and the overall character of the welfare state – which, in turn, affect educational and labour market systems (Neyer 2013). Across the low-fertility countries, governments have been investing more resources in families. However, the evidence on the effects of family policies on fertility is mixed, suggesting that their main impact is on the timing and spacing of births, often fuelling short-term cycles of baby booms and busts (Gauthier 2007; Thévenon and Gauthier 2011; Luci-Greulich and Thévenon 2013). Especially questionable are the effects of family policies designed with explicit pronatalist targets, which are often poorly conceived; have short-term aims; and are usually based on simple indicators affected by changes in the timing of births, especially the period TFR (Sobotka and Lutz 2011; Sobotka et al. 2019). In countries with comprehensive family policies, particular attention has been paid to policies supporting the reconciliation of work and family (OECD 2011), which are most beneficial to highly educated women (McDonald and Moyle 2019). Family policies might have neglected some of the needs and preferences of less educated women and men, who often have low-paid and unstable jobs, as well as more complex family trajectories and configurations (see also Lappegård this volume). In most countries, family policies have not yet fully “caught up” with the current diversity of family forms and living arrangements, and there is a need to extend more social protection rights and welfare benefits to children in non-traditional family configurations (Miho and Thévenon 2020).