The Oxford Research Encyclopedia of Global Public Health has moved behind the paywall. For information on how to continue to view articles visit the how to subscribe page.
Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Global Public Health. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 29 January 2023

The Demography of Fertilityfree

The Demography of Fertilityfree

  • Visseho AdjiwanouVisseho AdjiwanouSociology, Universite du Quebec
  •  and Ben Malinga JohnBen Malinga JohnPopulation Studies Department, University of Malawi Chancellor College


From the first billion people in the world in 1800 to the projected 9.7 billion people in 2050, the world’s population has passed through various stages. However, the different stages have not been the same for each global region or for every country within the same region. On one side of the spectrum is the fertility transition in Europe and North America, where the decline has been steady, with the median total fertility rate (TFR) declining from 2.80 children per woman in 1950–1955 to 1.66 in 2015–2020. In this region, childbearing is no longer the final goal of marriage, and this change has been accompanied by the emergence of new forms of union. The fertility rate is below the level of replacement in almost all the countries. On the other side of the spectrum is sub-Saharan Africa, where fertility has declined slowly and has stalled in various countries since the 2000s. The median TFR in the region declined from 6.51 children per woman in 1950–1955 to 4.72 in 2015–2020. In this region, this trend is associated with slower increase of the age at first marriage and in of the modern contraception. The fertility transition and its associated factors in the other regions of the world fluctuate between these two scenarios.


  • Sexual & Reproductive Health


From the first billion people in the world in 1800 to the projected 9.7 billion people in 2050, the world’s population has passed through various stages. As the fertility rate has started to decline in various part of the world, it has ignited fears and incomprehension, as well as optimism. However, although the demographic transition (which indicates the shift from a traditional regime where fertility and mortality are high and roughly in balance to a regime where birth and mortality rates are low and equally balanced) has spread to nearly all regions of the world, it has been slow to materialize in sub-Saharan Africa. Although fertility is declining, it is still very high in this part of the world. Indeed, between 2015 and 2050, the world’s population is expected to increase by 2.4 billion, and half of the increase will occur in sub-Saharan Africa. This demographic growth is, and will continue to be, characterized by a high demand for social services, a very high dependency ratio that limits the development possibilities of many countries in the region, and potential instability at the national and global level. Nevertheless, it is also a source of economic progress as fertility continues to fall, creating what is called a demographic dividend (May & Guengant, 2020). It is important that the decline in fertility be continuous and sustained for the opportunity to materialize that permits qualitative investment in social services (education, health, etc.). Creating employment opportunities and enhancing good governance are also critical to realizing the demographic dividend.

In other areas of the globe, particularly in developed countries, the first transition has ended and has continued with a second transition triggered by the decline in fertility to below 2.1 children per woman, or replacement-level fertility. This second demographic transition is characterized by structural (modernization, globalization, increased education), cultural (secularization, individual values, self-realization), and technological (modern contraception, assisted reproduction, new information technologies) changes (Reher, 2004; van de Kaa, 1987). Besides these changes, there are also shifts in the values and attitudes related to family life (cohabitation, common-law union), motherhood (childbearing is no longer the final goal of marriage), and sexuality (acceptance of premarital intercourse, use of contraception). For Sobotka (2008), these two groups of changes (behavioral and attitudinal) influence each other in strengthening the second demographic transition and subsequently influence the fertility variables. In addition, the demographic transition has been accompanied by major social changes in developed countries, including an aging population, family instability, a greater demand for immigrants—who increasingly come from predominantly non-Western populations as all Western countries face the same reality of (very) low fertility—and therefore increasing tension between native and immigrant populations (Vertovec, 2019).

Thus, the postulated fertility convergence has not (yet) taken place throughout the world. There are still huge differences between the North and the South, and between various population groups within the same region/country. Substantial research has documented various aspects of the fertility regime in different parts of the world (Stulp & Barrett, 2015).

This article aims to present an up-to-date description of the fertility dynamics in different regions of the world. Specifically, it aims to:

document the fertility dynamics in different regions of the world since the 1950s;

review the literature on the determinants of fertility decline; and

describe fertility differentials empirically according to their intermediate and distal determinants.

To achieve these objectives, the article draws on data from different sources. The data from the UN Population Division has been primarily used to describe global and regional fertility trends since 1950. The second source of data is the over 240 Demographic and Health Surveys collected in 84 countries since the 1980s. This second strand of data is appropriate to document fertility dynamics in developing countries, especially in sub-Saharan Africa. While presenting the theory of fertility decline and the various statistics associated with them, this article is not intended to assess the relevance of any theory. Instead, it should be recognized that different theories could be mobilized to better understand the fertility transition in different parts of the world (Mason, 1997).

Global and Regional Trends in Fertility Dynamics

Trends in the Total Fertility Rate

There are two ways to assess the pattern of fertility decline among different regions in the world: β- and σ-convergence. According to Strulik and Vollmer (2015):

β-convergence applies if countries of initially high fertility experience a stronger decline of fertility than countries of initially low fertility, while σ-convergence occurs if the cross-sectional dispersion, measured by the standard deviation of fertility, for a group of countries declines over time.

This idea is summarized in figures 1 and 2.

Figure 1 presents the trend in the total fertility rate (TFR) since 1950 in various regions of the world. It is based on estimates produced by the UN Population Division.

Figure 1. Trend of the total fertility rate in different regions, 1950–2020.

Figure 2 shows the disparities within each region and the change that has been achieved from the periods 1950–1955 to 2015–2020. Together, figures 1 and 2 show that during 1950–1955, fertility in sub-Saharan Africa was not different from that in most of the other world regions. In all other parts of the world (except Europe and North America), the level of fertility fluctuated between 5.6 and 6.6 children per woman. However, in 2015–2020, sub-Saharan Africa was the sole region of the world where fertility was still high, exceeding 4.5 children per woman. Thus, in early 1950, most of the world’s countries belonged to the high-fertility regime, with fertility exceeding 5 children per woman in 70% of the countries. By the year 2015, the picture reversed. The total fertility rate was < 4 children per woman in 80% of the countries, and it exceeded 5 children per woman in only 11 countries.

Figure 2. Fertility change in various regions of the world, 1950–2020.

Europe and North America (+ Australia and New Zealand)

Europe, North America, Australia, and New Zealand initiated fertility declines in the world that started in various parts of Europe in the 1870s (Watkins, 1991). As shown in figure 1, in 1950, the level of fertility oscillated between 2.5 and 4 children per woman, with various disparities between countries (figure 2). For example, the TFR was highest in Austria, at 6.2 children per woman, and lowest in Luxembourg, at around 2 children per woman. During 2015–2020, none of the countries in this region had a fertility higher than the level of replacement of 2.1 children per woman, with very few differences between the countries. During this period, many countries in the region reached what is commonly known as the lowest-low fertility regime—a persistent low level of fertility over a long period (Billari et al., 2009; Caltabiano et al., 2009; Sobotka, 2004). This low level of fertility has attracted diverse policy initiatives to reverse the course of the decline, with limited effects (Thévenon, 2008).

Sub-Saharan Africa

The fertility decline in sub-Saharan Africa has been slow and started only after 1980. During 1950–1955, the level of fertility was around 6.5 children per woman, with the highest fertility attained in Rwanda, at 8 children per woman. It was lowest in Gabon, at 4 children per woman (primarily due to sterility problems in this country rather than due to family control policies). After the independence of these countries in the 1960s, fertility increased to reach the level of 6.8 children per woman during the period 1975–1980. This period also coincided with an economic boom in many countries in the region. Since then, over a period of 40 years (from 1980), the average number of children per woman decreased from 6.8 to 4.8. In 2015–2020, 75% of the countries in sub-Saharan Africa had a fertility level below 5 children per woman, whereas only two countries had fertility below this level in 1950–1955. During the same period (2015–2020), 25% of the countries in the region recorded a TFR of fewer than 4 children per woman.

A closer view of fertility change in various regions in sub-Saharan Africa (not presented) reveals that the fertility decline has been dominant in the Southern and Eastern African countries while it has been modest in the West and Central Africa region. Another look at the fertility dynamic in the region was conducted by Tabutin and Schoumaker (2004), who demonstrated its different pathways. The authors showed how the fertility transition in sub-Saharan Africa has been affected by the HIV epidemic and political instabilities, including civil war. During the 2000s, the slow pace of the fertility decline in sub-Saharan Africa stimulated new research around the idea of fertility stalls (Bongaarts, 2008; Goujon et al., 2015; Schoumaker, 2019). Indeed, in various countries in sub-Saharan Africa, and especially among the leaders of fertility declines—Kenya and Ghana—fertility has stabilized around 4 children per woman and stopped declining further, prompting discussion over possible “exceptionalism” of the fertility transition in sub-Saharan Africa (Bongaarts, 2017; Bongaarts & Casterline, 2012).

Northern Africa and Western Asia

In Northern Africa and Western Asia, the level of fertility started to sharply decline during 1950–1955, until 2000–2005, when the pace of the decline slowed. During the period of 1950–1955, the level of fertility ranged from 2.8 children per woman in Georgia to 7.8 in Yemen. The median TFR in this region was around 7 children per woman. For the period 2015–2020, the TFR in the region was around 2.5 children per woman, with 75% of the countries having lower fertility than 3 children per woman. Furthermore, around one-quarter of the countries record a TFR of less than 2.1 children per woman. The patterns of stalling fertility have also been documented in some countries in this region (Cetorelli & Leone, 2012).

Central and Southern Asia

The pace of the fertility decline in Central and Southern Asia has been similar to the decline in Northern Africa and Western Asia. Nevertheless, it differs from the pattern in Northern Africa and Western Asia in the sense that the fertility level in 1950–1955 was much lower than in Northern Africa and Western Asia and the decline continued over the period 2000–2005. During 2015–2020, almost all the countries in the region had a TFR of over 2.0, with 75% of the countries registering a TFR of less than 3. Regionally, there is more convergence in the TFR between the countries from 1950–1955 to 2015–2020. Nevertheless, Afghanistan remains the country with the highest level of fertility, with 7.5 children per woman in 1950–1955 and 4.3 in 2015–2020.

Eastern and Southeastern Asia

While the level of fertility in Eastern and Southeastern Asia was very similar to the level in Central and Southern Asia in 1950–1955, the fertility decline has been more dramatic in the former region. The fertility level declined from 5.6 children per woman in 1950–1955 to below the replacement level in 2015–2020. More than half of the countries in the region have a fertility level below the replacement level. Timor-Leste remains the single country in the region with a fertility level above 3 children per woman (figure 2).

Latin America and the Caribbean

In Latin America and the Caribbean, the TFR decreased from 5.8 children per woman in 1950–1955 to around 2 children per woman in 2015–2020. This pattern of change characterized most countries in this region. The TFR varied between 2.3 (Uruguay) and 7.6 (Dominican Republic) in the period 1950–1955, to 1.6 (Puerto Rico) to 3.4 (French Guyana) during the period 2015–2020. As in the case of the European countries, almost half of the countries in the region have fertility rates below 2 children per woman. Higher fertility is found in Central American countries, such as Guatemala, Haiti, and French Guyana.

Global and Regional Trend in Fertility Intentions

Various factors determine the level of fertility in a country/region; the dynamic of fertility intentions is one of them. Thus, one way to forecast the fertility transition is to focus on fertility intentions, namely the desire for additional children or the desired number of children. The analysis of fertility preferences has indeed proven to be instrumental in identifying potential family planning demand, in understanding fertility inequalities between countries and regions, and in predicting future fertility trends. For example, in an attempt to understand the role of fertility preference in shaping fertility trends in sub-Saharan Africa, Casterline (2017) observed that demand for large family sizes was still prevalent in sub-Saharan Africa relative to other regions. He consequently concluded that marked fertility decline in the region would require major shifts in fertility intentions and implementation of such desires. Figure 3 illustrates the changes in the mean ideal family size (MIFS) in four major regions (figure 3, left panel) and assesses the relationship between the MIFS and the TFR in recent years (figure 3, right panel).

Figure 3 (left panel) indicates clearly that the MIFS has declined in all four regions but is still exceptionally high in various countries in sub-Saharan Africa. For the period 1990–1999, the median value of MIFS was 3.4 in North Africa and West Asia, 2.9 in Eastern and Southern Asia, 3.1 in Latin America and the Caribbean, and highest in sub-Saharan Africa (SSA), at 5.5. A similar pattern is revealed for 2010–2019, with sub-Saharan Africa registering a median value of the MIFS of 4.9, a mere 12% reduction relative to the 1990–1999 estimate.

The figure in the right panel depicts a clear one-to-one relationship between the MIFS and the TFR; in most countries, the MIFS reflects the TFR. Generally, desired fertility is higher than the TFR, suggesting that the fertility size documented did not always match the fertility desire at the individual level. This pattern is especially true for Niger (9 versus 7.8) and Chad (8.2 versus 6.3). At the opposite end, in Burundi, the TFR exceeds the MIFS by 1.5 children.

Figure 3. (a) Variation of the mean ideal family size (MIFS) over time. (b) Relationship between the Total fertility rate (TFR) and mean ideal family size (MIFS).

Theories of Fertility Decline: A Review

Several theories of fertility have been developed over the years to explain the factors that shape change in fertility and any decline, and more generally, the demographic transition. Even if the multiplicity of explanatory theories has not coalesced into a strong and coherent unifying theory, each theory in its way has shed some light on the fertility transition (Mason, 1997). The theoretical framework proposed by Davis and Blake (1956) and taken up by Bulatao and Lee (1983) presented the different factors that influence fertility and the levels of their relationships (figure 4), as well as the contribution of the various disciplines for the understanding of fertility (Leridon, 2015). However, it should be noted that if this conceptual framework is sufficient to explain the emergence of the fertility transition, it is imperfect to explain the very low fertility levels (less than 1.5 children per woman). The latter also mobilized a large corpus of theories structured around the notion of the second demographic transition (van de Kaa, 1987). For present purposes, the logic of researchers who have attempted to unify the two currents, in particular, Caldwell (2004) and Reher (2007), is discussed here.

Figure 4. Determinants of fertility.

Source: Adapted from Bulatao and Lee (1983, Vol. 1, p. 10).

The Stimulus: The Decline of Child Mortality

The decline in mortality (especially of children) has created a favorable environment for the fertility decline, by creating a discrepancy between the number of children attained and the number of children desired (Mason, 1997). But, admittedly, falling mortality does not automatically lead to falling fertility, for two main reasons. First, there are several other ways of dealing with excess children than just limiting birth, such as population migration. Second, there is an information asymmetry between the moment when couples realize that the phenomenon that is happening is not cyclical, and the moment when they start changing their behavior (Montgomery, 2000).

One of the pioneers of this view is Notestein (1945), who demonstrated that the central factor at work in the fertility decline had been the drop in mortality to a greater level than that of fertility. Consequently, the decline in mortality only appears as a stimulus or a mediating variable. However, the factors that created the stimulus also created an environment that would lead to “drastic social and economic changes.” Indeed, with the Industrial Revolution, the development of cities appeared and a new lifestyle emerged, with a turn toward individualism and changing family structure, as well as pursuit of self-fulfillment Thus, in the end, Notestein suggested it is urbanization and modernization that lead to lower fertility. The structural changes described by Notestein (1945) are the same that were described later by van de Kaa (2004) and Lesthaeghe (1995) in their formulation of the theory of the second demographic transition, but they are more profound. The work by Notestein and later by proponents of the economic determinants of fertility put a strong emphasis on economic development’s influence in shaping the fertility decline.

In the process of declining fertility, contraception plays an important role in giving couples the means to reach their “new aspirations” in terms of childbearing. Indeed, it becomes untenable for couples to arrive at the “ideal” of urban life with a large number of offspring. This pressure is stronger than the values of the past, which kept fertility at very high levels. Notestein (1945) did not deny the importance of values and attitudes, which he considered endogenous. For him, values change, but very slowly, so that he considered it “impossible to be precise about the various causal factors” (Chesnais, 1987, citing Notestein). The arguments against this view became the central counterpoint of diffusion theory. Although the declines in fertility in Europe followed this pattern, such was not the case in other environments that have experienced a decline in fertility without urbanization or modernization. While Notestein’s demographic transition theory has been criticized, it laid the groundwork for a theoretical framework that would evolve and be refined in subsequent years.

An important contribution was made by Davis (1963), who also aimed to explain the decline in fertility and to analyze its underlying factors. However, for Davis, a better understanding of the response to falling mortality in developed countries called for a broader perspective on the factors, including those that are not necessarily demographic and “quantifiable.” In his article, Davis (1963) showed that there is a single transition from one state to another and that the triggering factor is the individual perception of deterioration in living conditions (relative loss of status) and not a degraded economic situation. To arrive at this observation, he demonstrated that at the time of strong fertility decline, most of the countries studied (Northwest Europe in the 18th century and Japan in the 19th century) experienced very high economic growth rates in comparison to the past. Under these conditions, one cannot speak of difficult or degraded living conditions. However, given the anticipation of new economic environments, there is an increasing apprehension not to be able to fit in correctly, and therefore the perception of a degraded economic situation. Although the stimulus was the same for Notestein (1945) and Davis (1963), they disagreed on the ultimate fundamental determinants of fertility decline.

Intermediate (or Proximate) Determinants of Fertility Decline

Seminal papers by Davis and Blake (1956), and later by Bongaarts (1978), have made a profound impact on the study of the determinants of fertility around the world by distinguishing the intermediate variables (what Bongaarts called proximate determinants) that affect fertility directly, and the various distal economic, sociological, and institutional factors that affect fertility only through the intermediate variables. Although the intermediate variables play the same roles in all societies, their relative importance and impacts are different and have changed throughout the demographic transition.

Davis and Blake (1956) considered three intermediate determinants of fertility: those affecting the exposure to intercourse, those affecting the exposure to conception, and those affecting gestation and successful parturition. First, the factors that govern exposure to intercourse are mainly characterized by the formation and dissolution of the union and the exposure to intercourse within the union. While at the onset of the fertility transition the marriage regime played a decisive role in the level of fertility, it has had a much more limited effect on fertility since then. Throughout the world, not only has the age at first marriage increased, but also the meaning of marriage has substantially changed, especially in developed countries. Among the second set of intermediate factors documented in Blake and Davis’ conceptual framework, one specific variable has received tremendous consideration due to its important influence on fertility: the contraception regime. Fecundity and infecundity, either due to voluntary or involuntary causes, are also among the second set of factors. The final groups of factors include fetal mortality and abortion. According to Blake and Davis, in pre-industrial societies, abortion played a strong role, much more than contraception, in limiting fertility. The authors explained the finding by the fact that it is mostly women who bear the cost, revealing how gender disparities have played, and continue to play, a strong role in societies’ choices about fertility.

Reframing his argument in mathematical form, Bongaarts (1978) focused his attention on a small and quantifiable number of determinants. He began with the total fecundity rate, the theoretical number of births that a woman would have in the absence of any inhibiting factors, estimated to be 15.3. He showed that the total fecundity rate could be expressed as the product of four indices—an index of contraception (Cc), an index of proportion married (Cm), an index of lactation infecundability (Ci), and an index of induced abortion (Ca)—to estimate the actual level of fertility (TFR), with the formula:


The indexes vary between zero and one, reflecting their reducing effect on the total fecundity rate. Since Bongaarts’ publication, his framework has been extensively used to document various aspects of fertility change in various countries in the world or to compute some fertility-related outcomes (abortion, for instance), while at the same time being further developed and clarified (Stover, 1998).

Distal Determinants of Fertility

Economic Determinants of Fertility

From an economic and social point of view, two schools of thought have tried to identify the factors that motivate individuals (couples) to have children or that determine the number of surviving children and, therefore, the level of overall fertility. However, a fundamental point of these approaches is to treat “the child” as an economic good and to posit that its “acquisition” depends on the factors of supply and demand. The first school—the Pennsylvania school of thought—developed the works of Easterlin (1966, 2015) to explains the level of fertility through the role of external factors, both biological and social (supply-side factors) that limit the number of surviving children, either directly through the regulation of the length of fertile life or indirectly through mortality. Thus, any social or economic change that influences these factors (e.g., age at marriage or improvement in infant health) will have an immediate and obvious consequence on couples’ fertility. The sociological factors that condition the behavior of individuals are made up of social organization, cultural norms, and sanctions. The authors cited the example of several societies where these standards exist or have existed and how they have affected fertility. For example, they cited the cases of resumption of postpartum sexual relations in West African societies and religious norms for the Hutterites, the peoples of Lebanon (religion and dominance of other groups), Jews, and Muslims in India.

The second approach posits the demand side (Chicago School) and is led by the works of Becker (1991). In this case, couples, “Homo economicus,” decide to have children or not depending on the availability of other goods and the usefulness they derive from them under certain constraints. The Homo economicus hypothesis assumes that the number of children in a couple can be predicted once the other variables have been identified, that is, tastes or preferences are constant. In its first version (Chicago School), the constraints consisted of relative costs and available resources. However, the constraints can take several forms—biological, economic, security-related, and time-related (Hobcraft & Kiernan, 1995)—and they can come from social, religious, or ethnic influence (Folbre, 1994). Clearly, when they decide to have children, couples assess the trade-off between the quality of children they wish to have (and which therefore depends on their monetary and physical investments) and the quantity that allows them to ensure the expected level of quality. Becker’s economic theory of fertility has been widely embraced for understanding the decline in fertility, although it has also attracted criticism.

At the macro level, the economic transformation of society contains the seeds of declining fertility worldwide and may explain the (very) low level of fertility in Europe and North America. Indeed, for Caldwell (2004), the main factor that led to the first decline in fertility, in particular the free market economy, was greatly reinforced thereafter, resulting in a disparity between fertility levels and the requirements of the new economic order. Clearly, very low fertility or low fertility is explained by the global economic system driven by liberal economic policies that may prove in time to be the dominant force behind low, especially very low, fertility, and welfare regimes. Family systems and temporary economic crises are seen as factors of second-order importance (Caldwell, 2004; Caldwell & Schindlmayr, 2003). Other factors are used in this process to explain low or very low fertility levels. In fact, it is these secondary factors, such as education, the employment of women, laws, and policies, that explain the fertility differentials between countries. Like Notestein, Caldwell considered that values and attitudes are endogenous and must be classified as secondary factors, which change slightly. What is at stake is the strength of the stimulus that the liberal economic system has. Chesnais (1987) said, “cultural variables only appear as factors likely to advance or delay an evolution driven by the dynamics of modern economic growth.”1 Herzer et al. (2012) established a causal effect between economic growth plus mortality decline on fertility decline using country-level data for the 20th century. However, there are certainly other forces at work, since the specific pattern of the fertility transition differs substantially across countries and years (Lee, 2003; Reher, 2004).

An alternative explanation for low fertility levels has been developed by Lesthaeghe and van de Kaa (1986, 1987), who coined the term second demographic transition (Zaidi & Morgan, 2017). Simply stated, the second demographic transition, which aims to explain sub-replacement fertility levels, refers to a set of changes both in fertility behaviors and in the values that support traditional behaviors. Concretely, the second demographic transition is characterized by structural (modernization, globalization, increased education), cultural (secularization, individual values, self-realization), and technological (modern contraception, assisted reproduction, new information technologies) changes. Besides these changes, there are also changes in the values and attitudes relating to family life (cohabitation, common-law union), motherhood (child is no longer the final goal of marriage) and sexuality (acceptance of premarital intercourse, contraception). For Sobotka (2008), these two groups of changes (behavioral versus attitudinal) influence each other to reinforce the second demographic transition on the one hand and subsequently influence the levels of fertility. On the other hand, however, the order in which they influence each other depends on the prevailing economic conditions. Although there is much criticism of the second demographic transition, it appears to be the most coherent way of analyzing the changes taking place in fertility patterns in Europe (Sobotka, 2008).

Diffusion and Social Networks

Economic theories have played an important role in explaining the fertility decline, including relegating other explanatory mechanisms to the background. However, the variability over time and space in the influence of the economy on fertility, differentials in the magnitude, and the absence of a threshold have called into question accepting only an economic explanation for the fertility decline (Johnson-Hanks, 2015). In contrast to the economic approach, the cultural approach to explaining the decline in fertility is structured around the social organization, cultural norms, and sanctions. In pretransitional societies, the system of social organization, religious values, and moral codes governed all aspects of people’s lives (family, economic, political). These norms were organized to maintain a high level of fertility. The theory of diffusion of innovations, in contrast, offers a coherent framework for analyzing the non-economic determinants of the fertility transition.

The theory of innovation diffusion brings together several theoretical currents that have taken root in different disciplines, notably in sociology (Montgomery et al., 2003; Sampson, 2003), in the theories of social disorganization (Shaw & McKay, 1942), social capital (Coleman, 1988), and collective efficiency (Sampson et al., 1997, 1999). Diffusion is defined as a “process by which an innovation is communicated through certain channels over time among the members of a social system” (Rogers, 1983, p. 5). Applied to fertility, diffusion refers to

the spread or adoption of new information, ideas, beliefs, or social norms capable of influencing reproduction decisions and behavior that occurs through social interaction and influence, either at the interpersonal level or through impersonal channels, such as the mass media (National Research Council, 1999, p. 7).

The theory differs from classical theory by postulating that the decline in fertility is not (solely) due to demographic, economic, or social structure changes. The decline in fertility also comes from the diffusion of new forms of attitudes and behaviors that begin within a given group and which then spread throughout the population (Casterline, 2001; Montgomery & Casterline, 1996). Thus, according to this theory,

fertility decline is the consequence of the increased prevalence of attitudes and behaviors that were previously very rare or absent in the population (i.e., they are innovative), and their increased prevalence is the consequence of the spread of these attitudes and behaviors from some segments in the population to others (i.e., a diffusion process) (Casterline, 2001, p. 6).

Diffusion theory mainly distinguishes between two streams of thought: one specific to innovation (the innovation being diffused) and a second that concerns how the innovation spreads to the whole population. Each of the streams has been the subject of important theoretical developments. Concerning innovation, while the work of Knodel and van de Walle (1979) and later of Watkins (1990) focused on the new types of behavior that are diffused within the population and that will lower fertility (the cultural approach), the ideational theory of Cleland and Wilson (1987) focused on the importance of innovations in ideas (information, standards, etc.). Watkins’ work focused, for example, on the influence of culture and language on fertility differentials in pre-transitional Europe (Bongaarts & Watkins, 1996; Watkins, 1995). The fundamental criticism that can be made of this approach is that it considers individuals as patient, attentive actors and subjects without reluctance adhering to the dictates of society and the family (see Bledsoe’s, 1990, example of West Africa). On the contrary, according to McNicoll (1992), individuals appropriate the standards and try to adapt them to the institutional context in which they live. It is this set that generates differential levels of fertility between societies. On the other hand, according to the ideational approach to fertility, the timing of the demographic transition depends on the dissemination of information and new social norms concerning innovation—birth control, in this case. Caldwell’s works on the importance of Western values conveyed by education in the decline of fertility in Africa can also be aligned with this second view (Caldwell, 1982).

The second current of thought is concerned with how innovations appear and are diffused, as well as the role of the actors in this diffusion. Accordingly, innovations spread through two mechanisms: social learning and social influence (Behrman et al., 2002; Montgomery & Casterline, 1996). According to these authors, social learning “reduces the uncertainties about an innovation” and thus leads individuals to adopt the new change without great cost (Kohler, 1997; Kohler et al., 2001). This is the case with using a new contraceptive method in a given group. Members outside the group can also adopt the method based mainly on the beneficial effects observed for the first users or on the contrary by using it differently because of the problems that itgenerated in the first group. Social influence implies a reinforcement of social norms through examples of behaviors that others can copy (Keim et al., 2009; Montgomery & Casterline, 1996).

If interest is growing for the study of these mechanisms in the diffusion process, the lack of knowledge on their role in the demographic transition pointed out more than 20 years ago still holds (Keim et al., 2009). The current thought about diffusion will give an important place to social networks and the mass media as vehicles through which innovations will spread to the whole population. Its founding premise is based on the interdependence of individual actions and the structures (social networks) in which such actions take shape (Wasserman & Faust, 1994). Different types of networks have been studied in connection with fertility diffusion, including informal communication networks (Behrman et al., 2002; Entwisle et al., 1996; Kohler, 1997; Valente et al., 1997; Watkins & Danzi, 1995), and family networks (Keim et al., 2009, 2013).

Government Policy

The predominant perspective prescribed by the classical model of reduced fertility is economic development. Any other perspective appears unrealistic in this context, as evidenced by the slogan, “The best contraceptive is development.” Likewise, according to Pritchett (1994), family planning policies are not about reducing the desire for fertility (which is essential to trigger decline) but about meeting the unmet need for family planning. It should be noted, however, that most of the theories give an important place to institutions (family or state) in setting up the conditions that either favor economic development, reduce the influence of fertility norms (e.g., age at marriage), or allow better dissemination of information or the provision and circulation of the innovation. In both developed and developing countries, the success of these policies remains mixed. In fact, a precondition for the decline in legitimate fertility, according to Coale, is that “fertility limitation must enter into the ‘calculus of conscious choice’—that is, parents must be able to entertain the idea that they could choose, even approximately, the number of children that they bear.” Also, if childbearing motivation does not enter this conscious pattern, any external policy will be doomed to failure. We will briefly look at population policies in developing countries (which are interested in reducing fertility) and in developed countries (whose most recent policies aim to raise fertility). Ouedraogo et al. (2018) showed how population policies changed in the world from 1976 to 2013. In 2013, 51 countries had a pro-natalist policy (all of Europe) and 76 an anti-natalist policy, including almost the majority of African countries. Forty years earlier, these numbers were 18 and 55, respectively (and included very few African countries).

In developing countries, two particularly important policy areas are related to education and family planning. Caldwell’s pioneering work in West Africa has shown the importance of education as an engine of reproductive change, either by giving women access to information, by increasing women’s autonomy, or by adopting new Western values (Caldwell, 1982). Several studies have indeed shown empirically that women’s level of education is strongly correlated with their level of fertility (Bongaarts, 2003; Shapiro, 2012). The commitment of countries at the Jomtien education conference in 1990 and the international population conference in Cairo in 1994 resulted in increased investment in education to strengthen the education of women and their decision-making power within their household and the community.

While it was introduced without problems in developed countries, in particular with the support of states, family planning encountered more reluctance in developing countries at its inception. It continues to raise questions in certain countries (Ouedraogo et al., 2018), even if its importance no longer needs to be demonstrated with respect to fertility change, maternal and child health, and the economic opportunities it creates (DaVanzo & Adamson, 1998). Indeed, according to Bulatao, a vigorous family planning policy contributed 40% to the decline in fertility from 6 children per woman in the 1960s to 3 children per woman 30 years later (Bulatao, 1998). The economic success of the tiger economies of Asia (South Korea, Singapore, Hong Kong, Taiwan, and Japan) has also been partly explained by, and attributed to, the control of fertility. In sub-Saharan Africa, Zimbabwe, Kenya, and Ghana pioneered a vigorous family planning policy. Therefore, it is not surprising that the fertility transition started in these countries. More recently, Rwanda has been active in promoting family planning. Between 2005 and 2010, the use of modern contraception increased from 17% among married women and women in a union, to 52%. At the same time, fertility fell by 25%, from 6.1 to 4.6 children per woman (Westoff, 2013).

In developed countries, population policy aims to raise fertility above the generation replacement level. These policies are not uniform in content and include a set of objectives for raising fertility, work/child balance, reduction of economic inequality and family poverty, care and education of children, and balanced sharing of family tasks (Thévenon, 2008). However, most of these policies tend to affect the timing of the birth event (age at childbirth) rather than its intensity (number of children). According to McDonald (2002, p. 432), “successful (fertility) policy will almost certainly involve changes in social and economic organization on a much wider scale.”

An Empirical Assessment of the Determinants of Fertility in the World

This part of the article analyzes first the evolution of certain direct determinants, specifically marriage and contraceptive use, in different parts of the world. Then it focuses on two indirect determinants, the level of education and women’s autonomy, and empirically analyzes their influence on fertility.

Intermediate Determinants of Fertility

Marriage Patterns

Analysis of age at first marriage and level of marriage are some ways to understand marriage patterns. United Nations data make it possible to assess the trend in the average age at first marriage in different countries across all world regions. Figures 5 to 8 examine this evolution during the 1990s and for the more recent period (around 2015) for both men and women.

Europe and North America (+ Australia and New Zealand)

Figure 5 shows an exceptionally high average age at first marriage in Europe and North America. On average, in this region, the age at first marriage has increased from 24.90 years among women in the 1990s–29.12 years more recently (2010–2018). For men, the age increased from 27.90 to 31.76 years over the same period (1990–2018).

Figure 5. Singulate mean age at first marriage, 1990s versus 2020s, Europe and North America.

Sub-Saharan Africa

In sub-Saharan Africa, girls increasingly start their marital life later as well, but the change has been much slower than in the other regions in the world. Figure 6 shows that, on average, the mean age at first marriage for women increased by 2 years, from 20.48 years in 1990 to 22.19 years in 2015. This increase was less than 1 year for men, from 26.30 years to 27.27 years. Again, in most countries in this region, the mean age at first marriage is below 20 years for women. Marriage continues to play a vital role in the fertility dynamics in sub-Saharan Africa, more so than elsewhere in the world (John & Adjiwanou, 2021).

Figure 6. Singulate mean age at first marriage, 1990s versus 2020s, sub-Saharan Africa.

Northern Africa and Western Asia

Figure 7 presents the trend in the mean age at first marriage in Northern Africa and Western Asia. In most countries, the mean age at first marriage for women was 22.77 years in the 1990s, and it increased to 24.82 years in the second half of the 2010 decade. For men, the mean age at first marriage is above 25 years. It rose from the average of 26.40 years in the 1990s to 28.89 years during the 2010 decade.

Figure 7. Singulate mean age at first marriage, 1990s versus 2020s, North Africa and West Asia.

Central and Southern Asia

In Central and Southern Asia, the marriage patterns are similar to those observed in sub-Saharan Africa, with the average age at first marriage around 20.41 years for women and 23.80 years for men. However, the increase has been more substantial in this region than in sub-Saharan Africa. For example, in recent years (2006–2017), the average age at first marriage was 21.89 years for women and 25.53 years for men.

Figure 8. Singulate mean age at first marriage, 1990s versus 2020s, Central and Southern Asia.

The increase in the age at first marriage is reminiscent of a radical change in conjugal lifestyles worldwide, with the emergence of new forms of partnering and an increase in the rate of divorce and remarriage. This change in marriage patterns is significant and is differentiated by socioeconomic factors.


Cohabitation is a new pattern of union arrangement taking place in various countries globally, especially those in Europe and North America. Indeed, there is a disconnection between age at marriage and fertility in these regions, which is one aspect of the second demographic transition (Beaujot, 2000; Lesthaeghe, 1995; van de Kaa, 1987). Marriage is no longer the place of procreation. In fact, the prevalence of marriage has dropped considerably in most countries in Europe and North America, where children are increasingly born outside the union of both parents. The children are born into a new form of family structure: cohabitation. Difficulty in measuring marriage formation in Africa has often led to classifying marriages together with cohabitation, which does not allow understanding of the evolution of the two phenomena separately. Thus, figure 9 presents the trend in the proportion of women age 15 to 49 in different world regions who are in a union (i.e., either formally or informally married).

Figure 9. Average proportion of married and in-union women, 15–59 years old.

In all the regions, this trend has decreased over the years. Central and Southern Asia is the region in the world with the highest proportion of women age 15 to 49 in a union. The proportion of women in a union declined from 72.8% in 1970 to 68.7% in 2020. At the opposite end, the proportion of women in a union is the lowest in Latin America and the Caribbean. In Europe and North America, 65.2% of women were in a union in 1970 and 54.9% in 2020. More detailed data from the OECD (2011) further show that close to half of the unions are cohabiting relationships in recent years.

Contraception Patterns

The use of contraceptives is globally accepted as the most effective approach to fertility regulation. Consequently, the levels of contraceptive use within and between countries or regions are regarded as the dominant force shaping variation in country and regional fertility trends (Bongaarts, 2006). Contraceptive use is mainly composed of modern methods, defined as any modern method (pill, intrauterine device [IUD], injection, male or female condom, diaphragm, implant, or male and female sterilization) and traditional methods (rhythm [also known as periodic abstinence], withdrawal, and other folkloric methods). The traditional methods remain largely used in developing countries (Bertrand et al., 2021).

Figure 10 presents the trend in modern contraceptive use in various world regions from 1994 to 2019. As expected, modern contraceptive use has increased substantially worldwide, including in sub-Saharan Africa (even if its level is still very low in many countries in the region). In 1994, in half of the countries in sub-Saharan Africa, the median modern contraceptive use among married women of childbearing age was about 5%. In 2019, the median increased to 25%, with the proportion of modern contraception users among married women of childbearing age ranging from 5% in South Sudan to 52% in Eswatini. In Northern Africa and Western Asia, the increase in modern contraceptive use has been slower, where the median shifted from 20% in 1994 to 24% in 2019. Figure 10 depicts the trends in the rest of the regions, with Europe and North America showing the highest use of modern contraception.

Figure 10. Trend in modern contraceptive use, 1994–2019.

Figure 11 illustrates the prevalence of contraceptive use for various methods from 1994 to 2019 based on data from the United Nations, Department of Economic and Social Affairs, and Population Division (2019). First, there are marked inequalities in the popularity of contraceptive methods by region. Injectables are predominantly used in sub-Saharan Africa, while IUDs are commonly used in Eastern and Southeastern Asia. In Latin America and the Caribbean countries and Central and Southern Asia, women frequently use female sterilization, while pills are predominately used in Europe, North America, Northern Africa, and Western Asia and Oceania countries. The use of injectables and implants has risen across all regions, with a substantial increase registered in sub-Saharan Africa. On the other hand, the use of the rhythm method has declined in all the regions. Female sterilization has dropped, except in Central and Southern Asia and in Northern Africa and Western Asia. There appear to be fewer changes in the use of IUDs (except for the Oceania region, where it increased) and male sterilization (except for Eastern and Southeastern Asia, where it has decreased). In contrast, the use of male condoms and female sterilization has changed markedly, with significant increases observed in Eastern and Southeastern Asia.

Figure 11. Trends in use of contraceptive methods.

Data source: Calculations are based on the data compilation World Contraceptive Use 2019, additional tabulations derived from microdata sets, and survey reports and estimates of contraceptive prevalence for 2019 from Estimates and Projections of Family Planning Indicators 2019. Population-weighted aggregates.
Association Between Contraception and Fertility

Fertility is predominantly high in populations with low contraceptive prevalence, and it falls with increasing modern contraceptive use. Although the strength of this relationship may vary by region and the stage of the fertility transition, it appears to be universal. This pattern is depicted in figure 9, showing the association between fertility (measured by the TFR) and the percentage of all women of childbearing age using modern contraceptive methods. In both earlier and recent DHS surveys, most countries in sub-Saharan Africa had lower contraceptive prevalence corresponding to higher fertility levels. On the other hand, most countries in Latin America and the Caribbean, Eastern and Southern Asia, and Central and Southern Asia, reported high use of modern contraceptives, which corresponded to their low fertility level. However, we note exceptional cases in both earlier and recent surveys in Northern Africa and Western Asia, where low contraceptive use is linked with extremely low fertility. This pattern is explained by a high prevalence of induced abortion in these countries. Figure 12 also reveals that the slope of the regression line is steeper from earlier surveys than later surveys. This trend manifests the changing pattern of contraceptive use’s effects over the course of fertility decline, with stronger effects observed at the onset of the fertility transition.

Figure 12. Total fertility rate and modern contraceptive use in various periods (1990s versus 2015).

An Empirical Assessment of the Distal Determinants of Fertility in the World

Education and Fertility Level and Intention

The link between fertility and education is complex; nevertheless, one thing is evident—the fertility of women with no education is generally higher than that for women with more years of schooling. Some scholars have explained this relationship using the economic theory of fertility, which contends that educated women are more likely to emphasize quality (children with good health, education, and social interaction) rather than quantity, leading to a higher human capital investment per child. This inherently induces the desire to limit childbearing. An alternative argument along this line of economic thought relates to the high opportunity costs associated with childbearing. Educated women may find that raising children competes with attractive opportunities in the labor market; thus, they may deliberately have fewer children to realize their economic goals. Furthermore, substantial evidence reveals that attainment of high education levels is aligned with delaying the formation of marital unions and initiation of childbearing, which consequently shortens a woman’s reproductive lifespan. The shortened interval biologically constrains the supply of children, especially during prime childbearing years.

On the other hand, some researchers have considered the link between fertility and education within the context of an ideational theory of fertility. The underlining argument suggests that educated women desire smaller families, a desire that may be influenced by better information access and social networks. Better civil and social networks that come with higher education allow the diffusion of information and resistance to social norms that encourage, for example, large family size, short birth intervals, and use of traditional birth control methods. Such behaviors are considered an unpopular lifestyle for a modern woman. Thus, the norms, attitudes, and behavior concerning these childbearing ideals are strikingly different among educated women than among their less-literate counterparts. Indeed, there is extensive evidence in the literature indicating that, compared to women with no formal schooling, educated women not only have a desire for smaller family size but also have longer birth intervals, higher levels of knowledge, and adoption of modern birth control methods.

Because the mechanisms through which education influences fertility are so diverse (Adjiwanou et al., 2018; Vikram et al., 2012), its impact on fertility should be expected to vary by space and time and can better be explained within specific social, cultural, and economic conditions. This inequality is illustrated in figure 13, which shows the association between TFR and the percentage of women with secondary or higher education in six major world regions. The variation in TFR for countries with the same level of education is remarkable (see figure 13). For example, countries with 50% to 75% of women with secondary or higher education in Latin America and the Caribbean have TFRs ranging between 2 and 3 children. In contrast, countries in sub-Saharan Africa with a similar composition of women with secondary or higher education have TFRs that exceed 3 and reach as high as 5.1.

Figure 13. Total fertility rate and female education in 2015 in various regions.

Empowerment and Fertility Level and Intention

Women’s participation in decision-making regarding their health, major household purchases, and freedom of mobility are widely used as proxy measures of women’s empowerment. Researchers have attempted to understand the relationship between these measures and various indicators of fertility, including fertility intentions, birth intervals, use of modern contraceptive methods, and actual childbearing practices (Adjiwanou & N’Bouke, 2015; Upadhyay et al., 2014). The underlying assumption is that women who participate in household decision-making are equally likely to have a say in regulating their childbearing preferences and behavior. Upadhyay et al. (2014) provide a detailed literature review of 60 studies on fertility and women’s empowerment conducted between 1990 and 2012 in low- and middle-income countries. The authors observed a positive association between female empowerment and lower birth rates, longer birth intervals, and minimal rates of unplanned pregnancies. However, the authors also noted significant discrepancies in the measure of the women’s empowerment construct and variation in the strength of association across different study populations. A similar pattern was observed recently by Prata et al. (2017) in another systematic literature review, spanning the same period but focusing on women’s empowerment and family planning. The authors noted a positive association between women’s empowerment and selected family planning outcomes in certain studies, while a null relationship was observed in some cases. Consequently, they concluded that the relationship between women’s empowerment and family planning is complex, heavily dependent on the domain of women’s empowerment and the measure used as well as the subpopulation under consideration. This latter observation is illustrated in figure 14, which presents the correlation between total fertility and the percentage of women participating in the three domains of decision-making (own health, major household purchases, and a visit to family members or relatives). In general, lower fertility is associated with higher levels of participation in decision-making. However, this obscures a potential null relationship in Central and Southern Asia and Latin America and the Caribbean and even a negative correlation in Eastern and Southern Asia. In sub-Saharan Africa, the relationship is not clear when the level of participation in decision-making is less than 50%.

Figure 14. Relationship between total fertility rate and women’s empowerment.


The fertility transition is still underway in sub-Saharan Africa, whose population will reach 2 billion in 2050. In the other regions of the world, the challenge is to recover fertility that has reached extremely low levels. Thus, globally, fertility issues are no longer the same as in the middle of the past century.

In sub-Saharan African countries, robust fertility rates pose major challenges for investment in social services, which are fundamental for furthering fertility decline. The first signs of a decline in countries like Ghana, Kenya, and Zimbabwe have given way to fertility stalls, with rates declining very slowly, despite investments in family planning. For some authors, the fall in educational investments in the 1980s explains the slowdown in current fertility (Goujon et al., 2015). The pattern of a resumption of fertility decline or stagnation following a crisis, either an economic crisis, a disease-related epidemic, or armed conflicts, is not new. Tabutin and Schoumaker (2004), in their description of the demographic situation in Africa from the 1950s to 2000, revealed this explanatory schema. The region is still marred by various humanitarian crises that may erase the success of the previous decades, especially in Burkina Faso and Ethiopia. One thing is certain: the beginning of the transition in sub-Saharan Africa has been different from what was observed elsewhere (Timæus & Moultrie, 2020), and the direction it will take remains to be determined.

In Europe and North America, the decline in fertility was more pronounced than predicted by demographic transition theory (Vallin, 2000). In all the countries of this world region, the total fertility rate is below 2 children per woman, and public policies have failed to remedy it. It is estimated that several countries will observe a drop in their population size. For example, according to UN projections, the current population of Europe will decline by 37.1 million, reaching 0.7 billion by 2050. With 60.5 million people in 2020, Italy will have 54.4 million people in 2050. Hungary will see its population decrease from 9.5 to 8.5 million. A change in the economic or social paradigm is necessary to raise fertility in high-income countries. It should be noted that not all high-income countries experience the same realities. Fertility is higher in northern Europe than in the south, which can be explained by a gender structure at the institutional level that matches the gender structure at the individual level. According to McDonald (2002), these differentials in fertility can be explained by a dichotomy between institutions that concern people as individuals (such as education or employment) and those that consider the individual as a member of a family (family services, social security, taxation). He contended that the more a society adheres to a traditional family system, the higher the degree of divergence between social institutions, and the lower the fertility (McDonald, 2002). In the meantime, these changes and the associated fertility levels are not without consequences for these countries. In addition to an aging population, there will be increasingly expensive healthcare services, a need for international immigration (with its share of tensions between populations), and the rise of the populist right in several countries.

Between these two regional patterns lie the vast majority of fertility situations in the world, including Latin America, North Africa and the Middle East, and South and East Asia. Important questions remain about the future of fertility around the world. How will the very low fertility levels evolve with the higher fertility levels in sub-Saharan Africa and its modest provision of contraceptive access? Moreover, even for governments reluctant to promote family planning, its benefits have been demonstrated for women’s health, for their autonomy, and above all, for enabling them to achieve the objectives of fertility they set for themselves. On one side of the Mediterranean, in southern Europe, the economic crisis has induced a significant decline in fertility, while south of the Mediterranean, the crisis has not had the same consequences for fertility. In Asian countries, there are still pockets of strong demographic growth. Fertility, a vital human behavior, has transitioned from high to low levels nearly everywhere, and where disparities are large, they will continue to be studied.

Further Reading



  • 1. Chesnais (1987) said, “cultural variables now appear only as factors likely to advance or retard an evolution driven by the dynamics of modern economic growth.”