Human Reproduction Update Advance Access originally published online on June 14, 2006
Human Reproduction Update 2006 12(5):603-616; doi:10.1093/humupd/dml025
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Demographic effects of the introduction of steroid contraception in developed countries
Inserm, U569, F-94276 Le Kremlin-Bicêtre; INED, F-75020 Paris and Université Paris-Sud, F-94276 Le Kremlin-Bicêtre, France
To whom correspondence should be addressed at: U569, 82 rue du Général-Leclerc, 94276 Le Kremlin-Bicêtre Cedex, France. E-mail: leridon{at}ined.fr
 |
Abstract |
|---|
 TOP Abstract IntroductionSourcesTrends in fertility and...Country studiesAdolescentsTheories on fertility declineConclusionsReferences |
|---|
The use of the contraceptive pill increased very rapidly in the 1970s in many developed countries, and fertility almost simultaneously started to decline. We discuss here the possibility of a causal link between these two major changes. We first provide evidence for a relationship between the spread of oral contraceptive use and the change in fertility in many European countries over the last three or four decades. The situation of specific countries is examined more in depth on the basis of available literature. We then review the various theories attempting to explain these trends and see how the family planning variables are treated in these approaches. At the country level, the conclusion is unambiguous: within individual countries, there is no systematic negative correlation between fertility and contraceptive pill use. The development of hormonal contraception cannot be considered as responsible for either starting or the size of the fertility decline. A more subtle chain of causality must be considered, but there is no agreement on a general theory of fertility changes. Most authors, however, agree that the diffusion of modern contraception has certainly contributed to the reduction in the number of unwanted pregnancies and has also facilitated and favoured the adoption of new (more restrictive) norms for the ideal family size.
Key words: developed countries / fertility / hormonal contraception
|
Introduction |
|---|
|
TOPAbstractIntroductionSourcesTrends in fertility and...Country studiesAdolescentsTheories on fertility declineConclusionsReferences |
|---|
About 100 million women worldwide use the contraceptive pill (Population Reference Bureau, 2000a
,b). The first oral contraceptive was marketed in the USA in 1960. In the following years, it was released in many other countries. However, the product was often first available not as a contraceptive but as a means of ‘regulating the woman’s cycle’, making the situation regarding usage unclear. For example, in France, the local family planning association Mouvement Français pour le Planning Familial (MFPF) started prescribing the contraceptive pill in 1965, 2 years before the first law on contraception was passed. However, it was available and sometimes prescribed by physicians, since 1961. The use of the contraceptive pill increased very rapidly in the 1970s. Almost simultaneously, fertility started to decline in many developed countries: the total fertility rate (TFR, sum of the age-specific fertility rates) of the 15 member countries of the European Union decreased from an average of 2.72 children per woman in 1965 to 1.96 in 1975 and is currently below 1.50. In the USA, the TFR declined similarly from 2.91 in 1965 to 1.77 in 1975 but is now just over 2.0 children per woman. Some countries showed a striking coincidence. For example, in France (this case will be discussed later: see Figure 5), fertility peaked in 1964 at the end of the baby boom and then rapidly declined just as hormonal contraception was increasingly being adopted by French women. Therefore, we can ask whether there is a possible causal link: has the diffusion of oral contraception caused the recent decline in fertility observed in Europe and other developed countries? Before analysing this correlation in detail, we must first bear in mind that fertility had declined in most industrialized countries well before the pill existed. The parallel reduction in both mortality and fertility, usually called the demographic transition, began in many countries at the start of the 20th century and sometime during the last part of the 19th century or even earlier in France. This was well before modern contraceptives appeared. The mean number of children often dropped from about six per woman to three or less, when the contraceptive pill, the intrauterine device (IUD) or even the condom was either not available or not widely used, sterilization was limited to strong medical indications and abortion was prohibited (although abortions were often performed illegally in bad conditions). This highlights that a substantial decline in fertility can occur for reasons other than the availability of efficient means of birth control.
However, the appearance of the new contraceptives introduced a fundamental change in the behaviour of couples, which has been called the new (or a second) contraceptive revolution (Westoff and Ryder, 1977; Leridon et al., 1987). The key feature of modern contraceptives is not that they are more effective (traditional methods could be used quite effectively by motivated couples) but that they are under the sole control of women (Héritier, 1996; Bajos and Ferrand, 2004) and no longer linked to intercourse. A woman using the pill or an IUD is in a permanently infertile state (until she stops using the method), and if she wishes to conceive, she has to decide to end this situation. Therefore, she (or the couple) has to find good reasons for stopping her sterile state and start being exposed to pregnancy. However, in the past, having one or two children in the first years of marriage was quite natural, and the decision taken in this case was to stop being exposed to additional pregnancies.
We must also consider the many other changes that have occurred in these societies during the last few decades when looking more specifically at the emergence of new means of fertility regulation and their diffusion in industrialized countries. This will be examined in Theories on fertility decline section.
In this article, we will first provide evidence for a relationship between the spread of oral contraceptive use and the change in fertility in many European countries over the last three or four decades. For some of these countries, we will then report the explanations proposed by demographers and determine whether they assume a causal link between the increase in hormonal contraceptive use and the decline in fertility. Adolescent exposure to unwanted pregnancies will be addressed in a specific section. We will finally review the various theories attempting to explain these trends and see how the family planning variables are treated in these approaches.
|
Sources |
|---|
|
TOPAbstractIntroductionSourcesTrends in fertility and...Country studiesAdolescentsTheories on fertility declineConclusionsReferences |
|---|
For this review, we have consulted several bibliographic databases: POPLINE, which has the largest set of references on demographic trends and on contraceptive use; JSTOR, which gives access to full-text articles published by the main demographic journals; and MEDLINE, for more medical references. Owing to the ambiguous meaning of the word ‘fertility’ (Habbema et al., 2004
), the Medline search was not the most efficient because we were looking for articles analysing the trends in ‘fertility’ in the demographic sense (the number of births per woman) and not in the biological sense (the ability to conceive). Throughout this article, we will use fertility in its demographic meaning.
Hormonal contraception will be considered as synonymous with steroid contraception and with contraceptive pills because injections and implants were rarely used in the countries and periods considered here. According to UN data (United Nations Population Division, 2003), 0.3% of women of reproductive age used injections or implants in the mid-1990s, with the highest rate of use in the UK (3.0%).
|
Trends in fertility and oral contraceptive use in developed countries |
|---|
|
TOPAbstractIntroductionSourcesTrends in fertility and...Country studiesAdolescentsTheories on fertility declineConclusionsReferences |
|---|
In Europe in 1960, the average TFR (the mean number of children per woman under the prevailing conditions in a given year) was 2.6, and hormonal contraception was non-existent. Forty years later, the TFR was 1.4, and almost 18% of all women of reproductive age living in a marital or consensual union were using hormonal contraception, most often the contraceptive pill but sometimes implants or injectables (United Nations Population Division, 2000
, 2003). The contraceptive pill is the most widely used method (17.4%), and the IUD is second (12.2%); withdrawal and condoms are each used by 11–12% of couples and sterilization by 7% (most often the woman). Each of these percentages is not high in absolute terms because of the various methods used, but the records for some individual countries can be impressive: >58% of German couples and 49% of Dutch couples use the contraceptive pill, 26% of Finish women have an IUD, 30% of British couples have been sterilized (17% men and 13% women), withdrawal is used by 26% of Portuguese couples and 24% of Spanish couples use condoms. Therefore, it is much more useful to study fertility trends and pill use in individual countries rather than combined.
Figures 1–4 show the annual TFR from 1960 to 2002, and the proportion of women using hormonal contraception (among those living in a couple and aged 18–44 years) for 21 developed countries for which data were available. The graphs are drawn at the same scales to ease comparisons. Data on fertility are easy to obtain (e.g. on the Ined Web site: http://www.ined.fr; United Nations, 2003; Sardon, 2004), but data for contraceptive use are much more scarce. Data on the annual sales of contraceptive products, and more specifically the contraceptive pill, are sometimes available, but the number of countries and years is limited (Population Reference Bureau, 1988; for France, see de Guibert-Lantoine and Leridon, 1998; Toulemon and Leridon, 1998; for UK, Murphy, 1993; for Nordic countries, Makkonen and Hemminki, 1991). In most cases, the prevalence for contraceptive use in the population of any single country can only be obtained through specific demographic surveys, which are taken only occasionally. The 21 selected countries are those for which we were able to find at least two surveys during the last 30 years. In eight cases, there were only two surveys, in six cases three surveys and in seven cases at least four surveys. We sometimes adjusted the proportions of hormonal contraceptive users given in the various publications (Leridon, 1981; Berent, 1982; United Nations Economic Commission for Europe 1988, 1989, 1990, 1991a,b, 1992, 1993, 1994a,b, 1995a,b, 1996, 1997a,b,c,d, 1999; United Nations Population Division, 2000, 2003; Population Reference Bureau, 2002) to make the data truly comparable: in some cases, the published data referred to all women and not only those engaged in a union, or the age limits were not exactly 18–45 years. We have assumed that the rates of hormonal contraception use were close to zero in 1960 and 1961 in all countries. Also, the ‘morning-after pill’ (or emergency contraception) was not widely used during the years when fertility was declining fast.
|
|
|
|
The situations are quite diverse. In a few countries, particularly western European countries, we find the expected negative correlation between fertility and pill use. In Belgium, France, the UK and the Netherlands, and to a lesser extent in Germany (current boundaries except pill use in 1985) and Switzerland, fertility rates decreased by about one child between 1965 and 1975, with the contraceptive pill being used by 
30% of women by 1975 (Figure 1). Between 1980 and 2000, the TFR remained stable in several countries, although pill use increased considerably in Belgium, France, Germany and the Netherlands. The trends are somewhat similar in northern European countries (Figure 2), although pill use remained at lower levels. Hormonal contraception is not widely used in southern European countries and therefore cannot have played a role in the sharp decline in fertility in these countries. The USA, Canada and New Zealand show similar trends (Figure 3), with a peak in pill use around 1970–1975, followed by a decline when the TFR remained unchanged.
In Japan, hormonal contraception is still not used: Japanese couples use condoms and calendar methods as well as abortion to keep the fertility rate low. The abnormally low TFR in Figure 3e is not an error: we will comment on this accident in the next section.
In former eastern European (socialist) countries, the fertility rate was already close to two children in the 1960s (Figure 4). In these countries, modern contraceptives were not available, except in Hungary where, nevertheless, fertility was also close to two. The TFR fell below this level in the 1990s, although, with the possible exception of the Czech Republic, this cannot be related to a massive diffusion of oral contraception.
A complete picture of contraceptive use should take into account other contraceptive methods, at least modern methods such as IUD and sterilization, which have been a real alternative to the contraceptive pill in countries such as the USA and Canada. However, this is beyond the scope of this article.
|
Country studies |
|---|
|
TOPAbstractIntroductionSourcesTrends in fertility and...Country studiesAdolescentsTheories on fertility declineConclusionsReferences |
|---|
Even if oral contraceptive use is increasing when fertility is declining in the same country, there might be other reasons for this decline. This suggests to analyse in more detail the situation country by country, but this is not always feasible: trends in fertility have been extensively analysed by demographers but not often in relation to the availability and use of contraception.
Fertility dropped in Belgium between 1965 and 1985, and the age at which the woman gave the first birth rose. According to de Graaf and Lodewijckx (2000), ‘[t]hese changes were made possible by, among other things, the availability and use of modern contraceptives, enabling better planning of both the number and timing of births’ (p. 11). Indeed, the proportion of Belgian women using hormonal contraceptives increased rather quickly between 1965 and 1975, but also between 1985 and 1995, when fertility was no longer declining.
In the same article, the authors also show very similar trends for the Netherlands. Ketting (1982) agreed with their analysis of this country: ‘As might be expected, the introduction of modern contraceptive methods [...] resulted in an unprecedented drop in the birth rate. [...] In addition, the annual number of abortions has remained at a lower level than in any other country following legalization of abortion’ (p. 144). Ketting also reported a rapid decrease in pill use between 1978 and 1980 because of negative publicity on hormonal contraception in the media. Induced abortions increased by 38%, but the effect on the fertility rate seems to have been limited: there is only a short plateau in Figure 1e.
Several studies on UK are available. Cartwright (1979) was among the first to suggest a causal effect: ‘The contraceptive methods increasingly being used—the pill, the IUD and sterilization [...have] probably played a most significant part in the decline in the birth rate and in the average family size of British couples. At the same time, there has also been some reduction in the number of children that parents want’ (p. 135). This conclusion is also supported by Hobcraft (1996), using data assembled by Cartwright (1987) on unwanted pregnancies: the reduction in the number of these pregnancies could account for as much as 54% of the decline in fertility between 1967 (TFR = 2.61) and 1975 (TFR = 1.78). Hobcraft estimated that ‘improved contraceptive efficacy [mainly the pill] was alone responsible for a reduction in total fertility of perhaps 0.25’ (p. 508), being 30% of the total fertility decline. Abortion may have been responsible for 15%, with the remaining decline being due to the reduction of wanted fertility. Murphy (1993) also argued that the increased use of the contraceptive pill has played a major role in the decline of fertility, against the usual views of economists: ‘It must be concluded that the effect of contraceptive use, especially switching to the pill, or possibly induced abortion, was overwhelmingly responsible for the reduction in fertility during this period [1967–1983]’ (p. 224). ‘Thus the oral contraceptive pill was the principal determinant of fertility in Britain during this period. [...] Later, however, other efficient methods, especially sterilization, more than compensated for reductions in pill use’ (p. 229).
However, Langford (1991) took a long-range view and analysed fertility and contraceptive practice over the whole 20th century, cautiously concluding that ‘data on birth control practice are of little help in the understanding of fluctuations in fertility in Great Britain since the 1930s’ (p. 62). This opinion was in part due to the poor quality and the scarcity of data on contraception before the 1960s.
England has also experienced sporadic episodes of ‘pill scare’, when information on the possible risks of cancer associated with the use of hormonal contraception was widely publicized in the media. Data on pill sales per year are available (Murphy, 1993). The first alarm took place in 1969, although the effect on sales was limited. The second scare in 1977 resulted in the reduced use of the pill. The total number of conceptions, including those ending in abortions, increased soon after, showing that unwanted pregnancies and births occurred among women who had stopped hormonal contraception but had not been immediately able to use another effective method. However, Bone (1982) claimed that two other causes could explain the rise in fertility after 1977: ‘catching up’ after a long period of postponing births or an ‘opportunity’ effect, in which some women who stopped using the pill thought that it was a good occasion to have a child that had originally been planned for later in their lives.
In France, as in the UK, the decline in fertility and the rise in pill use have had a perfectly symmetrical relationship between 1960 and 1975 (Figure 5). Leridon (1985) concluded that ‘[t]he dramatic fall of fertility after 1965 resulted from both a reduction in the number of children desired and an improved effectiveness of family planning by couples. This improved efficacy was due to, or made easier by, the development and the diffusion of more effective contraceptive methods (pill and IUD), but also by a better use of traditional methods’ (p. 521). The French survey data made it possible to quantify the decline in the number of unwanted births. This counted for 40% of the overall fertility decline between 1963 and 1977 and even more when calculated by the mothers’ cohorts (60% between generation 1933 and generation 1947). Leridon stressed the importance of a couple’s motivation for avoiding unwanted pregnancies and warned against the idea that restricted access to contraception or abortion would cause an increase in fertility: ‘The two phenomena [the decrease in desired fertility and the decline in the number of unwanted births] are strongly linked. [...] If the motivation that drove couples to avoid such unwanted pregnancies remains unchanged, couples will find means for fulfilling their aims’ (p. 522). After 1975, hormonal contraceptive use continued to rise in France, although fertility stabilized at a level that most other European countries aspire to (between 1.9 and 2.0 children per woman).
|
We have not plotted any data for Ireland because of the lack of reliable data on contraceptive use. However, fertility began to decline in Ireland later than in other parts of Europe, from a record level of 4.0 children per woman in 1965. The Catholic Church has a strong influence on Irish society, with very traditional views on reproduction: abortion is still illegal, and modern contraception is still not easily accessible. However, fertility started to decline in the 1970s and is now close to reproduction level (1.9 in 2000). Den Draak and Hutter (1996) showed that >12% of women (aged 15–49 years) were using the pill in 1992 and that the number of Irish women having an induced abortion in the UK is such that the abortion rate for Irish woman may be greater than that of Dutch women and similar to that of German women. The authors concluded that ‘[t]he decline in fertility thus should be due to an increase in contraceptive use and number of abortions performed in England’ (p. 6).
For Norway, Ostby (1989) estimated the decline in fertility that would have been expected when modern and more effective forms of contraception spread (pill and IUD) between 1967 and 1977, assuming all other things remained unchanged. Ostby assumed that all women using a modern method were not at risk of becoming pregnant and that the fertility rate of non-users of modern contraceptives remained constant. The age-specific fertility rates of non-users were first calculated for 1965, when the pill and the IUD were not available to the population. The predicted reduction in fertility was higher than the observed reduction: ‘[T]he younger cohorts have not had changes in fertility as great as the changes in contraception would imply, whereas the older ones have had even greater changes’ (p. 42). Ostby concluded that ‘the changes in the age-specific fertility pattern have not been causally dependent on the contraceptive patterns. The contraceptive revolution has of course facilitated the process of change’ (p. 41). In a study using more limited data (a sample in Tromso), Jacobsen et al. (1992) emphasized the link between the level of education of women and their use of oral contraceptives to delay childbearing. In Finland, steroid contraceptives had an even more limited role in fertility decline than they had in Norway (Riihinen et al., 1979).
Data for Sweden have not been plotted because, surprisingly, no representative survey on contraception has been carried out since 1981. According to a small postal survey carried out in 1987, the use of the contraceptive pill by women aged 20–44 years increased from 19 to 27% between 1981 and 1987 (Riphagen and von Schoultz, 1989; Althaus, 1990). Fertility, however, also slowly increased during this period.
Despite its small population (300 000 inhabitants) and particular geographical characteristics, Iceland (not included in the graphs) has experienced similar trends to those of other European countries (Snaedal, 1981). The birth rate started to decline steadily in 1965, just as the contraceptive pill was introduced. Soon after, oral contraception was used by 40% of women of reproductive age.
The picture is different in Southern Europe. Italy was one of the first countries (with Germany, and Spain a little later) to experience a drop in fertility rate to well under 1.5 child per woman. Hormonal contraception has not had much to do with this trend, with its use still being very limited in 1995 and almost equal to its value 15 years earlier. For Livi-Bacci and Salvini (2000), the main factor is a form of traditional family that allows children to stay at home with their parents while they do not have stable employment: ‘Too much family, not enough children’. Other factors include the increased length of the time spent in education, the development of female employment (as in many European countries) and the great difficulty in reconciling work with raising children. The authors make no explicit reference to changes in birth control methods or to abortion.
In another article, Delgado Perez and Livi-Bacci (1992) included new data on these factors and discussed their effect. For abortion, they stated that the law ‘seems to have reclassified as legal procedures most of the previously clandestine abortions, and thus has had little effect on abortion and fertility trends’ (p. 167). Highlighting the limited use of modern contraceptives, they concluded that ‘extremely low levels of fertility have been achieved despite a still incomplete "contraceptive revolution" ’ (p. 166).
Dalla Zuanna et al. (2001) tried to analyse the paradoxical situation of a country with ‘low fertility and scarce diffusion of modern contraception’. In the 1960s and 1970s, withdrawal was by far the most widely used technique by Italian couples. After 1980, couples began using the condom and the contraceptive pill: in 1996, 34% of couples were still using coitus interruptus, although 25% were using the condom and 24% of women were taking the pill. The condom may have often been preferred to the contraceptive pill because it required a less drastic cultural change. Withdrawal and condoms are not completely effective techniques, but abortion became legal in 1978 to compensate for contraceptive failures. According to the same authors, the reduction in unplanned pregnancies accounted for 35% of the decline in fertility between the late 1970s and the early 1990s. They concluded that ‘unwanted conceptions [and fertility] among married women therefore fell thanks to a combination of two events: an increase in the use of more effective contraception, and a reduction in the number of person-years in which the woman was exposed to the risk of an unwanted pregnancy’ (p. 7). The authors added that if Italian women progressively adopt hormonal conception, it will not be to reduce further the number of children they produce (they are already successfully using other methods) but for reasons such as improving the women’s autonomy.
In Greece, fertility followed a similar pattern to that in Italy, reaching an equally low level in 2001 (1.25 children per woman; data not shown). However, modern contraceptive use by Greek women is still very limited according to the United Nations Economic Commission for Europe (1999) and Ioannidi-Kapolou (2004). The demographic situation of Spain is also similar to Italy for fertility trends but probably rather different for contraception. Aguinaga Roustan (1989) mentioned that the use of oral contraceptive almost doubled between the surveys of 1977 and 1985 in Spain, but traditional methods (withdrawal and condoms) were still widely used as in Italy. Roustan seemed to somewhat overestimate the increase in pill use, but on the contrary, sterilization was much more frequent in 1995 in Spain (20% of couples) than in Italy (6%) (United Nations Economic Commission for Europe, 1995a,b).
The demographic situation of the USA has been extensively analysed (e.g. Ryder and Westoff, 1967; Ryder, 1972; Blake and Das Gupta, 1975; Peterson, 1995; Westoff and Jones, 1977; Westoff and Ryder, 1977; Westoff, 1983; Mosher, 1990; Henry Kaiser Family Foundation, 2002). Ryder and Westoff (1967) suggested that ‘[t]he acceleration in fertility decline between 1964 and 1965 and the simultaneous acceleration in the adoption of oral contraception is unlikely to be a mere coincidence [...] (However) It is our hunch that what has been happening to fertility in the 1960s would have happened in direction if not in degree even if the oral contraceptive had not appeared on the scene’ (p. 3). Both Ryder and Westoff have constantly maintained that the spread of modern contraception (and sterilization) played a role in fertility decline but that motivation was the primary factor. Blake is as strong an advocate of this position: ‘We conclude, therefore, that the decline in American fertility since the early 1960s does not constitute an exception to the primacy, as demonstrated both historically and cross-culturally, of motivational factors in fertility reduction. [...] Individuals will employ such methods only when they have, in their judgment, good reasons for doing so’ (Blake and Das Gupta, 1975, p. 246). We will return to this in the next section.
As in the UK, the USA have experienced campaigns against the pill. Westoff and Ryder (1977) showed a rise in the rate of discontinuation of pill use between 1966 and 1970, which they related to adverse publicity given to the pill by several scientific papers on thromboembolic risks and the Nelson hearings in the Senate (pp. 46–48). They do not mention that the American TFR stopped declining in 1968, 1969 and 1970 and then continued its decline afterwards.
In Canada, the baby boom has been followed by a very fast change in social and demographic behaviour, especially in Quebec: fertility fell abruptly from 3.9 children in 1960 to 1.4 in 1985 in Quebec. In 1984, 25% of women (aged 18–44 years, living in Quebec) were using the pill, and 35% had been sterilized (or their partner was) (Rochon, 1989). The predominance of these two techniques is even more apparent when we look at specific age groups: more than half of the women aged 18–19 years were using the pill, and almost 40% of couples where the woman was aged 30–34 years were sterilized. The easy access to these techniques has certainly played a major role in the fall of fertility in this population.
Data for Australia are less detailed. The first analysis was by Caldwell and Ware (1973) from a survey in the city of Melbourne and allowed them to reconstitute contraceptive use from 1935 to 1970. The contraceptive pill appeared in 1960–1963, and its use reached a plateau in the mid-1960s. The decline of fertility started later, in the mid-1970s. Pool et al. (1999) regarded the pill as the main factor in changing reproductive behaviour in New Zealand.
The pill is still virtually not utilized in Japan. Japanese couples have managed to keep the fertility rate low using just condoms and periodic abstinence (Ogino-Knaus). The low value for the TFR in 1966 in Figure 3e should be explained. Fertility in that year fell to 1.58, compared with 2.14 in 1965 and 2.23 in 1967, a surprising temporary 26% reduction. In the absence of any major economic, social or political crises, the only plausible explanation to have been advanced is related to astrological beliefs of the Japanese (Biraben, 1968; Jitsukawa and Djerassi, 1994). The year 1966 was named as the year of the ‘Fire Horse’ (Hinoeuma), which is a conjunction of two astrological cycles that happens every 60 years. Girls born under the sign of the ‘Fire Horse’ are supposed to find it very difficult to get married. As couples had no reliable method for selecting the sex at conception, many preferred to avoid having a child that year. This explanation is supported by certain anomalies in the sex ratio in birth records. In January 1966, the sex ratio increased to 111 versus 101 in December 1965: many girls who were unfortunate enough to be born during the first few days of January were registered as being born in December to improve their future chances of marriage, thus increasing the sex ratio for January. A similar situation was observed in December 1966, with a sex ratio of 117 versus 99 in January 1967 (Biraben, 1968, 1969). This ‘accident’ in Japanese demography is another good example of the capacity of a population to modify abruptly its fertility, at least temporarily, with no change in the availability and use of contraceptive techniques. Goodking (1991, 1993) also mentioned this when reporting a similar case resulting in an increase in the birth rate among Chinese populations: the ‘Year of the Dragon’.
We will now look at some examples from former socialist countries of Eastern Europe. In this part of Europe, trends in fertility were not homogeneous. Hungary was the second country to cross the two-children threshold (preceded only by the former German Democratic Republic). Abortion was liberalized in the mid-1950s, and Hungary was the only socialist country to produce and market contraceptive pills (since the early 1970s). Despite this very particular situation and despite the high number of the pill users and abortions, Kamaras (2000) wrote that ‘no direct long-term relationship between the trends in the number of induced abortions and in the number of live births can be demonstrated’ (p. 271). Moreover, he stated that ‘per se, birth control, through either contraception or induced abortion, does not affect fertility: it is used to plan births and to prevent unwanted births’ (p. 270, both citations have been translated by the current author).
In Poland, as in Italy, where marriage remains the main form of cohabitation for couples, fertility remained at a high level (round 2.3 children) until the mid-1980s then dropped to very low levels (1.3 in 2001). Sobczak (2000) suggested two factors: the difficulty in finding a separate place to live, resulting in the postponement of marriage and the increasing desire of women to work (especially in private company) that cannot be reconciled with their ‘mother’s duties’ (p. 60). Despite the very low level of fertility, Kuciarska-Ciesielska (2000) reported that a 1995 survey showed that 23% of women who had been married for 10 years said they had had more children than they wanted. This was as many as those who had had fewer children than expected. Among those who did not want any further children, 20% were not using any contraceptive method, 44% used withdrawal and 43% used some kind of rhythm method (women could report more than one method). The contraceptive pill was not even mentioned in the list. A previous article by Kuciarska-Ciesielska (1993) based on less representative samples of 1985 and 1987 showed that contraceptive methods were classified as either ‘natural methods’ (in fact, Ogino-Knaus) or ‘anticonception methods’ (mainly coitus interruptus).
We did not include Albania in our comparative analysis because data on contraceptive use were not available before 1990. However, Albania is another example of drastic reduction in fertility without any modern methods of contraception. At the beginning of the Communist regime in 1945, the TFR was greater than six births per woman. ‘When Albania emerged from behind the "Olive Curtain" in the early 1990s, in the intervening period fertility had fallen to around three children per woman, despite a pro-natalist Marxist regime and in the virtual absence of contraception and abortion’ (Falkingham and Gjonca, 2001, p. 310). The decline in fertility started in 1961. ‘The decline must have been brought about by means of traditional forms of birth control—coitus interruptus and withdrawal. Both these methods require a high degree of motivation for successful use and, perhaps more importantly, both are methods that require the co-operation of the man involved’ (p. 315). This is not obvious in a strongly patriarchal society. (NB: we do not clearly see the difference between the two methods mentioned by the authors: coitus interruptus and withdrawal are usually considered synonymous.)
Fertility trends in Romania have been quite erratic. In 1967, the TFR jumped to 3.67 children, double that of the previous year. This was due to a sudden drastic restriction on abortion. In the 1960s, there were more abortions than live births (maybe twice as many as live births). President Ceaucescu suddenly repealed the liberal law, and thus, all pregnant women had to bring their pregnancy to term. The number of live births doubled in 1968 and even tripled during the first few months. With no reliable contraceptive techniques being available, women found it very difficult to adapt to this new situation (Johnson et al., 2004), although fertility started to decline in 1969 and was back to two children in 1983.
The conclusion of this section is unambiguous: within individual countries, there is no systematic correlation between fertility and contraceptive pill use. The development of hormonal contraception cannot be considered as responsible for either starting or the size of the fertility decline. However, the link may be more complex, as we shall discuss in Theories on fertility decline section.
|
Adolescents |
|---|
|
TOPAbstractIntroductionSourcesTrends in fertility and...Country studiesAdolescentsTheories on fertility declineConclusionsReferences |
|---|
Adolescents deserve special treatment when studying contraceptive use and fertility rates. The rate of pregnancy among adolescent women varies greatly between countries. Among western countries, it has always been especially high in the USA, with the UK coming second (Jones et al., 1985
, 1986; Murray, 1986; The Alan Guttmacher Institute, 2001). By 2000, the teenage birth rate for American girls was six times higher than that for their French or Swedish counterparts. Most of these births are not planned and are the result of either absence of contraception or ineffective use of contraceptive methods. Two comparative studies have recently been conducted. Darroch et al. (2001) showed that hormonal contraception (and IUD) was used much less by 15- to 19-year-old US women than in other countries. Jones et al. (1985) already suggested 20 years earlier that the use of these methods was low compared with many European countries. According to Bajos et al. (2003), ‘the drop in fertility [for teenagers] over the past few decades stems from the use of modern contraceptive methods in northern and western European countries’ (p. 10). Contraceptive use at first intercourse has indeed shown a rapid increase during the last 30 years. Among the nine western European countries reviewed, the prevalence of contraceptive use at first intercourse ranged from 20 to 50% among cohorts born between 1945 and 1950 and from 55 to 90% among cohorts born between 1975 and 1980. For women aged 15–19 years living with a partner, the pill accounted for between 50 and 96% of all methods used. However, the same authors stressed the important consequences of changing attitudes about family and sexuality, the effect of higher education and the increasing participation of young people in the labour force.
Netherlands is considered as a very good example of high prevention against unwanted adolescents births. De Graaf and Lodewijckx (2000) stated that ‘young Dutch people do not have less sexual intercourse nor do they start having sex later than their peers in other countries and there are no indications that Dutch teenagers are less fertile’ (p. 29). Three-quarters of first-time intercourse are well protected, by the contraceptive pill and/or condom, and contraceptive use is even higher afterwards (Rademakers, 1991). The morning-after pill may also contribute to the low rate of unwanted births (Ketting, 1982). A longitudinal study in Finland has shown a significant decrease in pregnancy rates among 16- and 18-year-olds, mainly because of a parallel increase in pill use (Rimpela et al., 1992). However, a more limited sample in Gothenburg, Sweden, showed that in the early 1980s, a high rate of contraceptive pill use at 19 years of age was followed by a high discontinuation rate (Andersch and Milsom, 1982), often resulting in accidental pregnancies and abortions.
In summary, easy access to effective contraception is usually considered crucial for reducing teenage fertility, but the analysis is often compromised by the competing use of condoms and by the varying recourse to abortion in cases of accidental pregnancy.
|
Theories on fertility decline |
|---|
|
TOPAbstractIntroductionSourcesTrends in fertility and...Country studiesAdolescentsTheories on fertility declineConclusionsReferences |
|---|
The recent decline in fertility has usually been associated with an increase in the age at which a woman first has a child. This may indeed reduce the TFR (which is a period variable), even if the final number of children in mothers’ cohorts remains unchanged. Therefore, if we explain the postponement of childbearing, we can possibly explain the overall decline in fertility. Sobotka (2004)
has extensively studied the situation in Europe and concluded that (i) none of the European countries could have reached its current level of fertility without postponement and (ii) the ‘lowest-low’ level of fertility in some countries is likely to be only a temporary phenomenon. However, the postponement itself has been driven by the availability of new contraceptives: ‘Oral contraception appears to have been instrumental in allowing and even facilitating fertility postponement’ (Sobotka, 2004, p. 25). Berrington (2004) has suggested the designation ‘perpetual postponers’ for couples who are ‘always maintaining either a positive or ambivalent intention to have a child but delaying to some date in the future and ultimately reaching the end of their reproductive years childless’ (p. 10). This suggests that postponement may be responsible for the low levels of both period fertility and cohort fertility.
We will now review the more theoretical approaches of the many researchers on how the availability and use of contraception, and especially of oral contraceptives, affect fertility. The dominant view is that there is no single theory to explain the changes affecting fertility over the last century. One reason may be that there are too many different situations over time and between countries. We can identify at least three major steps. During the first half of the 20th century, demographers began to realise that fertility decline was an irreversible reality in many industrialized countries, which they included in the general concept of a ‘demographic transition’ (Landry, 1934; Notestein, 1945). Later, historical data were gathered to analyse the transition, and several syntheses have been published (Coale and Watkins, 1986; Chesnais, 1992; Gillis et al., 1992). Although almost no data for the period are available, to most authors, it is obvious that couples were able to control their fertility more and more effectively through contraception and abortion. This occurred although the only existing techniques were withdrawal and periodic abstinence, and later, condoms and caps. The second phase started in the 1960s, when the rate of world population growth was accelerating because—in developing countries—fertility remained high, whereas mortality was already declining. The importance of the availability of contraceptives was central to many analyses, but this approach often underestimated the psychological barriers to contraceptive use: couples may want small families, and modern contraceptives may be known about and accessible, but couples can, however, be reluctant to use them. The third step in theoretical studies on fertility started in the 1980s, when it became apparent that fertility was not going to stabilize at two children per woman in many populations, the level that ensures a constant population over the long term. At this point, the ‘demographic transition’ was more or less assumed to have finished. Also, the different factors involved were no longer the same, with the drastic changes affecting the values, norms and family in modern societies: the growing individualism, secularization, mass education, women’s employment, new relations within couples etc. We will address these issues below. This is why it has been suggested that this should be called the ‘second demographic transition’ (Lesthaeghe and Van de Kaa, 1986; Van de Kaa, 1987). However, whether this period required a new name or not has been disputed (Cliquet, 1991; McDonald, 2001; Coleman, 2004), because the first theories of the demographic transition made no mention of the final levels of fertility and mortality.
Many observers became rather sceptical at this unexpected challenge. In the overview of a meeting on the determinants of fertility trends, Leridon (1982) wrote that ‘[a]t the present stage of research, we have to accept the fact that there is no general theory of fertility determinants, even tentative ones, which might explain fertility movements and differentials as a whole [...]. We have quite a comprehensive list of variables from which can be taken a part to suit the needs of a particular piece of research in a specific context’ (p. 304). This view was supported by Lesthaeghe and Surkyn (1988): ‘When surveying the recent literature on fertility and family formation, one is struck by the increasing fragmentation and even competition between the various social science disciplines’ (p. 1). Van de Kaa (1996) also stated along these lines: ‘The quest for the determinants of fertility behaviour and change during the last half-century can be best interpreted as the development of a series of sub-narratives from different disciplinary perspectives and orientations’ (p. 389). ‘There does not yet exist a single "good story", accepted by all knowledgeable scholars’ (p. 390). Kirk (1996) was no more optimistic towards the end of the 1990s: ‘[Our] review may leave an impression of chaos’ (p. 379). However, Kirk highlighted the ‘admirable effort to provide system and order’ in the two books, resulting from a working group set up by the National Academy of Sciences, although this work mainly dealt with the situation in developing countries (Bulatao and Lee, 1983). More recently, McDonald (2001) argued in favour of a multidimensional approach: ‘Transition theories, like other grand social theories, have not been resilient to empirical tests’ (p. 2). ‘I suggest that we release ourselves from the hegemony of the paradigm of the demographic transition (even more from the paradigm of a second demographic transition)’ (p. 3). ‘Explanations of low fertility are likely to be found in different weightings for different societies from among the range [of theoretical paradigms] that I propose’ (p. 4). In a recent book on reproduction and the family in Switzerland, Sauvain-Dugerdil reviewed the various dimensions of the theoretical framework for such a study (Le Goff et al., 2005, pp. 4–11). Contraception and abortion are almost totally absent, at least for the interpretation of the trends over the last decades. A full chapter is, however, devoted to contraceptive practices in Switzerland but with no attempt to estimate their impact on fertility at either the individual or the national level.
There are also technical problems for doing multifactorial studies. One is the type of data needed. Most fertility or sociological surveys are cross-sectional, and it is difficult in such studies to measure the original determinants of fertility decisions taken years before. But the absence of agreement on a general theory does not rule out the possibility and value of more specific theories, with each of them emphasizing the role of a limited number of variables. As Murphy (1993) wrote, ‘Three main areas have been suggested as possible determinants of post-war fertility changes in developed societies: economic factors, especially those concerned with women’s employment and relative incomes; cultural or attitudinal change, particularly associated with assertive individualism and secularization; and contraceptive and related technological factors such as improved contraceptive methods, induced abortion, and sterilization. Ideally, a full analytical model should include all these factors, together with some possible additional ones’ (p. 221). However, this approach is rarely taken, for several reasons. First, each specialist tends to give a high priority to data pertaining to his or her own discipline and to consider that all other variables are more or less dependent on these primary factors. Second, data are not always available for the entire period being studied: e.g. contraceptive use is only known when specific and representative samples are taken, which often occur 
10 years apart in a given country. Third, not all the variables listed above have the same status. Since the works of Davis and Blake (1956), Bongaarts (1978) and Bongaarts and Potter (1983), it is common for demographers to consider variables such as marriage, contraception, induced abortion and, where applicable, breastfeeding as being either ‘proximate determinants’ or ‘intermediate variables’ of fertility. These are ‘proximate’ because the level of fertility depends directly on them (e.g. a higher age at marriage most often results in a lower final number of children). However, they are only ‘intermediate’ variables because they are influenced, in turn, by factors such as the level of income, education or religion. Therefore, it is not recommended to mix all the types of variables in a single model: ‘Contraception is a proximate determinant of fertility and, therefore, is different in nature from the economic or cultural explanations of fertility change, which are remote determinants [...] Thus a demonstration that changes in a proximate determinant, such as contraception or marriage, account for changes in fertility may be regarded as a relatively low level of explanation, since one would naturally look for an explanation for why these changes in the proximate determinants took place’ (Murphy, 1993, p. 236).
What do the more restricted theories tell us? The economic approach is usually based on two major changes that took place in the decades after the Second World War. The first is the increased participation of women in the labour market. Consequently, the traditional roles of men (the bread winners) and women (the house keepers, taking care of the children) were less easily accepted. It became increasingly difficult for working women to take care of the house and the children, especially when there are more than two children, which caused a reduction in fertility. The ‘New Home Economics’ approach developed models to analyse the relationship between the proportion of working women and the level of fertility, taking into account the real and the relative wages of these women (Willis, 1973; Schultz, 1974; Becker, 1981). This approach has also been linked to Easterlin’s hypothesis, which related the actual income and expected income of the baby boomers to those of their parents (Lee, 1974; Butz and Ward, 1979; Easterlin, 1980), in a way that could generate cycles. Murphy has vehemently argued against the economic approach (Murphy, 1992, 1993), mainly because the changes that occurred in contraceptive methods and use were not taken into account.
A second approach looks to sociological factors such as values and norms. The declining influence of religion, the increase in individualism and the importance attached to the couple as a way of personal fulfilment have all diminished the traditional forms of marriage and of the family and the value of children (Lesthaeghe, 1983; Van de Kaa, 1987; Lesthaeghe and Willems, 1999): ‘The story of the Second Demographic Transition as told by its proponents, is the quintessential narrative of ideational and cultural change’ (Van de Kaa, 1996, p. 425). ‘The notions of individuation, embourgeoisement, and civil religion [...] form the cornerstone of our interpretation of changes in Western family formation’ (Lesthaeghe and Surkyn, 1988, p. 3). This was already being highlighted by Westoff (1983): ‘Such social changes [erosion of traditional and religious authority, growth of individualism, urbanization, mass education, increasing equality and independence of women, consumerism], when combined with modern contraceptive, [...] make very low fertility quite comprehensible’ (p. 101). Also, as these major sociological changes do not seem to be reversible, it is often concluded that their effects on fertility are not likely to disappear soon. As Murphy (1993) elegantly stated, ‘I argue that the change in fertility is essentially irreversible—the genie of efficient contraception has been let out of the bottle and cannot be reinserted’ (p. 239). However, Morgan (2003) has recently given a more optimistic view.
The general change in the norms regarding religion, sexuality and rules of life for couples also included a change in the number of children desired. Fahey and Zpéder (2004) summarized the Eurobarometer data on ideal family size. The reported ideal average number of children (for 28 European countries) was 2.47 for women aged 55 years and more, 2.31 for those aged 35–64 years and 2.14 for those <35 years of age. This decrease can be interpreted as a change in the successive birth cohorts, with a 13% reduction between cohorts born before 1947 and those born after 1967. For individual countries, the mean number for the younger age group ranged from 1.72 in Austria and 1.74 in Germany to 2.58 in Ireland. Zero was given as the ideal number of children by 17% of women aged 18–34 years in Germany (13% in Austria), and 19% of German women preferred only one child (14% in Austria). The emergence of less than two children being the ideal number has been analysed by Goldstein et al. (2003).
In fact, there is much agreement nowadays on the idea that we must accept multifactorial causes and, above all, the principle of recursiveness between the variables: ‘This article attempts to reintegrate the sociology and economics of fertility. [...] We also want to introduce recursiveness and reverse causation’ (Lesthaeghe and Surkyn, 1988, p. 2). Westoff et al. (1963) suggested this, and many authors now agree. ‘In fact it’s all connected. Couples want less children, and because they want less children they improve the effectiveness of their contraceptive practice [...] and they are more prone, in case of failure, to have an abortion’ (Leridon, 1985, p. 251; Bajos et al., 2004). ‘Obviously, the availability of more efficient contraception can change family size preferences (e.g. by allowing a woman to plan and conduct a career), while the intensity of family size preferences can influence the success of birth control practice [...]. The recent fall in fertility is related to declining proportions married, increased use of contraception and abortion, and a reduction in family size desires’ (Preston, 1986, p. 44). ‘But we also know that the possibility of regulating fertility itself creates a stronger wish for regulation, because such wishes can be fulfilled. Thus contraception has direct, as well as indirect, effects on fertility’ (Ostby, 1989, p. 38). ‘It is, moreover, quite possible that the fundamental changes in technology led to the sequence of changes in family formation behaviour so characteristic of that transition’ (Van de Kaa, 1996, p. 422). ‘All societies with unconstrained access to fertility regulation experience a rapid decline to replacement levels of fertility, and often lower [...]’. Unconstrained access requires, among other conditions, that ‘at least three reversible methods of contraception [are] available through a minimum of two channels of distribution’ (Potts, 1997, p. 8).
Even when all this is taken into account, some authors still stress the paramount importance of modern contraception. After reviewing seven major possible causal interpretations [mortality; economic theories; the Caldwell approach mixing economic, cultural and institutional variables (see Caldwell, 2001); cultural and ideational theories; historian’s views; governmental policies and diffusion processes], Kirk (1996) wrote that ‘[a]nother aspect of causation is the role played by contraceptive technology which, in some cases, is the most obvious cause of demographic change’ (p. 379). In this list, family planning policies are included in item 6. Murphy (1993) considered that the use of the contraceptive pill has played a dual role, as being both the main ‘proximate determinant’ and also the ‘principal remote determinant’, because contraceptive use can be driven by purely contraceptive considerations rather than by other remote determinants such as intending to have fewer children. For Van de Kaa (1996), ‘[t]here is evidence that improved birth-control technology played a central role in the reduction of unwanted fertility in these societies, and thus was instrumental in removing excess fertility’.
|
Conclusions |
|---|
|
TOPAbstractIntroductionSourcesTrends in fertility and...Country studiesAdolescentsTheories on fertility declineConclusionsReferences |
|---|
Despite the often-discordant views that we have reported above, some conclusions can be drawn:
- The diffusion of modern contraception has certainly contributed to the reduction in the number of unwanted pregnancies and, with abortion, to the reduction in the number of unwanted births.
- These new methods have been adopted by individuals and couples because of the major changes of attitude towards sexuality, the nature of marriage and other forms of union, the place of women in societies and, more specifically, the position of women in the work place.
- These contraceptive techniques have also facilitated and favoured the adoption of new (more restrictive) norms for the ideal family size.
- However, motivation always comes first: when couples are not worried about how many children they have, a baby boom may occur, with many ‘not really wanted’ births. When couples want to avoid births for any reason, they can largely succeed even without elaborate contraceptive technology, even if some unwanted births may still occur.
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionSourcesTrends in fertility and...Country studiesAdolescentsTheories on fertility declineConclusionsReferences |
|---|
The author thanks Martine Deville (INED) for her help in the preparation of the bibliography.
|
References |
|---|
|
TOPAbstractIntroductionSourcesTrends in fertility and...Country studiesAdolescentsTheories on fertility declineConclusionsReferences |
|---|
Aguinaga Roustan J (1989) Descenso de la fecundidad y modernizacion en la sociedad espanola. Analisis comparativo de las encuestas de fecundidad 1977 y 1985. Bol Asoc Demogr Hist 7,7–22.[Medline]
The Alan Guttmacher Institute (2001) Teenage sexual and reproductive behavior in develope