|
REGIONAL POPULATION MEETING |
|
| (Budapest, 7-9 December 1998) |
|
| |
| Location: Statistical Division, Regional Population Meeting | |
|---|---|
|
|
| STATISTICAL COMMISSION and ECONOMIC COMMISSION FOR EUROPE CONFERENCE OF EUROPEAN STATISTICIANS | GOVERNMENT OF HUNGARY | UNITED NATIONS POPULATION FUND (UNFPA) |
|
Distr. ENGLISH Original: FRENCH Regional Population Meeting Budapest (Hungary), 7-9 December 1998 Fertility in Europe and North America prepared by Gérard Calot Director European Demographic Observatory Saint-Germain-en-Laye, France Document prepared by Gérard Calot for the Regional Population Meeting (Budapest, 7-9 December 1998). The opinions expressed in this document are those of the author and do not necessarily reflect the opinions of the Government of Hungary, the Economic Commission for Europe or the United Nations Population Fund.
* The Regional Population Meeting is in the work programme of the Conference of European Statisticians. GE.98-32453 (E) NOTE
CONTENTS Page Overview of developments in western Europe and North America 4 Overview of developments in eastern Europe 16 Total fertility rate and completed fertility 18 Total first marriage rate and proportion of ever-married by age 50 in birth cohorts 19 Causes of variations in fertility 20 Possible government action to influence the fertility rate 22 Future fertility. Demographic projections 24 New forms of family life and social policy 26 For intensified international cooperation 27 Fertility in Europe and North America In the course of the last 50 years, fertility in Europe and North America has seen very contrasting developments, both in space and in time. Let us first consider the developments in western Europe and North America, and then those in eastern Europe. With regard to western Europe, we will distinguish four zones: the northern zone (Denmark, Finland, Iceland, Norway and Sweden), the western zone (Belgium, France, Ireland, the Netherlands and the United Kingdom), the central zone (Austria, Germany, Luxembourg and Switzerland) and the southern zone (Greece, Italy, Portugal and Spain). Overview of developments in western Europe and North America In western Europe and North America, where the demographic transition was largely completed or in the process of being completed, the period between the two wars, marked notably by the economic crisis of the 1930s, was characterized by generally low fertility rates, lower in a good many countries than the generation replacement rates (figure 1). The authors of the first population projections, made before the Second World War, did not anticipate any significant resurgence. However, shortly afterwards came what would eventually be called the baby boom. This went through its ascending phase during the war years themselves in two European countries - Sweden and Switzerland - that remained outside the conflict. Elsewhere in western Europe, the boom did not become fully apparent until the end of the hostilities, but it then manifested itself with great vigour, even if taken as a whole, in terms of total rates, the increase in fertility between the early 1930s and the late 1950s in a country like France (about +0.8 child per woman) was identical to that observed in Sweden or Switzerland (figure 1, graph at bottom right). In North America the baby boom assumed still larger dimensions than in Europe, the total rates for the United States and Canada reaching, respectively, 3.0 and 3.4 children per woman in 1950, then 3.7 and 3.9 in 1957 (figure 2). On the other hand, in Austria and Germany, where fertility had increased in the late 1930s, there was no rise from 1945 to 1955. In southern Europe, where the demographic transition came later, the baby boom was hardly noticeable, fertility being still high around 1950. During the period 1955-1965, total fertility rates generally showed an upward trend in western Europe and North America, owing partly to lower ages at marriage and, consequently, at maternity; added to the normal annual number of births (i.e. that which would have been observed all things being equal) was the surplus resulting from these lower ages of marriage and childbearing. The lowering of the age at marriage was by then a well-established phenomenon, which had started to become manifest shortly after the First World War; it was to continue until the mid-1970s (and even up to the beginning of the 1980s in southern Europe), with a shift towards earlier family formation. The increasing rarity of higher-order births (four or more) combined with the lowering of the age at marriage had the effect of reducing the average age at maternity. The late 1950s in North America, the mid-1960s in western Europe and the early 1970s in southern Europe were marked by a sudden and drastic shift in the pattern of fertility (figures 2 and 4). Within a single decade, the total rates fell by more than one child per woman, a variation probably without precedent in human history during untroubled times (with no wars, food shortages or epidemics). In the first place, it is especially fertility after 25 years or at fairly high parities (women already having at least two children) that is the most affected (figure 3A). At the end of 5 to 10 years, the fertility of the first part of fertile life begins in its turn to decrease, whereas the trend is reversed after 25 years: consequently, age at maternity not only ceases to fall but actually begins to rise very rapidly (figure 5) to the point where, after about 10 years, it reaches values close to those observed a half-century earlier. Furthermore, the combination of the lowering of the rates at younger ages of fertility and the rise in those rates at higher ages has had the effect for the past 20 years of keeping the total fertility rates for various countries (Austria, Belgium, France, Germany, Netherlands, Switzerland, United Kingdom) relatively steady or in moderate decline, but at a low level. Northern Europe towards the mid-1980s nevertheless registered a certain recovery in the total rates (graph at top left in figure 4), although this seems to have lost momentum after a few years. The most spectacular case is that of Sweden, where the indicator rose from 1.6 to 2.1 children per woman between 1983 and 1991, then fell back to 1.5 in 1997. In North America, while the total rate for Canada has remained at approximately 1.6-1.7 children per woman since the early 1980s, that of the United States, which never dropped below 1.8, has almost attained the generation replacement level since 1990. In southern Europe, where the rise in fertility rates at advanced ages is still not very marked and where the fall at the beginning of fertile life is rapid (figure 3A for Italy), total rates have reached the lowest levels ever observed (1.1 to 1.2 children per woman in Spain and Italy in 1997: graph at top right in figure 4). The fall in under-25 fertility coincided with a rapid decrease in the number of marriages. The total first marriage rate had usually ranged in Europe between 90 and 110 first marriages per 100 females, corresponding within female birth cohorts to a proportion of never-married by age 50 of around 5 to 10% (figure 6). The total first marriage rate initially began to fall in northern Europe (as from 1965 in Sweden) and then, from the first half of the 1970s onwards, the decline gradually spread throughout western Europe. Today, the rate in many countries is scarcely 50 to 60 first marriages per 100 females, and has even been lower than 50 since 1990 in Sweden (1997: 45). This fall in the total first marriage rate reflects a shift from legal marriage towards cohabitation, with marriages themselves being contracted at increasingly later ages (figure 7). While losing a large part of its social significance, marriage has at the same time become very sensitive to changes in social and tax legislation. Thus, in Austria on three occasions (1972, 1983 and 1989) and in Sweden (in 1989), sudden increases in the number of marriages were recorded mainly in the last quarter and especially in December of the years in question because of new laws entering into force on the following 1 January (figure 6). The fall in nuptiality has entailed an often very marked increase in the proportion of births occurring outside marriage (figure 8). This proportion is currently close to or above 50% in northern Europe (Denmark, Estonia, Iceland, Norway, Sweden) and near or higher than 30% in various other European countries (Austria, Bulgaria, Finland, France, Latvia, United Kingdom) and in North America. In southern Europe the present levels are significantly lower, but the increase is widespread. Divorces, unlike marriages, are becoming more frequent: the total divorce rate today is around 40 to 50 divorces per 100 marriages in northern Europe and generally about 30 to 40 elsewhere in Europe, with the exception of southern Europe; in the United States, the divorce rate has been above 50 since the 1970s, whereas in Canada it has hovered around 40. While the post-war period was characterized by forms of family life still conforming to the traditional model - almost universal and increasingly early marriage, low incidence of divorce, consensual unions and out-of-wedlock births - there is today a great diversity of family life cycles: numerous situations that at one time were barely tolerated or were reproved by society are today voluntarily chosen and socially commonplace. This has even led to a change of vocabulary, certain words with pejorative connotations (illegitimacy, concubinage) being dropped and new terms (reconstituted families, cohabitation) being adopted. Family law, and more generally social law, has not always managed to adapt to these changes. Systems of statistical observation are also experiencing difficulties in adjusting to them. Overview of developments in eastern Europe There were still only a few countries in eastern Europe which had completed their demographic transition by the end of the Second World War. The Communist regimes in those countries liberalized abortion during the 1950s, well before western Europe (except in the case of the German Democratic Republic, which did so only in 1970). As a result, the reduction in fertility was particularly accentuated: by 1965, the two countries with the lowest fertility rates in the world were Hungary and Romania (respectively 1.8 and 1.9 children per woman). Thus, a number of east European Governments retreated, at least in part, from this liberalization and instituted various policies to support the family (family allowances, day-care centres, assistance for housing) with the explicit aim of stimulating fertility. The trend in total fertility rates as a whole was therefore downward, but marked especially by sudden rises resulting from abrupt changes in legislation. We may cite the cases of Bulgaria (peaks in 1968-1969 and in 1974), Hungary (peaks in 1954, 1968 and 1975), Romania (spectacular peak in 1967, corresponding to a leap in the annual rate from 1.9 to 3.7 children per woman, and another peak in 1986), Czechoslovakia (1974) and Russia (peaks in 1983 and 1987). Lastly, during the 1970s and 1980s, fertility in eastern Europe, at slightly above the generation replacement level, exceeded that of western Europe. Early family formation was, and indeed still is, a feature common to most countries of eastern Europe. It was not unusual for the fertility rate at age 20 to be close to or even higher than the rate at age 25, a phenomenon never observed in western Europe (figures 3A and 3B). By contrast, the rates at age 30 or more are significantly lower in eastern than in western Europe. In the event of difficulties affecting young people more especially, a pattern of early childbearing is probably more likely than one of later childbearing to show a downward trend. Since the collapse of the Communist regimes, together with the economic crisis, rising unemployment and reduced social budgets to support families, total fertility rates have suffered a particularly drastic fall, bringing them within a few years to the levels observed in the West and, in some cases, to even lower levels. In 1997, the rates were 1.5 children per woman in Poland, 1.4 in Hungary, Lithuania, Slovakia and Ukraine, 1.3 in Estonia, Russia and the Czech Republic, and 1.1 in Latvia and Bulgaria. In the countries of the former USSR where fertility was generally higher than in Russia itself, particularly in the Republics of the Caucasus, the fall is roughly parallel to that observed in Russia or Ukraine. Only the States of the former Yugoslavia, where fertility is nevertheless continuing to decrease, seem to have escaped this drastic downturn. Slovenia, however, scarcely exceeded 1.2 children per woman in 1997. Albania, for which the fertility rate was still above 3.5 children per woman in 1980, is rapidly joining the other countries in its geographical area. The differences in fertility levels prevailing until the late 1980s have therefore been heavily blurred during the present decade. While the United States and Ireland in the West and Albania in the East are still special cases, no country of the ECE region today has a TFR higher than the generation replacement level. Total fertility rate and completed fertility Before discussing the present situation as regards fertility in Europe, it might be worth recalling the significance of the total fertility rate (TFR) and of completed fertility for female birth cohorts. Age-specific fertility rates are established for a given region and a given year: the fertility rate at, for example, age 25 is the ratio of the number of children born to mothers aged 25 to the number of women aged 25 residing in the region. The total fertility rate for the year under consideration is obtained by adding the various age-specific fertility rates (e.g. for ages 15 to 49) established for that year. Like each of these rates, the total rate is expressed as the number of live-born children per woman. It can be shown that the TFR is 2.05 children per woman when the size of the female birth cohort born during the year under consideration is equal to the mean size (weighted by the age-specific fertility rates) of the various female birth cohorts which that year are of childbearing age (cohorts then aged 15 to 49). This result can be refined somewhat and stated as follows: the TFR is 2.08 (usually rounded up to 2.1) when, taking account of mortality but not of any migration, the future size of the female birth cohort which has just been born is equal, when that cohort itself reaches childbearing age, to the current mean size of the female birth cohorts of an age to be mothers. In other words, the total fertility rate, divided by 2.1, measures the relationship between the size of the birth cohort just born and the mean size of the birth cohorts currently of childbearing age. That is why this value of 2.1 children per woman is called the generation replacement rate. A TFR of 1.4, for example, means that, in the absence of international migration, the size of the birth cohort just born, when it reaches childbearing age itself, will be a third lower (1.4/2.1 = 2/3) than the mean size of the birth cohorts currently of childbearing age. The completed fertility of a female birth cohort, i.e. all females who are themselves born in a given year, is the mean number of live-born children which these women have had during their fertile life. This number may be found by a census or survey conducted shortly after the birth cohort has passed the age of 50. It can be shown that the completed fertility of a birth cohort may also be obtained by adding the age-specific fertility rates for the same birth cohort observed over successive years, and thus at successive ages of the childbearing lifespan. The temporal pattern of completed fertility is much more regular 1/ than that of the TFR (compare the lower graphs in figure 4 with the corresponding graphs above): two neighbouring birth cohorts experience the same events at nearly the same ages and in general react to them in a similar way. There are many factors which condition the temporal pattern of the TFR. Variations in this indicator may derive from particular events which have the effect, for example, of lowering the rates of fertility at most ages, but without the completed fertility of the birth cohorts concerned being modified significantly, because corrective adjustments will be made later. They can also result from variations in completed fertility across birth cohorts. They may furthermore derive from variations in the timing of maternity. In particular, a lowering of the age at maternity, with completed fertility unvarying, has an inflationary effect on the TFR, while a rise in age at maternity has a deflationary effect. The factors contributing to variations in the TFR over a given period cannot be singled out until all information concerning the birth cohorts then of childbearing age becomes available, i.e. about 15 years at least after the period under consideration. Likewise, a birth cohort's completed fertility cannot be evaluated without too much uncertainty until it has completed the larger part of its fertile life. In practice, the cohort must be aged at least 30 years in the latest available statistics, and preferably at least 35 years. Whatever factors cause variations in the TFR, it has to be understood that this, combined with the legacy of past natality (possibly corrected slightly by mortality and migration) which is constituted by the members of the various birth cohorts currently of childbearing age, is what determines the number of births per year. 2/ It is therefore the TFR, and not completed fertility, that conditions the evolution of the magnitude essential for the functioning of the demographic machinery: the annual number of births, upon which the future age composition of the population will depend for many decades. Nevertheless, the durability or non-durability of given changes in the TFR is to be appreciated differently according to the change in completed fertility within the birth cohorts concerned, and that change will unfortunately not be known until about 15 years later. This is why monitoring changes both in the TFR and in completed fertility is indispensable for the analysis of fertility. It should, however, be kept in mind that the most recent data available to us concerning completed fertility inevitably lag behind the data available for the TFR by a good decade. Total first marriage rate and proportion of ever-married by age 50 in birth cohorts Rather as when studying fertility, each year we establish age-specific first marriage rates, for both males and females, i.e. the ratio of the number of never-married of the sex in question and of a given age who marry in the course of the year to the size of the resident population of the same sex and age during the year. The total first marriage rate for females, for example, is obtained by adding the age-specific first marriage rates for females. The ages taken into consideration are usually limited to less than 50 years. It may be shown that the proportion of ever-married by age 50 in a given birth cohort which can be observed in a census or survey is generally close to the result obtained by adding the before-50 first marriage rates for the same birth cohort over successive years. For this to be so, international migration must not be too large and the difference between the number of marriages of residents contracted outside the region and the number of marriages of non-residents contracted within the region has to be small. It can also be shown that the first marriage rate of a given sex for a given year is equal to the ratio of the number of first marriages for the sex considered which are contracted in the course of the year to the mean size (weighted by the age-specific first marriage rates) of the birth cohorts then of an age for first marriage. When, for example, the first marriage rate for females is 50 marriages per 100 women, this means that the number of never-married women who married in the course of the year represents half of the mean size of the female birth cohorts then of an age to contract a first marriage. Up to the generations born around 1945, which were not concerned by the development of cohabitation, the proportion of ever-married by age 50 varied quite little, remaining at around the 90% mark, from one birth cohort to another, and even from one sex to the other and from one country to another (lower part of figure 6). However, the total first marriage rate could vary much more widely under the influence of specific events (economic crises, wars, changes in legislation, etc.). Except during particular periods of tension in the "marriage market", resulting from discontinuities in the age pyramid, 3/ the first marriage rate for males changes in parallel with that for females, so that it suffices to follow the variations in only one of the two rates. The rate for females is the indicator more commonly chosen. Causes of variations in fertility It must be recognized from the outset that we do not know the root causes of the most striking variations in fertility. Are we today on the verge of a new baby boom? It would be very daring indeed for anyone to make such a prediction in the present circumstances, but excluding the possibility completely would surely not be reasonable. The baby boom of the 1940s and subsequent years surprised all observers at the time, who realized the magnitude of the phenomenon only after the event. The same applied also to the baby bust of the 1960s. We can, of course, identify without ambiguity the causes of sharp variations that follow major events, such as the collapse of the communist regimes in eastern Europe or the almost total ban on abortion in Romania in October 1966. However, other than with regard to such exceptional circumstances, we can only make suppositions about the causes on the basis of correlations and associations that are always uncertain. Hence it is extremely difficult for analysts to forecast medium- or long-term developments and for governments, when they want to, to influence those developments. We have no satisfactory explanation for the baby boom. Looking in particular at the Swedish and Swiss data, as compared with the data for other countries that were belligerents (Belgium, France, the Netherlands, Norway and the United Kingdom), one might think that it corresponded, in the late 1930s, to a kind of historical necessity: it became manifest first in the countries where the Second World War did not prevent it from happening, and then, after the end of the war, but very forcefully, in those countries where the war had in some way contained it. In any case, it was not the war itself that gave rise to the baby boom, whose magnitude far exceeded the recovery which had followed the First World War. Was it a delayed effect of that First War, and of the considerable losses and major social changes it had entrained, was it an effect of the economic and moral crisis of the 1930s, or was there some other cause? Concerning the baby bust, it does appear that control of reproduction, which became more and more total and increasingly widespread, played a crucial role. Various national studies indicate that the decline in the number of births went hand in hand with a certain decrease in the numbers of children who were wanted and with a marked, and significantly larger, reduction in the numbers of unwanted children, that being made possible by greater fertility control. Furthermore, the appearance of the baby bust in North America several years before western Europe and the decrease in fertility first of all after age 25 (easier access and greater motivation to use contraceptives, fewer reservations among the medical profession) and then, only a few years later, at the beginning of fertile life and in particular before marriage, corroborate this view. The later entry of the countries of southern Europe into the phase of rapid decline in fertility could be connected with a lesser degree of openness in southern societies to the idea of fertility control and to the later diffusion of modern means of contraception. As is frequently the case, a phenomenon that had for a certain time been contained later manifested itself with even greater vigour. Together with the dissemination of modern means of controlling reproduction, other major developments occurred within European societies, some causally linked and some not to such control, but most of them being facilitated or reinforced by it. First of all, we should note the development of the gainful employment of women outside the home. This greater participation by women in the labour force derives in part from the progress made in schooling for females and the desire of many couples to raise their standards of living whenever possible, given the huge difference that exists between a family budget sustained by one wage earner and a two-income family budget. Moreover, in a society where occupation is a determinant of social status, this increase in female employment has played a major role in improving the situation of women. With unions, both consensual and legal, tending to become more fragile, a gainful occupation for a woman represents a form of insurance against the risk of a break-up. The continued rise in female activity rates, despite the economic crisis and the persistently high level of unemployment, attests to the vigour of this change: women's participation in the labour force is an irreversible datum, something that is taken for granted. It seems very probable that the increase in female activity rates has depressed fertility. Thus, government action to enable couples, and especially women, to combine family and occupational responsibilities more easily is of the utmost importance. Indeed, it is in the countries of northern Europe, where social policies have been implemented to provide facilities for early childhood (nurseries, kindergartens, etc.), paid parental leave and part-time working, that fertility has generally declined the least. Some other major trends may also be referred to in connection with today's low fertility: the decline, or virtual disappearance, of farming, small-scale trading and handicrafts, sectors in which wives could work and still look after their children, the generalization of urban ways of life, the secularization of society and the reduced influence of religions, the rise of individualism and hedonism, the greater sensitivity to differences in standards of living and the encouragement of consumerism, particularly under the influence of the media, and so on. All of these changes have a common feature: they are very probably irreversible. It is therefore difficult, while recognizing our limited capacity to predict, to see how fertility could spontaneously, i.e. without deliberate government action, increase in a significant fashion. Indeed, an examination of the variations over a lengthy period in the three countries represented in figure 1 leads one to wonder whether European fertility has not quite simply reverted to its long-term trend, a trend from which it mysteriously departed during the interlude of the "Forty Glorious Demographic Years" (broadly the period 1935-1975). Must it then be deduced that post-transitional fertility could register large pulsations, of which the Swedish fluctuation between 1983 and 1997 would be the second example after the long interlude of four decades? We are clearly not in a position to answer those questions, but they are worth asking. Possible government action to influence the fertility rate Whatever the case may be, a basic human right has been progressively acquired: the right to have only the children one wants at the time one wants them. Even if, here and there, some imperfections remain, as witnessed by the still high proportion of adolescent pregnancies in some countries, it may be said that, overall, the European and North American populations have today attained a high degree of control over their fertility. Considering the monthly apprehension, or even anxiety, of so many women just three to four decades ago, such developments can only be welcomed. No democratic government can therefore seek to influence the fertility rate by limiting that basic right in any way whatsoever. Some people think that government intervention in family choices (encouragement of marriage and of larger families, particularly through social and fiscal policies) is not legitimate: the State must refrain from interfering in any way in what is a strictly private domain. Even advocacy of one family choice rather than another constitutes unacceptable pressure on the individual. Others believe, on the contrary, that the State, as guarantor of the collective interest, must be concerned with the nation's harmonious development in the short, medium and long term. The number of foot-soldiers that might have to be mobilized for a country's defence is, of course, no longer a crucial parameter in modern Europe. However, a government cannot be indifferent to population changes because the future functioning of society depends on them, at least in part. Social welfare systems, for example, have an important demographic component. If action by the State to stimulate fertility is not considered to be illegitimate, one must still make sure that it does not have the effect of in any way, directly or indirectly, increasing the number of unwanted children; it must be aimed primarily at enabling couples to have the number of children they actually want, and secondly at encouraging society itself to want children. The instruments for such a policy are naturally the same as those of social policy, and are designed to attenuate the economic and other disadvantages of having children. All fields of economic and social policy are concerned, but three are particularly important: employment and working conditions, action to deal with differences in the standards of living of households according to the number of children they have (allowances, taxation, community services, particularly for the care of small children, education, culture and recreation) and housing. The demographic argument is rarely put forward when government decisions in favour of the family and children are taken. The social concern is, of course, often the main or even the only argument. Above all, politicians are reluctant to invoke it for fear of appearing hostile to change that bears the hallmark of modernity. In addition, the political regimes which have insisted the loudest on the demographic argument are not among those one likes to remember and still less to recommend. Furthermore, a lowering of fertility for several decades is generally received very favourably by all the actors of social life. For households, it affords the opportunity of using for secondary, but more gratifying needs, the resources that would otherwise have been devoted to children. Employers have to cope with fewer absences for maternity or child illness. The State reaps an immediate benefit from a fall in the number of births: less expenditure on maternal and child health care for the social security system, less school facilities to be provided, fewer family allowances to be disbursed, less tax relief to be granted. However, this is probably a nation that is reducing its investment in human capital, rather like a business that cuts back on investment spending: it immediately improves its cash flow situation, but leaves itself exposed to problems at a later stage. Moreover, there is quite general scepticism about the demographic effectiveness of family policies, particularly in view of their cost for public finances. The impossibility in the present circumstances of undertaking controlled experiments, which would help to determine more precisely the effectiveness of such policies, further reinforces this scepticism. However, while it is true that we have virtually nothing but suppositions to work on, these are on the whole sufficiently convergent to allow us to think that there is indeed some effectiveness. We may cite the case mentioned earlier of the Scandinavian countries and particularly that of Sweden, where the total fertility rate and especially completed fertility have varied relatively little over a lengthy period (figure 1). Similarly, the fact that fertility in France has been slightly, but almost constantly, higher than that of its neighbours, especially in terms of completed fertility, is probably attributable to its family policy. Certain episodes of European history point in the same direction. Thus, Saarland, where French family legislation was applied from 1945 to 1955, before its return to Germany, had at that time one of the highest fertility rates among the 11 federal Länder. Thirty years later, it had one of the lowest, although developments in the other Länder largely paralleled one another. All the same, the demographic effect of family policies is limited. To give an order of magnitude in terms of completed fertility, one could suggest a figure of about 20 or perhaps 30 children per 100 women. On the basis of econometric calculations, one French researcher, Olivia Ekert, estimated at 50 children per 100 women the likely effect of a policy whereby the community would bear the full economic "cost" of children. It seems certain therefore that, if any demographer had been farsighted enough to announce the arrival of the baby bust by 1965 in western Europe, no policy would have been able to offset its effects. However, particularly in countries where the completed fertility of cohorts born after 1960 is likely to deviate significantly from the replacement level (160 children per 100 women in some countries for the cohort born in 1962; see the lower part of figure 4), the implementation of an active policy might be desired by some governments because it would make it possible to fill perhaps not all but at least a substantial part of the gap in the simple replacement rate. Some think that the demographic deficits created by low fertility can always be offset by immigration. It is true that for a century in some European countries international migration has been an appreciable and appreciated element of population change. This has generally been prompted by the desire to meet manpower needs rapidly. While candidates for migration to Europe, and even more so to North America, probably run little risk of not having a real future, it might be asked whether their skills will match the future requirements of the post-industrial economies. Moreover, the widening culture gap between the countries of origin and the host countries will not facilitate the migrants' integration. Above all, however, the demographic deficit of many European countries is on such a scale as to raise the fear that the volume of immigration necessary for its absorption might undermine the internal cohesion and balance of the host countries. Indeed, to ensure simple replacement with a fertility rate of 1.6 children per woman, a quarter of the adult population (1.6/2.1 • 0.75) would have to be comprised of immigrants. Thus, although immigration is bound to remain an adjuvant for population change, we cannot hope to use it, when fertility becomes very low, as a tool for redressing the balance. Future fertility. Demographic projections Despite our limited capacity to understand and consequently to predict, we are forced to make forecasts. Even if clumsy and imprecise, and even if periodically revised, as are the world demographic projections of the United Nations Population Division, they nevertheless help us to look forward, translating at least the idea that we have, at a given time, about the future. We should note that not all elements of a forecast have the same degree of uncertainty. The present weighs to a greater or lesser extent on the future, if only by reason of a certain continuity in temporal changes. In the demographic field, this weight of the present stems above all from the fact that the future size of a cohort already born at the starting point of the projection is, while the horizon is not too far removed, close to its initial size and linked to it by a relationship that is not too hard to evaluate with a satisfactory level of precision. A demographic projection, at least in its commonest form, involves translating, in terms of future numbers of the population (by sex and age), assumptions regarding the three factors of population change: fertility, mortality and migration. The projection method normally used, i.e. the cohort-component method, is the converse approach to that of analysis. Whereas the analysis of past observed data leads to the determination, for each year in the past, of fertility and mortality tables (collection of age-specific rates for fertility) and balances of migration by sex and age, the projection successively builds up the future numbers of the population on the basis both of the current observed numbers and of the series of tables and balances used to make assumptions for future years. A projection exercise usually considers not one but several temporal paths for each of these factors. The projections concerning the European Union countries made by Statistics Netherlands at the request of Eurostat thus envisage, for each of the 15 countries, three paths (middle, high and low) for each of the three factors. The combination of these various paths for any one country involves 33 = 27 projections, called scenarios, but only 5 have been drawn up and published to cover the period up to 2050. The central scenario combines the three middle paths for each of the three factors. A fertility path defined for a given country may be summed up as the development from 1995 to 2050 of its total fertility rate. Statistics Netherlands has systematically assumed the TFR to be invariable as from 2020 and change in the TFR, as observed before 1995, has been linked 4/ to the level projected for 2020. Lastly, the vision of the future incorporated into this projection exercise relies essentially on the values expected to be attained by the TFR as from 2020. In no European Union country is the TFR expected to rise above the generation replacement level (2.1 children per woman). Only on the high paths, and for only five countries (Finland, France, Ireland, Sweden and the United Kingdom), does the TFR attain this value in 2020. By contrast, the low paths of three countries (Germany, Italy and Spain) correspond to a TFR of 1.3 children per woman in 2020. This vision of a future of persistent low fertility is very generally shared by statistical offices when they make national projections. Thus, in the case of France, the latest levels are respectively 1.6, 1.8 and 2.1 children per woman in the Statistics Netherlands exercise and 1.5, 1.8 and 2.1 in the INSEE exercise. The Federal Statistical Office of Switzerland, a country which does not belong to the European Union, makes very similar choices - 1.35, 1.6 and 1.8 children per woman - while Statistics Netherlands has arrived at 1.4, 1.6 and 1.9 for Austria and 1.3, 1.5 and 1.8 for Germany, a country where change in fertility for the past half-century has been close to that observed in Switzerland (figure 4). These assumptions, which revise downwards very heavily the projections which the countries had made two or three decades ago, seem reasonable in the present context. Nevertheless, they are very fragile and are bound to undergo numerous revisions between now and 2050. Yet this fragility does not invalidate the analyses that can be made of the ageing of European populations. That is because the deformations in the age pyramid to be expected between 2010 and 2040 result from specific events that are of known magnitude, since they belong to the past. On top of the long-term ageing resulting from lower mortality and lower fertility, which many European countries have already been experiencing since the first decades of the twentieth century, there will be an exceptional ageing due to the following combination of circumstances: the 1945-1975 birth cohorts, as compared to preceding generations, are characterized by both higher numbers and lower fertility. This combination has had a stabilizing effect on the absolute number of births for the past 30 years, but in the first decades of the next century it will have the effect of amplifying the process of population ageing. Even if fertility were to increase markedly in the short or medium term, that increase would have no impact on the balance between the ages of gainful employment and the ages of retirement for at least 25 years. New forms of family life and social policy The transformations which have been recorded in family life over several decades, together with the development of cohabitation and the declining interest in legal marriage, the greater incidence of divorce and the rising number of births occurring outside marriage, call for various adjustments in social policy. At the time when couples were almost all married, social legislation very often referred to marital status as a way of defining an individual's social situation. This reference was convenient and made other evidence unnecessary because marriage is an event determining civil status that is duly recorded and verifiable, with divorce governing the alteration of that status. Thus, the rules for calculating income tax, for example, may refer to marital status and result in a total tax that differs according to whether the couple is married or not, with more favourable treatment varying in relation to the particular circumstances (existence of dependent children, number of incomes and difference between the amounts received). Some social benefits are provided to unmarried persons with low incomes and having children to care for; it is not uncommon for priority in assigning places in nurseries to be given to lone parents. With regard to social rights, the risk of illness of the non-working spouse of a wage earner participating in a social security scheme is covered by the wage earner's contribution to that scheme. A widow's pension is generally half the amount of the pension drawn by the husband before his death. These particular arrangements, cited by way of example, clearly correspond to the family model of the immediate post-war period: few consensual unions, few divorces, few births outside marriage and few gainfully employed women. Matrimony today no longer has the significance it once had in European societies. A large number of unmarried persons live together as couples and, among under-25 year-olds, in various countries, married couples by no means constitute the majority. Since society does not object to these situations, the State must respect the personal choices made in the private domain: strict neutrality must be its rule and the law has to treat all couples in the same way, whatever the form of their union. But the question which then arises is whether the law needs to be changed so that it can be applied on a purely individual basis, without reference to the concept of the couple, or whether it should continue to distinguish between couples and persons who are really lone individuals. In some fields, such as taxation, pension schemes or inheritance, for example, the question is debatable but, as regards social assistance to disadvantaged persons, the answer can only be negative. Furthermore, with a view to replacing the marriage certificate, which serves as proof in the case of the married couple, various alternatives have been proposed, such as a certificate prepared on the basis of a simple declaration, and perhaps attested by two witnesses, to the effect that the couple live together, or a contract drawn up with a notary or an administrative service. However, aside from the fact that the debate is made more complicated and heated by the demands of cohabiting homosexuals to be treated like heterosexual couples, including for purposes of adoption or medically assisted procreation, these alternatives can hardly have the same evidentiary value as a paper issued by a civil registry office that would document not only marriage but non-marriage. Furthermore, verifying the individuals' actual situation, while superfluous in the case of marriage, otherwise poses both moral problems and practical problems that quickly become inextricable. Also, in many fields of law, the European countries are faced with making careful changes in their legislation to adapt to the evolution of the family. There is, lastly, another area in which the government's vigilance is of the utmost importance, namely that of reproductive health. In particular, there is a need for ongoing and in-depth analysis of data not only about births but about abortions, obtained on the basis of a high-quality system of observation. If shifts become apparent, showing for example that younger generations are reducing their contraceptive practices and resorting more often to abortion, or that adolescent pregnancies are increasing significantly, the country's public health authorities should be alerted without delay. The same applies to the monitoring of sexually transmitted diseases, especially AIDS. The efforts of the educational system in the field of sex education should also be pursued unrelentingly. For intensified international cooperation In most areas directly or indirectly connected with population and the family, the rapidity and extent of the changes have already led or will increasingly lead States to amend their laws, with a view to adapting them to developments in society and perhaps also with a view to influencing some of those developments. Regular exchange of information between the European countries, as regards projects and accomplishments in the legislative and regulatory field, as well as concerning demographic trends themselves, is particularly useful. Various international organizations are helping to facilitate such exchanges: the Economic Commission for Europe, the Council of Europe and the Commission of the European Union. The creation of the Observatory on Family Policies within the European Union is also a response to that concern. Now that the continent is no longer divided politically, and now that demographic situations have become very similar and that their effects are substantially the same, as also very probably are their causes, efforts must be redoubled. While, as we have pointed out above, our difficulties of understanding and prediction are considerable, there is no doubt that common reflection can only be beneficial to all. Notes |