Fertility in Germany before and after the 2011 Census: Still no Trend Reversal in Sight Fertility in Germany before and after the 2011 Census: Still no Trend Reversal in Sight Olga Pötzsch Abstract: The population projections by the German Federal Statistical Offi ce show a likely decrease in the number of births in the 2020s. This development will be the result of a declining number of prospective mothers combined with an assumed continued low fertility rate. Given the available empirical fi ndings prior to the 2011 census, there was no indication that a possible distinct rise in the fertility rate in the next decade would compensate for the declining number of potential mothers. However, the 2011 census led to revisions in the population size, age structure and consequently in relative fertility measures such as period and cohort fertility rates. The objective of this article is to quantify the effects of the 2011 census on these fertility statistics and to check the validity of previous fi ndings for fertility trends on the census-adjusted data basis. A special focus is laid on analyses of the cohort fer- tility and the consequences of ever-later entry into motherhood on the completed fertility and on parity distribution. Using numerous fi ndings, we will show that a continuous rise in the completed fertility in the coming two decades cannot be real- ised without a reversal of fertility behaviour. A greater increase in fertility from the age of 30 onwards would be necessary to offset the decrease in fertility for ages un- der 30 – a trend which intensifi ed with cohort 1974 – and thereby stabilise the total cohort fertility rate at a relatively low level between 1.5 and 1.6 births per woman. A rise and subsequent stabilisation of the total cohort fertility rate at the level of at least 1.6 births per woman would, additionally, necessitate a trend reversal in the development of childlessness and distinct changes in birth timing. Keywords: Postponement and recuperation of births · Development of fertility rates · Total fertility rate · Cohort fertility · Childlessness · Census 1 Introduction Following the 2011 census, which enabled adjustments to population projections for the fi rst time in nearly a quarter century, demographic statistics also need to be reassessed. This also applies to the relative fertility measures, which, in addition Comparative Population Studies Vol. 41, 1 (2016): 87-118 (Date of release: 30.06.2016) Federal Institute for Population Research 2016 URL: www.comparativepopulationstudies.de DOI: 10.12765/CPoS-2016-02en URN: urn:nbn:de:bib-cpos-2016-02en2 • Olga Pötzsch88 to the absolute numbers, also describe the general development of fertility. They show the proportion of the actual number of children born to the number of poten- tial mothers, i.e. women of childbearing age (in this case between 15 and 49 years). Adjusted population fi gures thus have an infl uence on fertility rates via the female population (macro data level) while also forming the population for extrapolating the microcensus and thereby infl uencing its weighted sample results (micro data level). The microcensus is the largest random sample survey of private households in Germany which gathers information about the number of children per woman born until the time of survey every four years since 2008. This data provides the basis for the distribution of female cohorts according to parity as well as for the development of childlessness. This article examines birth trends incorporating the effects of the 2011 census on the main fertility measures. Emphasis is put on relevant aspects of fertility be- haviour, which, from today’s perspective, can also be signifi cant for future fertility developments. Respective fi ndings will be analysed here and combined to form a comprehensive picture of fertility. After presenting the data sources and methods in the second section of the ar- ticle, we describe effects of the 2011 census on the measurement of fertility in the third section. The choice of indicators employed is aligned to internationally com- mon and widely used fertility indicators. These include period-related fi gures such as the number of women in the respective cohorts who are potential mothers and thus belong to the so-called “population at risk” as well as the age-specifi c and total fertility rates. In the cohort perspective, these are mainly the cohort fertility rate and cumulated fertility rates at a selected age as well as the parity distribution among the female cohorts. The snapshot of the fertility level in the year 2011 before and after the census is the starting point for the adjustment of fertility trends and the assumptions for population projections. In the fourth section, current fertility trends and assumptions regarding future fertility developments based on new data after the census are discussed. While the number of prospective mothers in the next two decades will largely result from the previous demographic development, the future fertility rate has to be estimated on the basis of fi ndings about fertility behaviour. One approach chosen for this is based on studies of cohort fertility. The central question is how the completed fertility rate develops under the infl uence of a continued delay of fi rst births. The follow- ing analyses build upon the studies on the childbearing postponement and recu- peration (Frejka/Calot 2001; Frejka/Sobotka 2008; Sobotka et al. 2011; Frejka 2012; Pötzsch 2013), about interaction between parities and their impact on the cohort fertility (Statistisches Bundesamt 2009, 2013; Sobotka 2011; Bujard/Lück 2015a) as well as on the development of childlessness (Dorbritz/Ruckdeschel 2007; Sobotka 2011; Konizka/Kreyenfeld 2013; Statistisches Bundesamt 2013, 2015a; Beier et al. 2012; te Velde et al. 2012). Preliminary estimations of the completed fertility published so far go back to the cohorts born in the mid- to late 1970s (Goldstein/Kreyenfeld 2011; Pötzsch 2010a, 2013; Myrskylä et al. 2013; Statistisches Bundesamt 2015a). There is no doubt that the declining trend among women born in the late 1960s ends and is replaced by a Fertility in Germany before and after the 2011 Census • 89 slight increase in the completed fertility until cohort 1973. Afterwards, the cohort fertility in Western Germany appears to stagnate while it probably increases further in Eastern Germany (cf. Pötzsch 2013: 95-97; Myrskylä et al. 2013, Table 2 and Ap- pendix Fig. 1). In the following, we will analyse the overall fertility trends in Germany on the basis of the census-adjusted results and discuss perspectives of the fertility development up to the 1984 cohort. 2 Data and methods The relative fertility indicators usually refer to the so-called average population. Un- like the reference day population, the average population considers variations in the population size during the year, which usually and mainly are caused by migration. The average population therefore offers a better approximation to the “population at risk” as reference for fertility rates. The population at risk is an estimation of the lifetime in which this population was exposed to “the risk of a birth” (Luy 2016: 123- 125). To calculate the age-specifi c fertility rates, we also need the average popula- tion broken down by single ages. The offi cial statistics use two calculation methods for this: completed age and birth years (cohort). In this article we use the population broken down by cohorts because it permits a long-term view. Results according to completed age are only available from 2000 onwards. For clarifi cation’s sake, the cohort fertility is usually understood as the complet- ed number of children that a female cohort had born on average. The cohort total fertility rate (CTFR) is calculated as the sum of the age-specifi c fertility rates of a female cohort ascertained in the calendar years in which this cohort was in the ages between 15 and 49 years. The age-specifi c fertility rate is a ratio between the num- ber of births by mothers of a cohort c at a certain age, Bc(x), and the number of all women of this cohort at the same age, Pc(x), in one calendar year. The additive character of the CTFR allows us to subdivide these according to age spans by varying the parameters α and β. In this article, this character will be used to illustrate the different age-specifi c trends in the younger and older fertile ages as well as the interaction between the postponement and recuperation of births. This is done, for example, by adding both the age-specifi c fertility rates of the female cohorts aged between 15 and 29 years and of those aged between 30 and 49 years. The sum of the two results leaves us with the CTRF: The analysis of the cohort fertility rate in Section 4.1 is based on an approach by Frejka and Calot (2001) and its further development by Sobotka et al. (2011). This )x(P )x(B )x(fCTFR c c c 1 2CTFR)x(f)x(f)x(f 49 15 c 29 15 49 30 cc • Olga Pötzsch90 approach for calculating the recuperation index was simplifi ed here and slightly modifi ed in order to present the development of postponement and recuperation of fertility as clearly as possible. The two processes are observed separately and are compared with one another. Beforehand, the deviations between the cumulative age-specifi c fertility rates of the observed cohort and the cohort selected as a refer- ence were measured for each age year. The greatest deviation marks the “trough” at which the cumulative fertility of the observed cohort differs the most from the reference cohort. Once the trough is crossed, the recuperation phase begins. If the “postponed” births are not completely recuperated, there is a residual deviation between the completed fertility rates at the end of the fertile phase. However, if the deviation at the trough is compensated or overcompensated for by recuperation at an older age, then the cohort total fertility rate of the observed cohort will be the same or even higher compared to the completed fertility of the reference cohort. (For more details on the basic reference model and application examples for Ger- many see Sobotka et al. 2011, Section 3.2, and Pötzsch 2013). Theoretically, the completed fertility of the female cohorts could also be ascer- tained based on micro data, i.e. the information about the number of children born provided by respondents in the microcensus. The total number of children born to women of a cohort would be divided by the number of women in that cohort. This average fertility rate is, however, only comparable with the CTFR to a limited extent. For one, these two indicators refer to different populations (Pötzsch 2010b: 171-172; Kreyenfeld et al. 2012: 394). For example, because of migration and mortality it is highly probable that the composition of the 1950 female cohort at the time of the microcensus survey in 2012 was a different one than in the years 1965 to 1999, from which the 35 age-specifi c birth rates of this cohort (between the ages of 15 and 49) originated. Secondly, the microcensus results are marked by a sampling error and may be biased because of non-responses. An imputation method is employed in order to minimise systematic biases caused by non-responses (Statistisches Bun- desamt 2009, 2013). Nonetheless, some of the cases remain unknown. In the 2012 microcensus, this mainly affects the distribution of mothers according to the num- ber of children born. Hence, special attention needs to be paid when calculating average numbers of children per woman. If only known cases are included, the number of mothers without any information on the number of children is distrib- uted proportionally to all parities including the parity 0 (i.e. no child). This approach increases the weight of the childless in this average fi gure. In addition, the percent- age of missing information about the number of children born varies depending on the cohort, e.g. between 1 percent and 8 percent in the 2012 microcensus, which further limits the comparability of the results among the cohorts. In spite of these limitations, the microcensus is the main data source for ascer- taining the rate of childlessness and the distribution of mothers according to the number of children. The microcensus is also used to assess the effects of the cen- sus-based revision of the population on the cohort fertility rate. To do so, the results of the 2012 survey are compared based on the extrapolations before and after the census. Since the considered results refer to the same cohorts and are based on the same sample, the average fertility rate per cohort before and after the census can Fertility in Germany before and after the 2011 Census • 91 also be compared. In this case it is less a question of the average number of children per woman than of the extent of the shifts caused by the new extrapolation. For many indicators, it remains advisable to portray the development not just for Germany as a whole, but also for Western and Eastern Germany. In spite of con- verging trends, differences in women’s fertility behaviour remain signifi cant (Statis- tisches Bundesamt 2013). Although it is geographically located in Germany’s east, Berlin is not counted as “Eastern Germany” since Berlin’s fertility patterns are more similar to Western Germany’s. Due to its great weight, incorporating Berlin into the results for Eastern Germany would dilute the eastern German fertility trends and bias the comparison with the western part. Berlin is, however, included in the re- sults for Germany as a whole. 3 Effects of the census on the main fertility fi gures 3.1 Female population as reference value for the fertility rates The total number of potential mothers – i.e. women between the ages of 15 and 49 – in the year 2011 dropped in the census-adjusted population by 321,000, or 1.8 per- cent (Table 1). In Western Germany, the revision was at only -1.5 percent and thus lesser than in former Eastern Germany at -2.0 percent. While the number of women of German nationality remained almost unchanged, there was a greater need for revisions among women with foreign citizenships. The number of women with for- Tab. 1: Women aged between 15 and 49 years – revisions after the census 2011 Region Average annual population in 2011 Changes after the census BEFORE the census AFTER the census absolute relative 1,000 People % Women only Germany (as a whole) 18,209 17,888 -321 -1.8 Western Germany1 14,810 14,589 -221 -1.5 Eastern Germany1 2,554 2,504 -50 -2.0 German women Germany (as a whole) 15,959 15,895 -64 -0.4 Western Germany1 12,811 12,797 -14 -0.1 Eastern Germany1 2,461 2,429 -32 -1.3 Women with foreign citizenship Germany (as a whole) 2,250 1,993 -257 -11.4 Western Germany1 1,999 1,792 -207 -10.3 Eastern Germany1 92 75 -18 -19.0 1 without West Berlin or respectively without East Berlin. Source: Statistisches Bundesamt, intercensal population updates, own calculations • Olga Pötzsch92 Fig. 1: Women aged between 15 and 49 years1 – relative changes of the average annual population in 2011 after the census -5 -4 -3 -2 -1 0 15 20 25 30 35 40 45 Age Germany Western Germany* Eastern Germany* Percent -25 -20 -15 -10 -5 0 5 15 20 25 30 35 40 45 Age German women Women with foreign citizenship Percent 1 Age reached at the end of 2011. * without West Berlin or respectively without East Berlin Source: Statistisches Bundesamt, intercensal population updates, own calculations and design Fertility in Germany before and after the 2011 Census • 93 eign citizenships between the ages of 15 and 49 had been corrected by -11.4 percent when looking at Germany as a whole, in Western Germany by -10.3 percent and in Eastern Germany by -19.0 percent. Therefore, the percentage of women with a for- eign citizenship among all women of childbearing age dropped from 12.4 percent to 11.1 percent. Changes in the individual birth cohorts within the childbearing phase are rel- evant for the age-specifi c fertility rates. Germany-wide there is a relatively distinct pattern (Fig. 1). The revision of the number of women between the ages of 19 and 27 was above average (to -2.5 percent). As shown in the lower segment of Figure 1, these changes are mostly due to the revisions among the women of foreign nation- ality. By contrast, the deviations among women of German nationality are negligibly low. In Eastern Germany, there has been a greater need for revisions concerning the number of women aged 40 to 45 years. Since the number of births among this age group is rather low, it seems resonable to suggest only minor effects on the com- pleted fertility rate in 2011. At the same time, this might indicate that out-migration Fig. 2: Women aged between 15 and 49 years1 – relative changes of the average population size in 2011 after the census, in the federal states Berlin Hamburg Württemberg Saxony Thuringia Saarland axony-Anhalt ower Saxony cklenburg-V. e-Westphalia Brandenburg Bavaria Hesse wig-Holstein Bremen nd-Palatinate -5,9 -4,3 -2,6 -2,4 -2 -1,9 -1,9 -1,7 -1,7 -1,5 -1,5 -0,9 -0,9 -0,7 -0,6 0,3 Berlin Hamburg Baden-Württemberg Saxony Thuringia Saarland Saxony-Anhalt Lower Saxony Mecklenburg-V. North Rhine-Westphalia Brandenburg Bavaria Hesse Schleswig-Holstein Bremen Rhineland-Palatinate Germany -1,8 1 Age reached at the end of 2011. Source: Statistisches Bundesamt, intercensal population updates, own calculations and design • Olga Pötzsch94 of young women then aged between 20 and 25 years had not been fully included in population updates directly after German reunifi cation – due to missing registra- tion data. Therefore, the number of women of respective cohorts might have been slightly too high. For the cohort measures, such as the completed fertility, it is also important how the revisions affected the older female cohorts. The adjustments among the cohorts 1930 to 1961 – birth cohorts for which fertility rates are available – were far lesser, with an average of 0.9 percent, than among the younger women. We cannot preclude that the uncertainties were in part greater in each of the years between the censuses when the women of these cohorts were both of childbearing age and in their “mobile” phase. Nonetheless, the low deviation at the time of the census 2011 is an important indication that the cohort-specifi c fertility rates apparently do not contain any major biases. The need for revisions varied in the federal states between +0.3 percent in Rhineland-Palatinate and 5.9 percent in Berlin (Fig. 2). The city-states of Berlin and Hamburg recorded the greatest deviations between the census results and the pop- ulation updates between the censuses. 3.2 Effects of the census on the 2011 fertility indicators When calculating the fertility rates, the census adjustment means that only the refer- ence – the female population – in the denominator of the formula (1) changes, while the number of births in the numerator remains unchanged. The census revision therefore affects the relative birth fi gures in a comparable ratio to the revision of the number of women, but with reversed signs. From the demographic perspective, the age-specifi c and aggregated fertility rates are particularly interesting, such as the total fertility rate (from the period perspective) and the cohort total and cumulated fertility rate up to the respective age (from the cohort perspective). In 2011, when incorporating the census revision, the total fertility rate in Ger- many was 1.39 births per woman (Table 2). This fi gure was 2.0 percent higher than before the census (1.36 births per woman). In Western Germany the change was 1.8 percent (from 1.36 to 1.38 births per woman) and in Eastern Germany 2.0 per- cent (from 1.43 to 1.46 births per woman). For women of German nationality, the effects of the population revisions on the total fertility rate, with an average of 0.3 percent, were hardly noticeable. Among women of foreign nationality, in contrast, it rose by 15 percent from 1.58 to 1.82 births per woman. The revisions of the birth rates in the federal states ranged from 0.5 percent in Rhineland-Palatinate to 6.3 percent in Berlin. Besides Berlin, the federal states with a particularly great need for revisions included Hamburg (4.6 percent), partly due to the high percentage of women with foreign citizenships in the female populations in both city-states (approximately 16 percent each among women between the ages of 15 and 49). Since, according to the census, the foreign population dropped con- siderably, the fertility rates in the city-states rose to an above-average degree. Fertility in Germany before and after the 2011 Census • 95 The age-specifi c fertility rates for all women only changed slightly; therefore, the distribution of fertility rates according to the age of the mother remained almost constant. The average childbearing age in Germany even after the census revision was 30.5 years, the median was 31 years. However, the highest fertility rate, that of the 31-year-olds, rose from originally 98 to 100 children per 1,000 women. In Western Germany and in Eastern Germany only a slight increase in the maximum frequency was recorded. Tab. 2: Total fertility rate 2011 – necessary revisions after the census 2011 Region/Nationality BEFORE the AFTER the Changes census census absolute/relative Children per woman % Women (total) Germany 1.36 1.39 0.03 2.0 Western Germany1 1.36 1.38 0.02 1.8 Eastern Germany1 1.43 1.46 0.03 2.0 German women Germany 1.33 1.34 0.00 0.3 Western Germany1 1.32 1.32 0.00 0.1 Eastern Germany1 1.43 1.45 0.02 1.2 Women with foreign citizenship Germany 1.58 1.82 0.24 15.0 Western Germany1 1.60 1.81 0.22 13.5 Eastern Germany1 1.48 1.82 0.34 22.8 Federal states Baden-Württemberg 1.36 1.39 0.04 2.8 Bavaria 1.34 1.36 0.01 1.0 Berlin 1.31 1.40 0.08 6.3 Brandenburg 1.41 1.44 0.02 1.6 Bremen 1.27 1.29 0.01 1.1 Hamburg 1.26 1.32 0.06 4.6 Hesse 1.39 1.40 0.01 0.6 Mecklenburg-Vorpommern 1.42 1.44 0.02 1.5 Lower Saxony 1.39 1.42 0.04 2.6 North Rhine-Westphalia 1.36 1.39 0.02 1.8 Rhineland-Palatinate 1.37 1.37 0.01 0.5 Saarland 1.28 1.31 0.04 2.8 Saxony 1.48 1.51 0.04 2.5 Saxony-Anhalt 1.40 1.43 0.03 2.0 Schleswig-Holstein 1.38 1.40 0.02 1.5 Thuringia 1.41 1.44 0.03 1.9 1 without West Berlin or respectively without East Berlin. Source: Statistisches Bundesamt, Births statistics, own calculations • Olga Pötzsch96 The age-specifi c changes among women of foreign nationality occurred in par- ticular among those in their mid-twenties (Fig. 3). This led to a slight drop in the average childbearing age by 0.3 years to 29.3 years and in the median by 0.4 years to 29.6 years. The maximum fertility rate (at the age of 29) increased from 99 to 113 children per 1,000 women. These changes, however, have no relevant effects on the fertility of the women as a whole. Figure 3(a) clearly illustrates that they merely lead to a marginal shift in the of the age-specifi c fertility rates. 3.3 Effects of the census on the time series of the TFR The census-based adjustment does not result in a fundamental shift in the levels of the long-term development of the total fertility rate. The curves shown in Figure 4 contain the fi gures for the year 2011 both before and after incorporating the census results. The new, adjusted fi gures follow the curves without any signifi cant leaps. Even when we compare the annual changes in the total fertility rate it is clear that the “census effect” is no greater than the usual annual fl uctuations. Among women of German nationality, the “census effect” on the development of the TFR is negligible. It remains at the low level of 1.34 births per woman rather than 1.33 births per woman before the census (Fig. 5). Among women of foreign nationality, however, the census adjustment led to a visible shift in the level of the TFR for 2011 from 1.58 to 1.82 births per woman. The consequence is that a trend revision is necessary and that the assumption of a rapid convergence of the fertility rate of female foreigners to that of German women must be qualifi ed. At 11 percent, the decrease in the total fertility rate of women with a foreign nationality between 1991 and 2011 is, according to the new fi ndings, only half as strong as originally presumed (22 percent). In 2012 and 2013 the fertility rate among women of for- eign nationality stagnated further to the level of about 1.80 births per woman. The 0 20 40 60 80 100 120 15 20 25 30 35 40 45 Age before the census after the census a) Births per 1,000 women All Women 0 20 40 60 80 100 120 15 20 25 30 35 40 45 Age before the census after the census b) per 1,000 womenBirths Women with foreign citizenship Fig. 3: Age-specifi c fertility rate 2011 – necessary revisions after the census 2011 Source: Statistisches Bundesamt, Births statistics, own calculations and design Fertility in Germany before and after the 2011 Census • 97 Fig. 4: Total fertility rate 1990 – 2013 with „census-effect“ in 2011 * without West Berlin or respectively without East Berlin. Source: Statistisches Bundesamt, Births statistics, own calculations and design Fig. 5: Total fertility rate according to nationality 1991 – 2013 with „census- effect“ in 2011 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 1990 1993 1996 1999 2002 2005 2008 2011 Year Germany Western Germany* Eastern Germany* Births per woman 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 1990 1993 1996 1999 2002 2005 2008 2011 Year German women Women with foreign citizenship Births per woman • Olga Pötzsch98 F ig . 6: C u m u la te d f e rt ili ty r a te s a cc o rd in g t o w o m e n ’s a g e g ro u p s 19 91 -2 01 3 w it h „ ce n su s- e ff e c t“ in 2 01 1 * w it h o u t W e st B e rl in o r re sp e ct iv e ly w it h o u t E as t B e rl in . S o u rc e : S ta ti st is ch e s B u n d e sa m t, B ir th s st at is ti cs , o w n c al cu la ti o n s an d d e si g n 0 3 0 0 6 0 0 9 0 0 1 9 9 1 2 0 0 1 2 0 1 1 Y e ar 1 5 -2 9 3 0 -4 9 G e rm an y 0 3 0 0 6 0 0 9 0 0 1 9 9 1 2 0 0 1 2 0 1 1 Y e ar 1 5 -2 9 3 0 -4 9 W e st e rn G e rm an y * 0 3 0 0 6 0 0 9 0 0 1 9 9 1 2 0 0 1 2 0 1 1 Y e ar 1 5 -2 9 3 0 -4 9 E as te rn G e rm an y * 6 a) A ll w o m e n 0 3 0 0 6 0 0 9 0 0 1 ,2 0 0 1 ,5 0 0 1 9 9 1 2 0 0 1 2 0 1 1 Y e ar 1 5 -2 9 3 0 -4 9 6 b ) W o m e n w it h fo re ig n ci ti ze n sh ip B ir th s p e r 1 ,0 0 0 w o m e n B ir th s p e r 1 ,0 0 0 w o m e n B ir th s p e r 1 ,0 0 0 w o m e n B ir th s p e r 1 ,0 0 0 w o m e n Fertility in Germany before and after the 2011 Census • 99 percentage of births by mothers with a foreign nationality among all children born remained a steady 17 percent to 18 percent since 1996. Long-term age-specifi c trends are decisive for the development of the total fertil- ity rate. They can basically be illustrated using only two age groups. As Figure 6a shows, the birth rate among women under age 30 drops while the fertility of the 30 to 49-year-olds rises. After the census, there is a slight revision for the younger age group for the year 2011 but even the revised TFR for 2011 is lower than in 2010. Thus, the trend in the fertility rate of 15 to 29-year-olds is still declining even after the census. The adjusted fertility rate among the older age group fi ts perfectly into the previous development and does not disturb the curve progression. This is true for Western Germany and – with a minor variation in the curve progressions – also for Eastern Germany. The census-based revision resulted in a leap in the curve progression of the fer- tility rate among younger women of foreign nationality (Fig. 6b). The accumulated fertility rate of the 15 to 29-year-olds in 2011 rose by 20 percent. However, there is apparently still a declining trend. Among the age group of the 30 to 49-year-olds, the revisions are comparatively low at +9 percent. The main reason for the falsely esti- mated decrease in the birth rate among women of foreign nationality is thus due to the fertility rates of the 20 to 29-year-olds. The apparent systematic overestimation of the number of women with foreign citizenships, in particular for this very mobile age group, led to much too low fi gures. 3.4 Effects of the census on the cohort fertility rate In Germany, the cohort total fertility rate per woman is available for the cohorts from 1930 onwards.1 A cohort-by-cohort back-calculation would be necessary to revise the age-specifi c cohort fi gures after the 2011 census. However, the census only caused a slight change in the number of women in the cohorts until the early 1960s (average 0.9 percent) so we can assume a marginal bias to the cohort fi gures and therefore, a time-consuming and highly uncertain back-calculation does not seem worthwhile. Among the cohorts at childbearing age, in particular the 1980s and 1990s cohorts, the number of women dropped a bit more after the census than among the older cohorts. But the changes in the individual age-specifi c fertil- ity rates are also only minor as already shown in Section 3.2. The cumulated cohort fertility rates of the women in their fertile ages therefore only slightly changed after the census. 1 Cohort fertility, in this case, refers to all women. A separate analysis for German women and women of foreign nationality does not seem advisable, mainly because the foreign population constantly changes due to migration and their cohorts are subject to diffi cult-to-control exog- enous infl uences. A cohort-specifi c analysis would be possible with restrictions for German women but since it hardly changed after the census, such an analysis would not provide any additional information with regard to cohort fertility among Germans. • Olga Pötzsch100 Fig. 7: Completed or achieved average number of children per woman according to birth cohorts – results based on the microcensus 2012 according to the new (after the census) and the old (before the census) extrapolation 0,00 0,50 1,00 1,50 2,00 0.00 0.50 1.00 1.50 2.00 199519901985198019751970196519601955195019451940 Birth cohort before the census after the census Completed average number of children Achieved number of children until 2012 Source: Statistisches Bundesamt, Microcensus 2012, own calculations and design Fig. 8: Absolute differences of completed or achieved average number of children per woman according to birth cohorts between the results of the microcensus in 2012 with the new (after the census) and the old extrapolation (before the census) -0,02 0,00 0,02 0,04 -0.02 0.00 0.02 0.04 199519901985198019751970196519601955195019451940 Birth cohort Source: Statistisches Bundesamt, Microcensus 2012, own calculations and design Fertility in Germany before and after the 2011 Census • 101 The microcensus is another possibility for checking the census impacts on the cohort fertility rate (Statistisches Bundesamt 2013). The two questions about chil- dren born are asked in the microcensus every four years. The last survey took place in 2012. The results of the 2012 microcensus published so far were based on the old extrapolation that did not yet incorporate the census revisions. The results from the new census-adjusted extrapolation are now also available and enable us to make a comparison with the results from the old extrapolation. In the following, we will compare these for the average number of children per woman (see Section 2) and for the distribution of the female cohorts according to number of children born (par- ity). The development of the average number of children per woman is shown in Figure 7. For the cohorts of 1963 and older this development is completed, and for the younger cohorts only shown until 2012. The differences between the results according to the old (dotted line) and to the new (solid line) extrapolation are barely perceptible to the eye. The absolute differences are, as Figure 8 reveals, minor. The highest deviations are the almost 30 children per 1,000 women in the 1976 to 1981 cohorts. Among the Western German female cohorts, the differences are similar to those for all women in Germany. In Eastern Germany, the average number of children among the young cohorts after the census rose a little more. In the 1976 to 1983 cohorts the deviations are 40 to 60 children per 1,000 women but these differences are lesser than the doubled absolute standard error and are therefore within the 95 percent confi dence interval (Statistisches Bundesamt 2013: 66). Among the old- er cohorts from 1964, in contrast, the deviations among the Eastern German women are negligible. The basis of the average number of children is the respective structure of the women of one cohort according to parity distribution. The interaction between the parities infl uences the completed fertility rate. The new extrapolation incorporating the census results only caused minor changes to the parity distribution. The upper part of Table 3 shows the percentage parity distribution of the cohorts according to the new extrapolation including the census-adjusted population. The lower section of the table shows the differences to the results of the “old” extrapolation. These tend to show a somewhat lower rate of childlessness and somewhat higher percent- ages of mothers with one or two children. The individual differences are, however, very slight and only marginally alter the previous basic fi ndings concerning the par- ity distribution of the female cohorts. 3.5 Intermediate conclusions According to the 2011 census, the female population of childbearing age was 321,000 lower than indicated by intercensal population updates. Therefore, the fertility indi- cators based on the number of women or on the extrapolated sample results of the 2012 microcensus only slightly changed. The 2011 total fertility rate increased by about 2 percent from 1.36 to 1.39 births per woman. With regard to long-term age- specifi c trends, the census-based revision was within the range of average annual • Olga Pötzsch102 T a b . 3: S tr u c tu re o f fe m a le b ir th c o h o rt s a cc o rd in g t o t h e n u m b e r o f ch ild re n – r e su lt s b a se d o n t h e m ic ro ce n su s 2 01 2 an d n e ce ss a ry r e v is io n s a ft e r th e c e n su s- a d ju st e d e x tr a p o la ti o n B ir th c o h o rt To ta l C h ild le ss M o th e rs a cc o d in g t o b ir th n u m b e rs (a g e in 2 01 2 ) w o m e n 1 c h ild 2 c h ild re n 3 c h ild re n 4 o r m o re c h ild re n N o t a p p lic a b le R e su lt s o f th e c e n su s a d ju st e d e x tr a p o la ti o n , in % 19 9 3 -1 9 9 6 ( 16 -1 9 ) 10 0 9 9 1 / / / 0 19 8 8 -1 9 9 2 ( 2 0 -2 4 ) 10 0 91 7 2 / / 0 19 8 3 -1 9 8 7 ( 2 5 -2 9 ) 10 0 71 18 8 2 1 0 19 7 8 -1 9 8 2 ( 3 0 -3 4 ) 10 0 4 4 2 7 2 2 5 2 0 19 7 3 -1 9 7 7 ( 3 5 -3 9 ) 10 0 2 7 2 5 3 4 10 3 1 19 6 8 -1 9 7 2 ( 4 0 -4 4 ) 10 0 2 2 2 3 3 5 11 4 5 19 6 3 -1 9 6 7 ( 4 5 -4 9 ) 10 0 2 0 2 3 3 5 11 4 6 19 5 8 -1 9 6 2 ( 5 0 -5 4 ) 10 0 18 2 3 3 8 12 5 4 19 5 3 -1 9 5 7 ( 5 5 -5 9 ) 10 0 16 2 4 3 9 12 5 3 19 4 8 -1 9 5 2 ( 6 0 -6 4 ) 10 0 14 2 7 3 9 13 5 2 19 4 3 -1 9 4 7 ( 6 5 -6 9 ) 10 0 12 2 6 4 0 14 6 1 19 3 7- 19 4 2 ( 7 0 -7 5 ) 10 0 11 2 3 3 7 17 10 1 D if fe re n ce s co m p a re d t o t h e r e su lt s o f th e e x tr a p o la ti o n b e fo re t h e c e n su s 2 01 1, p e rc e n ta g e p o in ts 19 9 3 -1 9 9 6 ( 16 -1 9 ) - 0 .0 0 .0 / / / 0 .0 19 8 8 -1 9 9 2 ( 2 0 -2 4 ) - -0 .3 0 .2 0 .1 / / 0 .0 19 8 3 -1 9 8 7 ( 2 5 -2 9 ) - -1 .1 0 .6 0 .4 0 .1 0 .0 0 .0 19 7 8 -1 9 8 2 ( 3 0 -3 4 ) - -1 .4 0 .4 0 .7 0 .2 0 .1 0 .0 19 7 3 -1 9 7 7 ( 3 5 -3 9 ) - -1 .0 0 .1 0 .7 0 .2 0 .0 0 .0 19 6 8 -1 9 7 2 ( 4 0 -4 4 ) - -0 .6 0 .0 0 .4 0 .1 0 .0 0 .0 19 6 3 -1 9 6 7 ( 4 5 -4 9 ) - -0 .1 0 .0 0 .1 0 .0 0 .0 0 .0 19 5 8 -1 9 6 2 ( 5 0 -5 4 ) - 0 .1 0 .0 0 .0 -0 .1 0 .0 0 .0 19 5 3 -1 9 5 7 ( 5 5 -5 9 ) - 0 .1 0 .1 0 .0 -0 .1 -0 .1 0 .0 19 4 8 -1 9 5 2 ( 6 0 -6 4 ) - 0 .1 0 .1 0 .0 -0 .1 -0 .2 0 .0 19 4 3 -1 9 4 7 ( 6 5 -6 9 ) - 0 .1 0 .1 0 .1 -0 .1 -0 .1 0 .0 19 3 7- 19 4 2 ( 7 0 -7 5 ) - 0 .0 0 .0 0 .0 0 .0 0 .0 0 .0 S o u rc e : S ta ti st is ch e s B u n d e sa m t, M ic ro ce n su s 2 0 1 2 , o w n c al cu la ti o n s Fertility in Germany before and after the 2011 Census • 103 fl uctuations. The effects of the 2011 census on the cohort fertility rate were also mi- nor. The census-based adjustment had varying effects on the birth rates according to nationality. While the total fertility rate of women of German nationality remained nearly unchanged at 1.34 births per woman, it rose signifi cantly among women of foreign nationality from 1.58 to 1.82 births per woman. This revision demands that previous fi ndings on the fertility of women with foreign citizenship be re-examined. However, the infl uence of women of foreign nationality on total fertility trends has been rather low so far. 4 Development of the cohort fertility rate in Germany The comparison of the main indicators showed that the general demographic pic- ture of fertility developments largely remained the same after the 2011 census. The fertility studies can therefore be continued without major revisions. But what is the essence of this picture? In the following, we will attempt to approach this question by supplementing the results based on the data from the ongoing fertility statistics with the results of the 2012 microcensus. When put into context, these fi ndings of- fer us insights into the principle features of fertility behaviours and an outlook on future fertility developments. 4.1 No stable increase in cohort fertility rate foreseeable On average, women born in the 1930s – mostly the mothers of the so called “baby- boom generation” – had more than two children (Fig. 9). Their family formation phase coincided with the economic upswing of the 1950s and 1960s. Yet, within the 1930s cohorts there are already indications of a steady decline in the completed fertility rate. It dropped particularly rapidly because of the decreases of higher order births – four or more children – between the cohorts of 1934 and 1944 (Statistisches Bundesamt 2009, 2013). Thereafter, the parity distribution of mothers largely sta- bilised but childlessness continued to increase. This has led, until today, to a less rapid yet steady decline in the cohort fertility rate. The 1964 cohort, with a total of 1.57 births per woman, is the most recent cohort to reach its fi ftieth birthday and whose cumulated fertility rate is statistically considered completed. It is already predictable that the completed fertility rate of the late 1960s cohorts will drop further to approximately 1.49 births per woman. Afterwards, a slight re- covery of the completed cohort fertility is anticipated (Pötzsch 2010a; Goldstein/ Kreyenfeld 2011; Sobotka 2011). One aspect contributing to this recovery is the sta- bilisation of the cohort fertility up to the age of 29 among the 1969 to 1973 cohorts. In addition, though, the fertility rates of these cohorts have continually risen for the ages of over 29. Together, these two effects created a favourable constellation, which did not, however, continue from the 1974 cohort. Between the 1974 and 1984 cohorts, cumulative fertility up to the age of 29 again dropped steadily. It remains to be seen how it will develop at an older age since these cohorts have not yet reached the end of their childbearing phase. Nonetheless, we can derive some indications by • Olga Pötzsch104 analysing the postponement and recuperation of births, the so-called P&R process (Frejka//Calot 2001; Sobotka et al. 2011; Frejka 2012; Pötzsch 2013). For P&R analysis, we must understand cohort fertility as a process in which the earlier fertility biography of women signifi cantly infl uences their future fertility de- cisions. Every female cohort has its own birth timing resulting from its fertility be- haviour. For example, if most of the women of a cohort tend to have children in the younger childbearing phase, the cumulative fertility rises quickly, reaches its peak relatively soon and then remains almost constant (Fig. 10: 1944 cohort). If instead the majority of women of a cohort decide to have children in later ages, the cumu- lative fertility rises more slowly (1962 cohort). Whether the completed fertility re- mains lower than among the cohorts with a relatively “young” birth timing depends on whether all births that have been omitted at a young age are, in fact, recuperated in later ages. In Germany, the shift in timing to a later age has so far resulted in a trend towards a lower completed fertility (Pötzsch 2013). Fig. 9: Completed as well as cumulative fertility until the age of 29 by female birth cohorts 0.0 0.4 0.8 1.2 1.6 2.0 2.4 1934 1944 1954 1964 1974 1984 Births per woman At the age between 30 and 49* At the age between 15 and 29 Completed completed number of children* * Estimated fi gures for the cohorts 1965-1974. Source: Statistisches Bundesamt 2013, own calculations and design Fertility in Germany before and after the 2011 Census • 105 In comparison to the 1944 cohort, the 1962 cohort2 gave birth later. The larg- est difference in cumulative fertility was reached in their late twenties. Although the distance between the two curves lessened, the extent of recuperation was not suffi cient to compensate for the decrease in fertility at the young age and the CTFR fell by 0.2. By the age of 28, the women born in 1973 had fewer births (on average) than the women of the 1962 cohort with a completed fertility of 1.61. Starting at age 29, they have recuperated the “postponed” births. By the age of 40, the distance between the two curves lessened by three-fourths to 70 births per 1,000 women. They will probably not be able to completely close this gap at ages of 41 and above but it will be reduced considerably (cumulated cohort fertility rate at the ages of 41 to 49 is presently approx. 35 births per 1,000 women). The completed fertility rate of the 1973 cohort will be an estimated 1.57. The even younger 1983 and 1988 cohorts represent the current trends in cohort fertility rates. It becomes evident that ever fewer births are taking place during the younger fertile period. In order to be able to assess effects of the most recent birth postponement on the completed fertility rate, we attempt to estimate the future recuperation potential Fig. 10: Cumulative fertility rate of selected female birth cohorts by age 0 400 800 1.200 1.600 2.000 0 400 800 1,200 1,600 2,000 15 20 25 30 35 40 45 Age in 2013 1944 1962 1973 1983 1988 Births per 1,000 women Source: Statistisches Bundesamt, Births statistics, own calculations and design 2 The 1962 cohort will be used as reference for further comparison since it is preliminarily the last completed cohort with a completed number of children of 1.6, for later cohorts the completed number of children will most likely drop below 1.6. • Olga Pötzsch106 using a graph (Fig. 11). For this purpose, we defi ne the 1962 cohort with the CTFR of 1.61 as reference cohort. For every further cohort we calculate the difference be- tween the cumulated cohort fertility rates of this cohort and the 1962 cohort (refer- ence). These differences are negative because the birth rates decrease from cohort to cohort. They usually intensify up to the age of 29 years, which is the so-called trough, and then decrease, as shown previously using the example of the 1962 and 1973 cohorts in Figure 10. The phase until reaching the trough corresponds to the postponement of births; after this trough, the recuperation of births begins. In or- der to compare the postponement of births with the extent of recuperation, the age-specifi c fertility rates per cohort are added in the postponement and in the recuperation phase, respectively. The postponement phase ranges from 15 to 29 years. Multiple age limits are chosen for the recuperation phase in order to be able to make assertions even for the relatively young cohorts: from 30 to 35, from 30 to 37 as well as from 30 to 40 years. For these phases we look at the deviations in the respective cumulative fertility rates of the 1963 to 1984 cohorts in comparison to the 1962 reference cohort (Fig. 11, upper part). As expected, the deviations in the postponement phase are negative up to the age of 29 years while they are positive in the recuperation phase because the co- hort fertility rate of the younger cohorts increases starting at the age of 30. Here, the developments described above are portrayed in a different way: the decrease in the cohort fertility at a young fertile age was especially strong among the 1960s cohorts. In spite of an increasing fertility at the age of 30 and older, it can hardly be compensated for by the end of the fertile period. Instead, the remaining deviation at the age of 40 rose until the 1968 cohort, indicating a drop in the completed fertility (Fig. 11, lower part). For the 1969 and 1973 cohorts, the graph illustrates the favourable constellation of the stabilised cumulated fertility of 15 to 29-year-olds on one hand and the in- creasing recuperations between the ages of 30 and 40 on the other hand (described above). As a result, the completed fertility will slightly rise. Starting with the 1974 cohort, fertility under the age of 30 dropped steadily, indicating a renewed increase in childbearing postponement. How would this decrease affect the cohort fertility? Using Figure 11 (above) we can see the extent of recuperation that would be nec- essary in specifi c ages to compensate for the dropping fertility among the younger fertile age and reach the completed fertility rate of 1.6. This objective would not be reached until the 1980 cohort if the recuperation trends of the past 16 years were to continue in a linear way (cf. dashed lines). This scenario does not seem unrealistic at fi rst glance. But when we look closer, it becomes apparent that this is in fact a major challenge. On one hand, there are signs that recuperation tends to stagnate until the age of 35 in the case of Western Germany. The linear trend extrapolation used here is therefore already an optimistic assumption. On the other hand, such a shift in the cohort fertility rate to a higher childbearing age would mean that fertility at the ages of 36 to 40 years would have to increase by 250 percent between the 1962 and 1984 cohorts and recuperation at ages of 36 and older would have to equal that of the strong fertile period between 30 and 35 years of age. Fertility in Germany before and after the 2011 Census • 107 Residual deviance at the age of 40. Estimated fi gures for the 1974 and later cohorts with a formal extrapolation of the recuperation trend -200 -100 0 -200 -100 0 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 -50 -133 -73 -62 -39 Note: CFR x to (x+n) years is the total age-specifi c fertility rate of a cohort in the respec- tive ages. Absolute changes have been calculated as the difference between the CRF of cohort c and the CFR of cohort 1962 for the respective age range. The dotted line shows the linear trend. Source: Statistisches Bundesamt, Births statistics, own calculations and design -500 -400 -300 -200 -100 0 100 200 300 400 500 -500 -400 -300 -200 -100 0 100 200 300 400 500 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 Births per 1,000 women CFR 30-40 years CFR 30-37 years CFR 30-35 years CFR 15-29 years Recuperation Postponement Fig. 11: Absolute changes of the age-specifi c cumulative cohort fertility rate (CFR) in the postponement and recuperation phase in comparison to the cohort of 1962 • Olga Pötzsch108 To stabilise the completed fertility rate at the relatively low level of 1.5 to 1.6 births per woman would require that the fertility rate of the late 1970s and 1980s cohorts rise continuously from the age of 30 onwards. With regard to parity, this means both more fi rst births at older fertile ages and an even distribution of moth- ers according to number of children in spite of later family formation. Actually, the completed fertility rate of 1.6 births per woman is only reached when about 80 per- cent of the women of one cohort bear an average of two children. In the following, we will determine the extent to which these conditions can be fulfi lled. 4.2 Stability of the fertility rate per woman is fragile Between the 1930s and 1940s cohorts the average fertility rate per mother dropped from 2.3 to 2.0 and then stabilised at that level. During the decrease, the percent- age of mothers with four or more children dropped from 12 percent to 6 percent. Between the late 1940s and early 1970s cohorts the parity distribution of mothers exhibited only minor fl uctuations. By 2012, 31 percent of the 45 to 49-year-old moth- ers had one child, 48 percent had two children, 15 percent had three children and 6 percent four or more children. On average, mothers had two children in the course of their lives (Statistisches Bundesamt 2015c). However, given the long-term trend towards an older age at fi rst birth the sta- bility of the mothers’ parity distribution among the younger cohorts is at risk. This is mainly due to the fact that the share of those mothers who start a family before their thirtieth birthday is gradually getting smaller. But these were the mothers who would potentially have had many children: The microcensus survey showed for the 1960s and 1970s cohorts that on average, the more children a woman has had during her lifetime, the younger she was at the birth of her fi rst child (Statistisches Bundesamt 2015: 30). The mothers of only children had their child on average at the age of 30. The mothers of two children were already three years younger when they began family formation at the age of 27. On average, mothers under the age of 25 formed families with three or more children.3 Therefore, a future stable parity distribution requires that the fertility behaviour at an older fertile age changes by increasing the percentage of mothers who start a family during their thirties and then nevertheless have three or more children. Since the fertile window is limited, the spacing between a mother’s births would also have to be shorter. This is also confi rmed by the results of the microcensus. The older a mother with at least three children was at the time of the fi rst birth, the shorter were the intervals between her births. For women who had their fi rst child at the age of 20, the interval between the fi rst and third child was a little under 10 3 These microcensus results refer to mothers who were between the ages of 35 and 44 at the time of the microcensus survey in 2012 and who lived in the same household with their children. Using this method, the average age at fi rst birth may tend to be a bit too high since in some cases the oldest child may already have moved out of the mother’s household (Source: micro- census 2012, author’s calculations). Fertility in Germany before and after the 2011 Census • 109 years (Fig. 12). Among women who did not have their fi rst child until they reached the age of 33, this interval was 5 years which is only half as long. On average, 7.4 years passed between the fi rst and third birth. Since an increasing number of women do not become mothers until their early thirties and, if they have more than one child, they consequently have shorter birth intervals, then the median spacing should also tend to be shorter. However, the analysis of birth intervals to the previous child of a mother at her second, third or subsequent births so far does not confi rm this assumption. The information about the completed birth order of mothers available since 2009, reveal instead that the median intervals between births remained unchanged from 2009 until 2014 – in spite of the continued postponements of family formation. “In 50 percent of cases a second child is born within 3.3 years after the fi rstborn. The other 50 percent of sec- ond born children have a further spacing to their older siblings. The third children usually follow at an even greater interval after the birth of the second child. Half of all third births take place within 3.9 years, the other half even later.” (Statistisches Bundesamt 2014, quotation translated by CPoS). These results also confi rm the fi ndings of the 2012 microcensus, according to which the interval between the fi rst and third birth is greater than 7 years on average. Consequently, mothers of three or more children are still largely starting their family early. The percentage of women Fig. 12: Average spacing between the fi rst and the third birth, according to the mother’s age at fi rst birth* 0 3 6 9 12 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34-40 9.9 7.6 5.7 5.0 in years * Includes mother between 35 and 44 years with (at least) 3 children who live together in one household Source: Statistisches Bundesamt, Microcensus 2012 (census adjusted projection), own calculations and design • Olga Pötzsch110 who have their fi rst child relatively late within their fertile period is apparently low among mothers of three or more children. Therefore the distribution of birth spac- ing is hardly infl uenced by shorter intervals between their births. The development of the intervals between births is an important early indicator of whether not only the fi rst and second, but also subsequent births are “recuper- ated.” As Figure 13 shows, the percentage of second or third children born within 2 years after the mother’s previous child remained steady at about 16 percent to 18 percent between 2009 and 2013. About 44 percent of second and approx. 36 per- cent of third children were born at an interval of under 3 years. It is also worth not- ing that the percentage of second children born at an interval of 2 to under 3 years after the fi rstborn was almost exactly as large as the percentage of second children born after an interruption of fi ve and more years. Until the third birth the intervals tended to be even greater: 37 percent of third children in 2013 were at least fi ve years younger than the previous sibling. Present fertility behaviour results from a combination of the trend towards late family formation, large birth intervals as well as the negative correlation between the time of the fi rst birth and the total number of children. If this constellation does not change in the near future, it would result in a decrease in the percentage of mothers with three or more children. 4.3 Trend towards higher childlessness continues Within the past twenty years, the development of the cohort fertility rates was greatly infl uenced by the rise in childlessness – especially in Western Germany (Statistisches Bundesamt 2013). According to Sobotka’s calculations, the complet- Fig. 13: Share of births which occurred with a respective spacing to the birth of the preceding child 0 25 50 75 100 2009 2010 2011 2012 2013 18 27 18 11 26 17 27 18 11 26 17 27 19 11 26 18 27 19 12 25 17 27 19 12 25 under 2 years 2 - slightly under 2 years 3 - slightly under 4 years 4 - slightly under 5 years more than 5 years Second birth 0 25 50 75 100 2009 2010 2011 2012 2013 17 20 15 11 37 16 20 15 11 37 16 19 15 11 38 17 19 15 12 37 16 20 16 12 37 Third birth * Only one child was included in the case of giving birth to twins or multiples. Source: Statistisches Bundesamt, Births statistics, own calculations and design Fertility in Germany before and after the 2011 Census • 111 ed fertility of the 1965 cohort would be 1.7 births per woman instead of 1.5 if the rate of childlessness had remained at the level of the 1940 cohort (Sobotka 2011: 273). Childlessness is also increasing in Eastern Germany, where the share of childless women was only 11 percent until the 1960s cohorts, just half as high as in Western Germany. From the age of 41 years onwards the rate of childlessness is hardly changing. Although the number of over 40-year-olds who have a second or third child is in- creasing, it is still very rare for women to have their fi rst child when they are over 40. In 2013, only 3 percent of fi rstborns were children of women aged between 40 and 49, and only 1 percent from the age of 42 onwards. A comparison of the rate of childlessness in the respective female cohorts between the 2008 and 2012 mi- crocensus surveys also only shows marginal changes from the age of 39 years with the exception, however, of women with university degrees, whose family formation phase may last longer (Statistisches Bundesamt 2013: 22 and 36). In the 1971 cohort, who reached the age of 41 in 2012, 22 percent of women remained childless. The rate of childlessness was therefore 10 percentage points higher than in the 1940 to 1946 cohorts (12 percent). Between the 1946 and 1971 cohorts, the increase in childlessness was almost linear, with the exception of some minor fl uctuations (Fig. 14). The extent to which and how long the fi nal level of childlessness will rise in the future remains to be seen. On one hand, there are positive stimuli, for example among Western German women with academic degrees. Their fi nal rate of child- lessness recently stabilised at 29 percent and will probably even drop somewhat in the coming years (Statistisches Bundesamt 2013: 37; Statistisches Bundesamt 2015a). On the other hand, some fi ndings indicate that the peak of the childless- ness has not yet been reached. This includes, for instance, the rising percentage of women with academic degrees who exhibit above average levels of childlessness. In Eastern Germany as well as among Western German women without higher edu- cations, childlessness will probably increase in the coming years. This would lead to a further overall increase of the childlessness level. Socioeconomic, biomedical and cultural factors also indicate a continuation of the trend rather than a reversal. With the introduction of parental leave benefi ts and the expansion of childcare, family policy measures were taken that might work against this trend (Bonin et al. 2013), however, they are nowhere near solving the problem of reconciling work and family life (Bujard/Lück 2015a). For biomedical rea- sons, continuing to postpone the birth of the fi rst child additionally leads to more women not being able to realise their desire for at least one child (Beier et al. 2012; te Velde et al. 2012). At the same time, the latest research on family-related role models shows that lifelong childlessness has become an appealing way of life and is “hardly ever perceived as a shortcoming anymore” (Dorbritz/Diabaté 2015: 131, translated by CPoS). • Olga Pötzsch112 5 Conclusions and discussion The comparison of the main indicators revealed that the fertility development after the 2011 census essentially remained the same. The previous fertility trends are therefore a reliable starting point for assumptions on future developments of fertil- ity. These considerations are not only relevant to research; the future annual fertility rate also has a real effect on the number of births and thus on population develop- ment. Given a total fertility rate of about 1.4 births per woman, a distinct decline in the number of births in the near future is predestined on the basis of the present age structure of the female population (Statistisches Bundesamt 2013: 11-14). From 2017 onwards, the decreasing 1990s and younger cohorts will reach their most fertile age during the mid-twenties to mid-thirties. The declining number of potential mothers will probably only partly be compensated by net immigration.4 Based on the results 4 In 2013, the difference between the number of 30-year-olds and 10-year-olds was 150,000. With a positive net migration and a low mortality rate, the number of girls who are 10 years old will increase over the next 20 years. However, even given the high migration assumption of the 13th Coordinated Population Projection, they would rise by a maximum of 60,000. Thus the number of 30-year-old potential mothers in 2033 would still be smaller by 90,000 in comparison to 2013 (Statistisches Bundesamt. 13th coordinated population projection). Fig. 14: Share of childless women by the respective birth cohort, aged between 41* and 69 years in 2012 (in %) 0 4 8 12 16 20 24 0 4 8 12 16 20 24 1971196619611956195119461941 * From the age of 41 childlessness rates only marginally change. Source: Statistisches Bundesamt, Microcensus 2012 (census adjusted projection), own calculations and design Fertility in Germany before and after the 2011 Census • 113 of the 13th Coordinated Population Projection, a relatively steady numbers of births could only be reached if, in addition to a higher net immigration rate, the annual total fertility rate would rise to at least 1.6 births per woman (Statistisches Bundesamt 2015a: Variant 6 “Relatively young population”). For this, the completed fertility would have to rise to at least 1.6 births per wom- an and then stabilise. The previous projections showed a probable stabilisation of the cohort total fertility rate among the cohorts of the late 1970s at a level slightly below 1.6 (Myrskylä et al. 2013; Statistisches Bundesamt 2015a). In order to be able to assess how probable a further increase in the cohort fertility rate would be, we studied the trends in the fertility behaviour of the female cohorts. The focus was on the effects of increasing postponement of the fi rst birth on the total cohort fer- tility rate. “In the latter part of the 20th century and early in the 21st century, the most prominent demographic mechanism determining fertility trends has been the extent to which childbearing postponement has been counterbalanced by birth re- cuperation.” (Frejka/Sobotka 2008: 30). The experiences in France and Sweden, for example, revealed that a rising fi rst birth age does not necessarily lead to a decline in completed fertility. However, in the case of Germany, these developments were apparently also associated with changing attitudes towards family (Ehmer et al. 2012). While the average age at fi rst birth rose since the cohort of 1947 onwards (Pötzsch/Sommer 2009), the completed fertility rate per woman dropped from 1.80 (1946 cohort) to 1.57 (1964 cohort). The postponement of entry into parenthood was accompanied by a rapid rise in childlessness (from 12 percent to 20 percent) as well as by a decreasing percentage of women with 3 or more children, from 21 percent to 17 percent (Bujard/Lück 2015b, Table 1). To prevent the completed fertility rate from falling in spite of childbearing post- ponement, the decrease in fertility at a younger age would need to be compensated or overcompensated by the recuperation of births. Using postponement and recu- peration analysis, we were able to gain insights into the interaction effects of the age-specifi c fertility of younger cohorts up to 1984. This allows us to explain the expected slight recovery of the completed fertility rate among the 1969 to 1973 co- horts. This recovery is due to a special constellation: the birth rate of these cohorts stabilised at a younger age below 30 years, while it continually rose from the age of 30 onwards. In contrast, it becomes apparent that fertility among under-30-year- olds from the 1974 cohort again drops steadily, which results in another rise in child- bearing delay. This fertility decline at a younger age can only be compensated if the recupera- tion trends of the past 16 years would continue in a linear way. In addition to a signifi cant rise in births between the ages of 30 and 35, which appears consistent and realistic, this also means that fertility at the age of 36 and older would need to increase signifi cantly. From today’s perspective this is a very optimistic assump- tion. There are still “reserves” in fertility at older childbearing ages. For instance, the fertility rate of the over 36-year-olds in Germany as a whole is 26 percent lower than, for example, in Hamburg (Statistisches Bundesamt 2015d) and about 12 per- cent lower than in Sweden, the country with the oldest average childbearing age in Europe (Eurostat 2015: http://ec.europa.eu/eurostat/data/database). Furthermore, • Olga Pötzsch114 Goldstein et al. point out improved family policy frameworks, greater acceptance of female employment and the rising family focus of younger birth cohorts as factors that could result in a reversal of the trend (Goldstein et al. 2012). Although these fac- tors apparently have a positive effect on fertility development at older childbearing ages, at the same time other fi ndings limit the extent of the compensation effects starting at the age of 36. In particular, the transition to the fi rst child during the last third of the childbear- ing period is still rare. According to Beier et al. the ideal time for a fi rst pregnancy is – from the biological point of view – still between 18 and 30 years of age (Beier et al. 2012: 303). Afterwards the probability of fi rst pregnancy gradually decreases. It re- mains to be seen how further medical advances will be refl ected in the age-specifi c fi gures concerning fi rst births in the future. With regard to family size, Neyer et al. reach the following conclusions: “Regarding third births, there is a relatively uni- form picture: in all the Nordic countries the tendency of mothers with two children to have a third decreases with increasing age, [...]. [...] the relative risk of a second birth only drops signifi cantly for mothers from their mid-thirties. For these moth- ers we can assume that at least some of them did not have another child owing to decreasing fertility although they still wanted a child” (Neyer et al. 2006; quotation translated by CPoS). In addition, the results presented show that mothers with sev- eral children are a special group. It is characteristic for them to have their fi rst child before their mid-twenties and have relatively large intervals between individual births. In order for the percentage of mothers with three or more children to remain stable in the future, these patterns in fertility behaviour would have to change: more women who fi rst started a family at the age of 30 and older would have to have third and subsequent children and/or the age at fi rst birth would need to remain stable. Empirically these changes cannot yet be ascertained. Increasing childlessness is another obstacle. To achieve and maintain the com- pleted fertility rate per woman of at least 1.6 – with the average number of children per mother of 2.0 – the childlessness rate should not exceed 20 percent. In 2012, the rate of childlessness, however, was already 22 percent and the peak has prob- ably not yet been reached. Despite some positive stimuli such as a consolidation of childlessness among highly educated women, there are no indications for a reversal of the increasing trend. All in all, we can say that even the stabilisation of the completed fertility rate on a relatively low level between 1.5 and 1.6 births per woman requires that the increase in the fertility of the 1970s and 1980s cohorts over the age of 30 must be suffi ciently strong to compensate for the increasing fertility decline under the age of 30. To achieve this, the fertility behaviour of women would both have to change in the transition to the fi rst child and with respect to family size. The positive momentum in the fertility rate in recent years is insofar consistent and shows that the recupera- tion process has intensifi ed. An increase and subsequent stabilisation of the total cohort fertility rate at the level of at least 1.6 births per woman, however, would also require a stable fertility below the age of 30 and a decline in the childlessness rate. Fertility in Germany before and after the 2011 Census • 115 Acknowledgements The author would like to thank Felix zur Nieden, Bettina Sommer, Martin Bujard, Norbert F. Schneider and two anonymous reviewers for their helpful comments regarding earlier versions of this article. References Beier, Henning M. et al. 2012: Medizinische und biologische Aspekte der Fertilität. In: Stock Günter et al. (Eds.): Zukunft mit Kindern: Fertilität und gesellschaftliche Ent- wicklung in Deutschland, Österreich und der Schweiz. Frankfurt am Main: Campus: 294-310. Bongaarts, John; Feeney, Griffi th 1998: On the quantum and tempo of fertility. In: Popu- lation and Development Review 24,2: 271-291. Bongaarts, John; Sobotka, Tomas 2012: Demographic explanation for the recent rise in European fertility. In: Populaltion and Development Review 38,1: 83-120 [doi: 10.1111/j.1728-4457.2012.00473.x]. Bonin, Holger et al. 2013: Lehren für die Familienpolitik: Zentrale Resultate der Ge- samtevaluation familienbezogener Leistungen [URL: ftp://ftp.zew.de/pub/zew-docs/ gutachten/Fampolit_Leistungen_HBO.pdf, 14.01.2016]. Bujard, Martin; Lück, Detlev 2015a: Kinderlosigkeit und Kinderreichtum: Zwei Phäno- mene und ihre unterschiedlichen theoretischen Erklärungen. In: BiB Working Paper 2015,1. Wiesbaden: Bundesinstitut für Bevölkerungsforschung [URN: urn:nbn:de:bib- wp-2015-011]. Bujard, Martin; Lück, Detlev 2015b: Kinderlosigkeit und Kinderreichtum: Gründe und Daten für eine paritätsspezifi sche Fertilitätsforschung. In: Zeitschrift für Familienfor- schung 3,27: 255-269 Dorbritz, Jürgen; Ruckdeschel, Kerstin 2007: Kinderlosigkeit in Deutschland – Ein euro- päisches Sonderweg? Daten, Trends und Gründe. In: Kreyenfeld, Michaela; Konietzka, Dirk (Eds.): Ein Leben ohne Kinder. Wiesbaden: VS Verlag für Sozialwissenschaften: 45-82. Dorbritz, Jürgen; Diabaté, Sabine 2015: Leitbild und Kinderlosigkeit: Kulturelle Vorstel- lungen zum Leben ohne Kinder. In: Schneider, Norbert F.; Diabaté, Sabine; Ruckde- schel, Kerstin (Eds.): Familienleitbilder in Deutschland. Verlag Babara Budrich: 113- 132. Ehmer, Josef; Ehrhardt, Jens; Kohli, Martin 2012: Fertilität in historischer Perspektive. In: Stock, Günter et al. (Eds.): Zukunft mit Kindern. Frankfurt/New York: Campus Ver- lag: 41-52. Eurostat 2015: [URL: http://ec.europa.eu/eurostat/data/database, 14.01.2016]. Frejka, Tomas; Calot, Gérard 2001: Cohort Reproductive Patterns in Low-Fertli- ty Countries. In: Population and Development Review 27,1 [doi: 10.1111/j.1728- 4457.2001.00103.x]. Frejka, Tomas; Sobotka, Tomáš 2008: Overview Chapter: Fertility in Europe: Diverse, delayed and below replacement. In: Demographic Research 19,3 [doi: 10.4054/Dem- Res.2008.19.3]. Frejka, Tomas 2012: Die Auswirkungen des aktuellen Aufschubs und Nachholens von Geburten auf die Ausprägung der Periodenfertilitätstrends. In: Comparative Popula- tion Studies 36,4 [doi: 10.4232/10.CpoS-25011-20de]. • Olga Pötzsch116 Goldstein, Joshua R.; Sobotka, Tomáš; Jasilioniene, Aiva 2009: The end of ‘lowest-low’ fertility? In: Population and Development Review 35,4: 663-699 [doi: 10.1111/j.1728- 4457.2009.00304.x]. Goldstein, Joshua R.; Kreyenfeld, Michaela 2011: Has East Germany overtaken West Germany? Recent trends in order-specifi c fertility. In: Population and Development Review 37,3: 453-472 [doi: 10.1111/j.1728-4457.2011.00430.x]. Goldstein, Joshua R. et al. 2011: Fertility Forecasting in the German-speaking World: Recent Experience and Opportunities for Improvement. In: Comparative Population Studies 36,2-3: 661-692 [doi: 10.4232/10.CPoS-2011-09en]. Goldstein, Joshua R.; Kreyenfeld, Michaela; Rößger, Felix 2012: Gibt es eine Trendum- kehr in der Kinderzahl nach Geburtsjahrgängen in Deutschland? Working Paper: Berli- ner Demografi e Forum, Ausgabe 4, Januar 2012. Kreyenfeld, Michaela et al. 2010: Order-Specifi c Fertility Rates for Germany: Estimates from Perinatal Statistics for the Period 2001-2008. In: Comparative Population Studies 35,2: 207-224 [doi: 10.4232/10.CPoS-2010-06en]. Kreyenfeld, Michaela et al. 2012: Fertilitätsdaten für Deutschland, Österreich und die Schweiz: Wo liegen die Möglichkeiten? Was sind die Begrenzungen? In: Comparative Population Studies 36,2-3: 381-416 [doi: 10.4232/10.CPoS-2011-06de]. Konietzka, Dirk; Kreyenfeld, Michaela 2013: Kinderlosigkeit in Deutschland. Theoreti- sche Probleme und empirische Ergebnisse. In: Konietzka, Dirk; Kreyenfeld, Michaela (Eds.): Ein Leben ohne Kinder: Ausmaß, Strukturen und Ursachen der Kinderlosigkeit. 2. Aufl age. Wiesbaden: VS Verlag für Sozialwissenschaften. Kohler, Hans-Peter; Ortega, José A. 2002: Tempo-adjusted period parity progression measures, fertility postponement and completed cohort fertility. In: Demographic Re- search 6,6: 91-144 [doi: 10.4054/DemRes.2002.6.6]. Luy, Marc 2010: Tempo-Effekte und ihre Bedeutung für die demografi sche Analyse. In: Comparative Population Studies 35,3: 447-482 [doi: 10.4232/10.CPoS-2010-11de]. Luy, Marc; Pötzsch, Olga 2010: Estimates of the Tempo-adjusted Total Fertility Rate in Western and Eastern Germany, 1955-2008. In: Comparative Population Studies 35,3: 569-604 [doi: 10.4232/10.CPoS-2010-14en]. Luy, Marc 2016: Demografi sche Kennziffern und Methoden. In: Niephaus, Yasmin; Kreyenfeld, Michaela; Sackmann, Reinhold (Eds.): Handbuch Bevölkerungssoziologie. Springer VS Verlag, Wiesbaden 2016. Myrskylä, Mikko; Goldstein, Joshua R.; Cheng, Alice Yen-Hsin 2013: New Cohort Fertil- ity Forecast for the Developed World: Rises, Falls, and Reversals. In: Population and Development Review 39,1: 31-56 [doi: 10.1111/j.1728-4457.2013.00572.x]. Neyer, Gerda R. et al. 2006: Fertilität, Familiengründung, Familienerweiterung in den nordischen Ländern. In: MPIDR Working Paper WP 2006-022 [http://www.demogr. mpg.de/papers/working/wp-2006-022.pdf, 15.01.2016]. Pötzsch, Olga; Sommer, Bettina 2009: Generatives Verhalten der Frauenkohorten im langfristigen Vergleich. In: Wirtschaft und Statistik 5/2009. Statistisches Bundesamt: Wiesbaden. Pötzsch, Olga 2010a: Annahmen zur Geburtenentwicklung in der 12. koordinierten Be- völkerungsvorausberechnung. In: Wirtschaft und Statistik 1/2010. Statistisches Bun- desamt. Pötzsch, Olga 2010b: Kohortenfertilität: Ein Vergleich der Ergebnisse der amtlichen Ge- burtenstatistik und der Mikrozensuserhebung 2008. In: Comparative Population Stu- dies 35,1: 165-184 [doi: 10.4232/10.CPoS-2010-05de]. Fertility in Germany before and after the 2011 Census • 117 Pötzsch, Olga 2013: Wie wirkt sich der Geburtenaufschub auf die Kohortenfertilität in West und Ost aus? In: Wirtschaft und Statistik 2/2013. Statistisches Bundesamt. Prioux, France; Mazuy, Magali; Barbieri, Magali 2010: Recent demographic develop- ments in France: fewer adults live with a partner. In: Population-E 65,3: 363-414 [doi: 10.1353/pop.2010.0016]. Ryder, Norman 1980: Components of temporal variations in American fertility. In: Hiorns, Robert W. (Ed.): Demographic Patterns in Developed Societies. London: Tay- lor and Francis: 15-54. Sobotka, Tomáš 2003: Tempo-Quantum and Period-Cohort Interplay in Fertility Changes in Europe: Evidence from the Czech Republic, Italy, the Netherlands and Sweden. In: Demographic Research 8,6: 151-214 [doi: 10.4054/DemRes.2003.8.6]. Sobotka, Tomáš 2011: Fertility in Austria, Germany and Switzerland: Is there a Common Pattern? In: Comparative Population Studies 36,2-3 [doi: 10.4232/10.CPoS-2011-12en]. Sobotka, Tomáš et al. 2011: Postponement and Recuperation in Cohort Fertility: Austria, Germany and Switzerland in a European Context. In: Comparative Population Studies 36,2-3 [doi: 10.4232/10.CPoS-2011-10en]. Statistisches Bundesamt 2009: Imputation von Werten bei fehlenden Angaben zur Mutterschaft und zur Zahl der geborenen Kinder im Mikrozensus 2008 [https://www. destatis.de/DE/ZahlenFakten/GesellschaftStaat/Bevoelkerung/Geburten/Methoden/ Downloads/Imputationsverfahren.html, 14.01.2016]. Statistisches Bundesamt (Ed.) 2012: Geburten in Deutschland, Ausgabe 2012. Wiesba- den. Statistisches Bundesamt (Ed.) 2013: Geburtentrends und Familiensituation in Deutsch- land 2012. Wiesbaden. Statistisches Bundesamt 2014: Pressemitteilung 434/14 vom 8.12.2014 „682 000 Kinder kamen im Jahr 2013 zur Welt“. Statistisches Bundesamt (Ed.) 2015a: Bevölkerung Deutschlands bis 2060. Wiesbaden. Statistisches Bundesamt 2015b: Geburtenraten und Tempoeffekt: [https://www.desta- tis.de/DE/ZahlenFakten/GesellschaftStaat/Bevoelkerung/Geburten/Geburtenraten- Tempoeffekt.html, 14.01.2016]. Statistisches Bundesamt 2015c: Daten zu Geburten, Familien und Kinderlosigkeit – Er- gebnisse des Mikrozensus 2012 – Tabellen mit neuer Hochrechnung anhand der Be- völkerungsfortschreibung auf Basis des Zensus 2011 – Ausgabe 2015. Statistisches Bundesamt 2015d: Tabellen „Altersspezifi sche Geburtenziffern 2014_Ge- burtsjahr_Länder“. Stock Günter et al. (Ed.) 2012: Zukunft mit Kindern: Fertilität und gesellschaftliche Ent- wicklung in Deutschland, Österreich und der Schweiz. Frankfurt 2012. te Velde, Egbert et al. 2012: The effect of postponement of fi rst motherhood on perma- nent involuntary childlessness and total fertility rate in six European countries since the 1970s. In: Human Reproduction 27, 4: 1179-1183 [doi: 10.1093/humrep/der455]. Wood, Jonas; Neels, Karel; Kil, Tine 2014: The educational gradient of childlessness and cohort parity progression in 14 low fertility countries. In: Demographic research 31,46: 1365-1416 [doi: 10.4054/DemRes.2014.31.46]. • Olga Pötzsch118 Date of submission: 15.10.2015 Date of acceptance: 29.01.2016 Olga Pötzsch (). Federal Statistical Offi ce. Wiesbaden, Germany. E-mail: olga.poetzsch@destatis.de URL: https://www.destatis.de/EN/Homepage.html Published by Prof. Dr. Norbert F. Schneider Federal Institute for Population Research D-65180 Wiesbaden / Germany 2016 Managing Editor Frank Swiaczny Assistant Managing Editor Katrin Schiefer Copy Editor (Selected Articles in German) Dr. Evelyn Grünheid Layout Beatriz Feiler-Fuchs E-mail: cpos@bib.bund.de Scientifi c Advisory Board Paul Gans (Mannheim) Karsten Hank (Köln) Johannes Huinink (Bremen) Michaela Kreyenfeld (Rostock) Marc Luy (Wien) Notburga Ott (Bochum) Peter Preisendörfer (Mainz) Nikola Sander (Groningen) Zsolt Spéder (Budapest) Comparative Population Studies www.comparativepopulationstudies.de ISSN: 1869-8980 (Print) – 1869-8999 (Internet) Board of Reviewers Martin Abraham (Erlangen) Laura Bernardi (Lausanne) Hansjörg Bucher (Bonn) Claudia Diehl (Konstanz) Andreas Diekmann (Zürich) Gabriele Doblhammer-Reiter (Rostock) Jürgen Dorbritz (Wiesbaden) Anette Eva Fasang (Berlin) E.-Jürgen Flöthmann (Bielefeld) Alexia Fürnkranz-Prskawetz (Wien) Beat Fux (Salzburg) Joshua Goldstein (Berkeley) Sonja Haug (Regensburg) Hill Kulu (Liverpool) Aart C. Liefbroer (Den Haag) Kurt Lüscher (Konstanz) Emma Lundholm (Umeå) Nadja Milewski (Rostock) Dimiter Philipov (Wien) Roland Rau (Rostock) Tomáš Sobotka (Wien) Jeroen Spijker (Barcelona) Olivier Thévenon (Paris) Helga de Valk (Brussel) Heike Trappe (Rostock) Michael Wagner (Köln)