Keynote Sobotka

FERTILITY TRENDS IN EUROPE:Is below-replacement fertility an inevitable outcome of the Second Demographic Transition?: 

FERTILITY TRENDS IN EUROPE: Is below-replacement fertility an inevitable outcome of the Second Demographic Transition? VUB Colloquium on “Demographic challenges for the 21st century,” Brussels, February 15-16, 2007 Tomáš Sobotka VIENNA INSTITUTE OF DEMOGRAPHY

INTRODUCTION (I): 

INTRODUCTION (I) Three ‘stylised facts’ on European fertility Fertility rates in Europe are very low and further declining Current low fertility will necessarily lead to a rapid population ageing and to a decline in population size These trends are unsustainable in the long run and constitute serious threats to economy, labour market, welfare system, and thus also to the foundations of European societies

INTRODUCTION (II): 

INTRODUCTION (II) Renewed fears of “fertility implosion,” “baby deficit” and population decline 1920s and 1930s: spreading concerns about low birth rates and their consequences 1970s-?: renewed fears of declining fertility, rapid population ageing and declining population size J. C. Chesnais (2001): population implosion in the 21st century may be particularly pronounced in Europe Pope Benedict XVI (Christmas 2006): Europe “…seems no longer wants to have children” (…)and “seems to be wishing to take its leave of history” Increasing political concerns & pronatalism The “Green Paper” (European Commission, 2005): low birth rate is a “challenge for the public authorities”; “return to demographic growth” is one out of “three essential priorities”

TOPICS & HYPOTHESES (I): 

TOPICS & HYPOTHESES (I) Aims of this presentation: Outline the recent shift towards low & late fertility in Europe Discuss selected determinants of fertility Future prospects: Implications of current fertility & migration for long-term population trends in the EU

TOPICS & HYPOTHESES (II): 

TOPICS & HYPOTHESES (II) Main Hypotheses Extremely low period fertility rates are linked to fertility postponement and are likely to be temporary Pronounced regional differences are likely to prevail Second Demographic Transition is not necessarily linked to below-replacement fertility Immigration can substitute most or all of the births ‘missing’ due to below-replacement fertility Very low fertility and the prospects of population decline do not constitute an all-European problem and, therefore The fears of European population implosion seem exaggerated

ANALYSIS (1): ‘QUANTUM’: The shift to low and lowest-low fertility: 

ANALYSIS (1): ‘QUANTUM’: The shift to low and lowest-low fertility Period Total Fertility Rate in Western, Northern, and Southern Europe (1950-2005): rising regional differences McDonald (2006): Dividing line: TFR below 1.5 (need for policy interventions?)

Slide7: 

Period Total Fertility Rate in Central-Eastern Europe (1950-2005): decline to the very low levels Data sources: Council of Europe (2006), EUROSTAT (2006), and national stat. offices NOTE: Data are weighted by population size of given countries & regions

The spread of “lowest-low fertility” in Europe (Kohler, Billari, and Ortega 2002): 

The spread of “lowest-low fertility” in Europe (Kohler, Billari, and Ortega 2002) EUROPEAN FERTILITY CRISIS? Data sources: Council of Europe (2006), Eurostat (2006)

The shift to very low fertility is often linked with the spread of one-child family model(Important exception: German-speaking countries): 

The shift to very low fertility is often linked with the spread of one-child family model (Important exception: German-speaking countries) Data sources: Council of Europe (2006) Share of first births on the total births (%) To understand low fertility trends and differences it is crucial to look at parity-specific fertility

Parity-specific differences manifested in fertility ‚recuperation‘ at higher reproductive ages: 

Parity-specific differences manifested in fertility ‚recuperation‘ at higher reproductive ages Lifetime probability of having a(nother) child among women aged 30

ANALYSIS (2): ‘Tempo’: The shift to an ever later family formation: 

Data sources: Council of Europe (2006), EUROSTAT (2006), and national sources ANALYSIS (2): ‘Tempo’: The shift to an ever later family formation

Proportion of first birth rates realised at ages 40+ (%): 

Pushing the age limits: Increase in first-time motherhood at very late reproductive ages Proportion of first birth rates realised at ages 40+ (%) Data sources: computations based on Eurostat (2006), CBS Statline, Statistics Austria and NCHS (2006)

…and increasing ‘polarization’ in first birth timing: 

…and increasing ‘polarization’ in first birth timing England and Wales: Age when a given proportion of first birth rates realised Source: Own computations based on data estimated by Smallwood (2002)

ANALYSIS (3). ‚Tempo‘ vs. ‚Quantum‘: The link between fertility postponement and lowest-low fertility: 

ANALYSIS (3). ‚Tempo‘ vs. ‚Quantum‘: The link between fertility postponement and lowest-low fertility CONSENSUS on tempo distortions: The shift to later childbearing distorts commonly used indicators of period fertility, especially the TFR which is ‘deflated’ Possible misinterpretations of trends and cross-country differences in the TFR: apparent cross-country differences may be caused by different pace of fertility postponement This may lead to erroneous projections of the future period & cohort fertility Tempo distortions may affect the TFR for many decades (thus, also the gap between the period TFR and the cohort CTFR may persist for 30-40 years)

ANALYSIS (3). ‚Tempo‘ vs. ‚Quantum‘: The link between fertility postponement and lowest-low fertility: 

ANALYSIS (3). ‚Tempo‘ vs. ‚Quantum‘: The link between fertility postponement and lowest-low fertility LACK OF CONSENSUS: How to measure the ‘underlying’ period fertility levels There is no unambiguous alternative to the TFR Two major approaches: Shift to the more sophisticated, parity-specific measures (life table indicators based on age & parity or parity & duration since the previous birth) Adjustment approach: Attempts to correct the period measures for tempo distortions (Bongaarts-Feeney, Kohler-Ortega, Kohler-Philipov); not universally accepted How to interpret these indicators? Do they come closer to the corresponding cohort measures? Do they indicate the likely future extent of the ‘recuperation’ of postponed fertility and of the future TFR?

More sophisticated measures (1): Period Average Parity (PAP), based on parity progression ratios: 

More sophisticated measures (1): Period Average Parity (PAP), based on parity progression ratios Period TFR and PAP in Austria, 1984-2006 Mean values 1984-2006: TFR: 1.429 PAP: 1.630 CTFR (cohort 1966): 1.68

More sophisticated measures (2): adjusted TFR in Europe, 2002-2003 (Bongaarts-Feeney adjustment): 

More sophisticated measures (2): adjusted TFR in Europe, 2002-2003 (Bongaarts-Feeney adjustment)

Tempo-adjusted measures: Main conclusions: 

Tempo-adjusted measures: Main conclusions Different size of tempo effects across Europe: some cross-country differences in the TFR linked to tempo effects Fertility ‚quantum‘ recently stable in most regions in Europe – surprising stability in the adj. TFR There is no reason to consider the TFR of 1.5 as a threshold, below which countries fall into a low-fertility trap (McDonald) No country with the lowest-low fertility (TFR<1.3) also has the lowest-low adjustedTFR Lowest-low fertility appears to be a temporary phenomenon, caused by tempo effects

ANALYSIS (4): Insights based on completed cohort fertility: 

ANALYSIS (4): Insights based on completed cohort fertility Cross-country variation; long history of sub-replacement fertility Some countries of Northern & Western Europe likely to retain completed CTFR close to replacement threshold

Huge swings in the period TFR often have only a small impact on the period TFR: 

Huge swings in the period TFR often have only a small impact on the period TFR Period and cohort TFR in the Czech Republic; 1900-2005

Delayed childbearing and completed fertility: 

Delayed childbearing and completed fertility Individual-level data: strong association between the timing of first birth and completed fertility Aggregate-level data: conflicting evidence on the impact of delayed childbearing on completed fertility rates Mean age at childbearing by birth order and parity progression rates to first, second, and third birth among Swedish women born in 1940-65

Childless societies?: 

Childless societies? Childlessness is on the rise in almost all advanced societies In many countries record-high levels of childlessness reached among women born around 1900 Germany stands out for the highest childlessness in Europe

Cohort progression rates to second birth: 

Cohort progression rates to second birth The main differentiating factor for the overall completed fertility rates

ANALYSIS (5): Changing family context of childbearing: 

ANALYSIS (5): Changing family context of childbearing Rapid rise in non-marital childbearing in all parts of Europe

...but vast cross-country differences: 

...but vast cross-country differences No East-West divide in non-marital childbearing; history and diverse cultural and institutional factors matter most

Non-marital childbearing in east and west Germany: no convergence after the Unification: 

Non-marital childbearing in east and west Germany: no convergence after the Unification

In most countries, non-marital childbearing first spreads among the lower-educated people: 

In most countries, non-marital childbearing first spreads among the lower-educated people Czech Republic: non-marital births by education of mother, 1990-2005 (%)

...and takes place within increasingly diverse living arrangements, especially in cohabiting unions : 

Distribution of non-marital births in England and Wales by the recognition of father (%) ...and takes place within increasingly diverse living arrangements, especially in cohabiting unions Source: ONS 2006

Discussion (1): How are declining marriage rates, rising cohabitation and partnership instability linked to fertility?: 

Discussion (1): How are declining marriage rates, rising cohabitation and partnership instability linked to fertility? Macro-level associations: reversal of the previous patterns in the 1990s (Billari and Kohler 2004) Divorce rates and non-marital ratio became positively associated with the TFR, first marriage rates show slight negative association Countries with the highest fertility rates also have high divorce and partnership dissolution rates, low marriage rates, high prevalence of cohabitation and high frequency of extramarital childbearing

Slide30: 

Micro-level analysis: less conclusive, many different factors play a role Eckhard (2006): Shorter, less stable partnerships and increasing ‘partnerlessness‘ may partly explain declining cohort fertility in Germany But: Kravdal (1997): many Norwegians deliberately enter prenthood even if their partnership situation is uncertain Moreover: Women and men entering their second union have a strong motivation to have a shared child with their new partner -> positive effects of stepfamily fertility, especially on third births (Vikat, Thomson, Hoem 1999 and other studies) Billari (2005: 80): “If the rule is one child per couple, the only way to reach replacement is to have individuals experience two couple relationships“

HOW MANY CHILDREN WITH HOW MANY PARTNERS?Multipartner fertility among Danish men born in 1960: 

HOW MANY CHILDREN WITH HOW MANY PARTNERS? Multipartner fertility among Danish men born in 1960 Notes: Data recorded for the period through 2003, some men have not completed their ‘childbearing‘ No records about fathers are known for ca. 2-3 percent of children Sources: Own comptations from the Danish registry data

Discussion (2): Is the Second Demographic Transition (SDT) ‘good‘ for fertility?: 

Discussion (2): Is the Second Demographic Transition (SDT) ‘good‘ for fertility? Three main links between the SDT concept and fertility trends (van de Kaa 1987, 2001, Lesthaeghe 1995): SDT linked to a massive postponement of parenthood (facilitating role of modern contraception emphasised) SDT leads to a marked rise in non-marital childbearing SDT leads to a decline of period and eventually also cohort fertility rates below replacement level Period fertility may eventually recover a bit (‚recuperation‘), but not enough to bounce back to the replacement level (van de Kaa 1997 & 2001: an ideal scheme of 15 stages of the SDT) Lesthaeghe and Neels (2002): long-term subreplacement fertility one of the more recent characteristics of the SDT Some studies simplistically equal SDT with very low fertility

Slide33: 

Macro and micro-level associations: Coleman (2004: 18): surprisingly, “SDT has nothing to do with low fertility on a cross-national basis today“ Van de Kaa (2001, individual-level data): in 1992, postmaterialism and subjective well-being did not show any association with fertility intentions Lesthaeghe and Neiderd (2006): regional-level TFR in the US negativelly associated with SDT factor on a county level (3141 units) small, but positive correlation on a state level (50) HOWEVER: The positive link between divorce rate, non-marital births, and the TFR suggests that some aspects of the SDT in Europe may be positively linked to fertility

Analysis of the SDT-fertility link: 

Analysis of the SDT-fertility link Indexes capturing different dimensions of the SDT: Index SDT1 (demographic factors in 2004): mean age at first birth, first marriage, teenage fertility, non-marital ratio, total divorce rates, total first marriage rate, prevalence of cohabiting unions; 34 countries Index SDT2 (ideational & value factors around 2000); based on the European Values Study in 29 countries 8 questions: Family values, non-conformism, permissiveness, secularism Both indexes and their components can range from 0 (= no STD) to 10 (=max. score on STD factors) A combination of the mean values of the SDT1 and SDT2: SDT-C index

Results (1): SDT-TFR association: 

Results (1): SDT-TFR association This association remains identical if only the values- dimension index SDT2 is used

Results (2): Does the association hold if the TFR is adjusted for tempo effects?: 

Results (2): Does the association hold if the TFR is adjusted for tempo effects? Adjusted TFR computed for the period 2001-2003

Results (3): Does the association hold also for the intended family size?: 

Results (3): Does the association hold also for the intended family size? Mean intended family size computed from Eurobarometer 2006 data (Testa 2006)

Discussion (3): The multifaceted impact of immigration on childbearing and population trends: 

Discussion (3): The multifaceted impact of immigration on childbearing and population trends Two dimensions analysed: Impact of immigration on childbearing trends and patterns ‘Replacement migration‘: Impact of immigration on population size Focus: regions with large immigration streams in the last 2-3 decades (western, southern, northern Europe)

1) Impact of immigration on childbearing: 

1) Impact of immigration on childbearing Lack of comparative cross-country data; few studies (Coleman 1994; Schoorl 1995; Haug, Compton and Courbage 2002) Definitions: (all) immigrant women (men), foreigners only, first vs. second and third generation, legal vs. illegal The use of data for foreigners problematic, especially in countries with high rates of naturalisations TFR may be problematic as well (Toulemon 2004, Andersson 2004) Analysis: Childbearing of legally resident immigrant (foreign) women; 1st generation: Proportion of births to immigrant women Fertility differences between immigrant vs. ‚‘native‘ women ‘Net impact‘ of immigrant women on the TFR

Slide40: 

Births and fertility among immigrant women, 9 countries I: immigrant women, F: foreign women only

2) Replacement migration: can immigration substitute ‚missing‘ births?: 

2) Replacement migration: can immigration substitute ‚missing‘ births? UN report (2000): Replacement migration: Is it a solution to declining and ageing poplation? Often misinterpreted; different concepts of ‚replacement migration‘ CONSENSUS: migration cannot stop population ageing; only modest impact on slowing-down the process BUT: Can immigration substitute most of the births ‚deficit‘ even in the countries with very low fertility?

Measuring replacement migration: 

Measuring replacement migration Importance of immigration: need to rethink the traditional concepts of replacement fertility (Calot & Sardon 2001, Smallwood & Chamberlain 2005) Measuring longer impact of immigration by combining period & cohort measures: GRE (=Gross REplacement rate) tracing the change in the relative cohort size of females from their birth through their prime reproductive period At birth: GRE(t) = GRR(t) (Gross Reproduction Rate) = = TFR * (female live births / all live births) At age 30 (final GREF) GREF(t) = PF(t+30)/PF(t) * GRR(t) Advantage, interpretation

Switzerland: a textbook example of very low fertility combined wih replacement migration: 

Switzerland: a textbook example of very low fertility combined wih replacement migration The GRE of selected cohorts by the time elapsed since birth (age)

GRE: A crucial contribution of ages 20-30: 

GRE: A crucial contribution of ages 20-30 Switzerland: GRE at the time of birth (GRR) and at the duration (age) 20, 25, and 30

Cross-country comparison: different combinations of fertility and migration levels: 

Cross-country comparison: different combinations of fertility and migration levels 5 countries: GRE for 1978 (1984 for Spain) By age 27, GRE up to 0.90-0.98 Younger cohorts: further decline in the GRE at age 0 and faster subsequent increase

Conclusions:: 

Conclusions: Very low fertility is not an all-European problem New heterogeneity and new cleavages in Europe (also van de Kaa 1999) ‚High fertility belt‘: Nordic countries + North-western Europe (Benelux, FR, UK, IRE) TFR 1.7-2.0, adj.TFR and cohort CTFR around 1.9, replacement or above-replacement migration Low fertility-high migration: Southern Europe, Switzerland, Austria, Western Germany Very low fertility (adj. TFR 1.5) combined with replacement migration Low fertility-emigration: dangerous mix; negative population momentum, rapid population decline (East Germany, Baltic countries, Bulgaria, Ukraine, Belarus, Moldova, Russia) In these regions, very low fertility may seriously undermine social dynamic and prosperity

Conclusions:: 

Conclusions: (2) SDT is positively linked to fertility Advanced SDT does not necessarily imply sub-replacement fertility (France, Sweden, Norway) Especially the close association with the values dimension (SDT2) surprising Cautious interpretation necessary (ecological fallacy) Can STD in its later stage become positively liked to fertility? Advanced SDT countries may be transformed into more gender-equal, child-friendly, women-friendly and family-friendly societies, which give individuals more freedom to better realise their childbearing plans alongside their competing aspirations (Strong enabling role of welfare policies?) This interpretation may change our perspectives on the likely future fertility trends in Europe Need for more sophisticated research!

Conclusions:: 

Conclusions: (3) Family instability appears positively linked to fertility (4) Replacement migration may occur even in societies with sustained very low fertility rates (Spain, Switzerland) Immigration continent: EU25 has recently attracted more immigrants than the US Substantial immigration now seen as a part of the SDT (van de Kaa 1999, Lesthaeghe and Neiderd 2006) The importance of immigration: A combination of high immigration, younger age structure and higher fertility rates of immigrants Dalla Zuanna (2006): sustained low fertility-high immigration cycle in Northern Italy. Low-skilled migration mixed with high aspirations for social mobility of migrants‘ children, leading to their fertility reduction

FUTURE OUTLOOK: 

FUTURE OUTLOOK Selected reasons why a modest fertility increase may take place in many countries End of tempo distortion, modest recuperation More family friendly policies, accomodated to the SDT context More immigrants coming from higher-fertility settings Improvement in assisted reproduction Further advancement in the SDT? More family instability? Increase in religiosity? (Many reasons to expect further decline; better known – e.g. Lutz 2006)

FUTURE OUTLOOK: European population implosion postponed?: 

FUTURE OUTLOOK: European population implosion postponed? Eurostat (2004) projection: main scenario – EU population decline after 2025 (also other projections) BUT: higher fertility and higher immigration may postpone the EU population decline well after 2050 (also projection by Alho et al 2006) What about European global demographic and economic marginalisation? It may not have sense to compare continents with fragmented geopolitical units. More meaningful is an EU-US comparison EU territorial expansion enabled it to surpass the US in population growth and to keep pace with its economic power

Slide51: 

Population: US vs. EU-15 and EU (actual borders)

Slide52: 

Economy: GDP in PPP; US vs. EU (actual borders)

Sweden: Most stable fertility level in Europe: Completed cohort fertility among women born 1735-1963 : 

Sweden: Most stable fertility level in Europe: Completed cohort fertility among women born 1735-1963 Source: Statistics Sweden

Future outlook (2):Will the end of fertility postponement bring an increase in the period fertility rates?: 

Country-specific and parity-specific differences in the extent of ´catching up´ at higher reproductive ages First birth probabilities increasing after age 30 in almost all countries Future outlook (2):Will the end of fertility postponement bring an increase in the period fertility rates? Stages of fertility postponement and ´recuperation´: the case of first births in the Netherlands