Uncertainty of Fertility Expectations

Transcription

1 Petra Buhr & Anne-Kristin Kuhnt Uncertainty of Fertility Expectations Longitudinal Analyses with the German Family Panel Presentation at the International User Conference of the German Family Panel in Jena, May

2 Introduction Uncertainty is part of fertility process [Morgan 1981, 1982] Uncertainty of fertility goals influences fertility outcomes [Morgan 1981, 1982, 1985; Schaeffer/Thomson 1989] Recent research does not take this into account uncertainty excluded from analyses [e.g. Berrington 2004; Iacovou/Tavares 2011; Buhr/Kuhnt 2012] uncertainty considered as positive [e.g. Testa/Toulemon 2006] or negative intention [e.g. Miettinen et al. 2005; Mitchell/Grey 2007; Kuhnt/Trappe 2013]

3 Research Questions (1) How large is the proportion of individuals who are uncertain about their fertility plans? (2) How stable is uncertainty over the life course? (3) What are the reasons for instability?

4 What does uncertainty of intentions mean? Uncertainty about whether to have a (further) child or not [Morgan 1981, 1982; Bernardi et al. 2015; Miller 2011] Uncertainty about the timing of childbearing [Morgan 1982; Bernardi et al. 2015] Indifference or ambivalence about fertility intentions [Bernardi et al. 2015; Miller 2011 ] a child maybe a child at times yes, at times no a child yes, but later not sure (how to answer) don t know...

5 Previous Research Prevalence of uncertainty depending on how uncertainty is defined varies between 10% [e.g. Morgan 1981; Kuhnt/Trappe 2013] and 40% [e.g. Morgan 19981; Ní Bhroclcháin/Beaujouan 2011; Sobotka 2009] Reasons for being uncertain middle age [Morgan 1981; Ní Bhroclcháin/Beaujouan 2011; Sobotka 2009] low parity [Morgan 1981; Ní Bhroclcháin/Beaujouan 2011] parenthood status [Sobotka 2009; Sobotka 2009] partnership status [Qu et al. 2000; Ní Bhroclcháin/Beaujouan 2011; Bernardi et al. 2015] labor force status [Bernardi et al. 2015]

6 Previous Research Uncertainty is not a stable condition [Bernardi et al. 2015; Qu et al. 2000; Morgan 1981] Reasons for instability of uncertainty, i.e. changes from uncertain to certain and vice verca: increasing age [Morgan 1981] change in partnership status [Qu et al. 2000] postponement of next birth [Morgan 1981]

7 Theoretical Considerations [Huinink et al. 2011] 1. According to the time dependence of the life course fertility decision making is influenced by past experiences and anticipations of future risks. 2. Parenthood is a long-binding relationship. Children will only be intended and realized if important prerequisites for parenthood are met (e.g. financial security, partnership). 3. A change in conditions (labor force status, relationship status) can lead to a change of intentions.

8 Hypotheses (H1) There is a certain percentage of people who are uncertain about their fertility intentions. (H2) Uncertainty is not stable over the life course. (H3a) Finding a partner reduces uncertainty. (H3b) Becoming unemployed increases uncertainty. (H3c) Uncertainty decreases with age.

9 Data, Variables & Methods German Family Panel wave 1-5 (incl. Demodiff) Dependent variable: Realistically expected number of children Without children: When you think realistically about having children: how many biological or adoptive children do you think you will have? 0,1,2,3,4 and more; not sure, haven t thought about that With children: When you think realistically about having additional children: do you think that you will have more biological or adoptive children? Yes, No, not sure; haven t thought about that

10 Data, Variables & Methods Independent Variables birth cohort ( , , ) partnership status (partner/no partner) employment status (employed/not employed/unemployed) Further variables of interest sex (men/women) parenthood status (children/no children) highest school degree (still at school, low, medium, high) region (east/west) Methods descriptive analyses (prevalence, bivariate analyses) fixed effects models (stability & change)

11 Descriptive Findings Table 1: Prevalence of certainty and uncertainty (about expected number of children) in Wave 1 & 5 by birth cohort Wave 1 Wave Certain Uncertain Not thought about that n 4,314 4,749 4,771 2,398 2,286 2,540

12 Descriptive Findings Table 1: Prevalence of certainty and uncertainty (about expected number of children) in Wave 1 & 5 by birth cohort Wave 1 Wave Certain Uncertain Not thought about that n 4,314 4,749 4,771 2,398 2,286 2,540

13 Descriptive Findings Table 1: Prevalence of certainty and uncertainty (about expected number of children) in Wave 1 & 5 by birth cohort Wave 1 Wave Certain Uncertain Not thought about that n 4,314 4,749 4,771 2,398 2,286 2,540

14 Descriptive Findings Table 2: Change of expected number of children from wave 1 to wave 2 n Percent Stable certainty 7, Stable uncertainty 66 1 Certainly yes uncertain Certainly no uncertain 44 1 Uncertain certainly yes Uncertain certainly no 79 1 Total 8,

15 Descriptive Findings Table 2a: Change of expected number of children from wave 1 to wave 2 (uncertain persons in wave 1 only) n Percent Uncertain uncertain Uncertain certainly yes Uncertain certainly no Total

16 Descriptive Findings Table 2a: Change of expected number of children from wave 1 to wave 2 (uncertain persons in wave 1 only) n Percent Uncertain uncertain Uncertain certainly yes Uncertain certainly no Total

17 Descriptive Findings Table 2a: Change of expected number of children from wave 1 to wave 2 (uncertain persons in 1 only) n Percent Uncertain uncertain Uncertain certainly yes Uncertain certainly no Total

18 Multivariate Findings Table 3: Determinants of change from certain to uncertain (fixed effects logit regression, odds ratios presented) Certain <> uncertain Certainly yes <> uncertain Certainly no <> uncertain No partner partner 0.72*** 0.62** Employed unemployed Employed non-working 0.72** 0.66** 0.91 Number of children 2.35*** 5.77*** 0.57** *** < 0.000; ** < 0.001; * < 0.05, + < 0.1 N of Obs=7503; N of Groups=1817; LR chi2(4) = 84.77; Prob > chi2 = N of Obs=4276; N of Groups=1154; LR chi2(4) = ; Prob > chi2 = N of Obs=2092; N of Groups=624; LR chi2(4) = 12.44; Prob > chi2 =

19 Multivariate Findings Table 3: Determinants of change from certain to uncertain (fixed effects logit regression, odds ratios presented) Certain <> uncertain Certainly yes <> uncertain Certainly no <> uncertain No partner partner 0.72*** 0.62** Employed unemployed Employed non-working 0.72** 0.66** 0.91 Number of children 2.35*** 5.77*** 0.57**

20 Multivariate Findings Table 3: Determinants of change from certain to uncertain (fixed effects logit regression, odds ratios presented) Certain <> uncertain Certainly yes <> uncertain Certainly no <> uncertain No partner partner 0.72*** 0.62** Employed unemployed Employed non-working 0.72** 0.66** 0.91 Number of children 2.35*** 5.77*** 0.57**

21 Multivariate Findings Table 3: Determinants of change from certain to uncertain (fixed effects logit regression, odds ratios presented) Certain <> uncertain Certainly yes <> uncertain Certainly no <> uncertain No partner partner 0.72*** 0.62** Employed unemployed Employed non-working 0.72** 0.66** 0.91 Number of children 2.35*** 5.77*** 0.57**

22 Summary & Conclusion Uncertainty is a broadly defined concept Uncertainty is of relevant prevalence in our sample (= H1) Uncertainty is not stable across life course (period of 5 years) (= H2) Finding a partner increases certainty of fertility intentions (= H3a) Becoming unemployed does not increase uncertainty (!= H3b) Uncertainty is not related to younger ages (!= H3c) Birth of child increases uncertainty

23 Summary & Conclusion It is worthwile to look at uncertainty Surveys should include an answer option related to uncertainty We have only scratched on the surface Next steps: Separate timing & number effects Add more variables (e.g. personality traits) Sequences Consequences of incertainty for family formation & enlargement

24

25 Appendix

26 Table A1: Determinants of being uncertain in Wave 1 (multinomial logistic regression) Uncertain Certainly No Child *** Female *** Birth cohort ** Birth cohort *** 11.19*** Not employed Unemployed Still in school Medium school degree * High school degree *** Partner 0.42*** 0.92 East Germany ** N=13,098; LR chi2 (22)= ; Prob >chi2=0.000; Pseudo R2 = *** < 0.000; ** < 0.001; * < 0.05, + < 0.1

27 Table A2: Change of expected number of children from wave 1 to wave 5 n Percent Stable certainty 5, Stable uncertainty 63 1 Certainly yes uncertain Certainly no uncertain Uncertain certainly yes Uncertain certainly no 81 1 Total 6,

28 Table A3: Determinants of Change from Certainly Yes to Uncertain between Waves 1 & 2 (logistic regression) Child in W Female 1.15 Birth cohort Birth cohort *** Not employed W Unemployed W Still in school W Medium school degree W High school degree W Partner W East Germany W * Certainly Yes Uncertain N=4936; LR chi2 (11)=53.50; Prob > Chi2=0.000; Pseudo R2= *** < 0.000; ** < 0.001; * < 0.05, + < 0.1

29 Table A4: Determinants of Change from Certainly Yes to Uncertain between Waves 1 & 5 (logistic regression) Child in W1 3.25*** Female 1.00 Birth cohort * Birth cohort *** Not employed W Unemployed W Still in school W Medium school degree W High school degree W Partner W East Germany W Certainly Yes Uncertain N=3496; LR chi2 (11)=263.97; Prob > Chi2=0.000; Pseudo R2= *** < 0.000; ** < 0.001; * < 0.05, + < 0.1

30 Table A5: Determinants of Change from Certainly Yes to Uncertain between Waves 1 & 5 (logistic regression) New child born 2.16** Female 1.08 Birth cohort *** Birth cohort *** Not employed W1, unemployed W Unemployed W1, not unemployed W Unemployed W1 & W5 1.4 No Partner W1, Partner W Partner W1, No Partner W No Partner W1 & W5 1.58** East Germany W Certainly Yes Uncertain N=3481; LR chi2 (11)=209.67; Prob > Chi2=0.000; Pseudo R2= *** < 0.000; ** < 0.001; * < 0.05, + < 0.1