Despite the increase in the segment of populationstaying childless, the mechanisms accompanying this choice remain unclear andmainly untested (Tanturri & Mencarini,2008). This paper contributes to the understanding of which factors lead toliving a life without children by focusing on negative fertility intentions ofmen and women and on their later fertility realization. Following the frameworkof the Theory of Planned Behavior (Ajzen, 1985; 1991) we analyse the extent to which intentionsto be childless are stable over the life course and whether these intentionsare later realized.

Inorder to track the link between intentions to be childless and fertilityoutcome, analyses are achieved using the Generations and Gender Survey (GGS) andits corresponding follow-up study. The final sample includes men and women aged 18-45 who were childless atthe first interview, sexually active and living in Austria, Bulgaria, CzechRepublic, France, Georgia, Germany, Italy, Lithuania and Russia (N = 4,505).

To answer the first research question we observe whether and howindividualsÕ intentions to be childless at the first wave change in the shorttime. As shown in Figure 1 we distinguish between stable positive intentions,stable negative intentions and those who change their minds in the period. Followinga similar structure, intentions are matched with the fertility status afterthree years, differentiating between those who realize their negativeintentions, those who realize their positive intentions and those who did notrealize their intentions.

Results of a confirmatory factor analysis, which also tests for the measurementinvariance in the different countries, confirm that the available items act asvalid and reliable measures of the TPB variables. Other key explanatoryvariables capture the information on partnership and working history, the ageof the respondent and the socio-economic characteristics. To identify thefactors associated with stable intentions to be childless, the relationship isexpressed in terms of a linear equation by applying a logistic transformation.

Descriptive results show that intentionsto be childless are less often maintained in the short run compared tointentions to be parents among all the group of ages, and are also much moredependant on age than intentions to be parents.

The preliminary multivariate evidence (Figure 4) showhow people with stronger attitudes against parenthood tend to have more durableintentions to be childless and less stable intentions to be parents, as well asa higher probability to realize this intention in the short run, while theymore often quit their plans toward parenthood.

At the same time, stronger subjective norms related toparenthood strongly reduce both the stability and the realization of childlessplans, showing how a greater perception of normative pressure on having a childreduces the chances of maintaining and realizing childless intentions. Lastly,people perceiving a higher behavioural control on having children tend to have slightlylower chances to shifting plans toward parenthood, as well as weaker chances torealise their plans, especially among women.

Looking at the stratification of the phenomenon(Figure 5), we find a reversal in the association between education andchildlessness that show – in line with recent findings (Kreyenfled & Konietzka, 2016)– a reduced effect of education on voluntary childlessness, especiallyfor women. Moreover, changes in family life also play a role in intendedchildlessness and its realization: being in a stable or new partnershipstrongly reduces voluntary childlessness, still without eliminating theprobability of giving up on having children. This indicates how voluntarychildless is a choice that is also taken within the couple. Contrarily to ourexpectation, changes in the working condition does not seem to play asignificant role in realizing or giving up on fertility intentions.

Figure 1 – Stability of intentions andcoherence of realization to be childless. Multinomial Logistic Regression (AME)

Figure 2 - Stability of intentions andcoherence of realization to be childless. Multinomial Logistic Regressions(AME)

Concluding, previous research has shown how socio-demographic andpsychological factors covered by the TPB influence childbearing intentions (Billari et al., 2009). We show how such factors also play arole in influencing a) the intentions to be childless; b) thestability of both intentions toward parenthood and intentions towardchildlessness; c) the realization of bothpositive and negative intentions. We provide new insight about the link betweenintentions and actual behaviour: given that failure in realizing intendedparenthood happens more often than failure in realizing intended childlessness,we conclude that intentions are a stronger predictor of behaviour in the caseof intended childlessness than in the case of intended parenthood. By the timeof the EPC 2018 additional analyses and robustness checks will be completed.