www.ssoar.info Having power, having babies? Fertility patterns among German elite politicians

Members of the political elite have far-reaching influence on the overall society. In this paper, we analyse fertility patterns among the German political elite for two reasons: First, we learn more about the living circumstances of a subgroup that makes crucial decisions and could serve as a role model for the general population. Second, we gain insight into the association between social status and fertility patterns at the top tier of the status distribution. We collect biographical data from all high-rank politicians in Germany in 2006 and/or 2017, comprising 184 women and 353 men. We compare fertility patterns in this subgroup to the general population, as well as we differentiate the number of children by politicians’ gender, region (eastern/western Germany), party affiliation, and other variables. Results show that, on average, male politicians have relatively many children: 2.0 in western Germany, and 2.2 in eastern Germany. Female politicians have very few children in western Germany (1.3) and relatively many in eastern Germany (1.9). The east-west gap between men and women is entirely driven by differences in childlessness. For men, the observation of high fertility in this high-status group could hint towards a positive association between social status and fertility at the top of the status distribution. For women, large east-west differences in this subgroup could mean that the association between social status and fertility at the top of the status distribution might be negative or positive, depending on macro-level characteristics such as gender norms and work-family reconciliation policies.


Introduction and background
This paper aims to provide first novel insights into the demographic behaviour of a subgroup of the population relevant for its far-reaching influence on society. Our three main motivations for this descriptive analysis of fertility patterns of elite politicians are as follows: First, we can observe the lives of people who make influential decisions that influence society. Political decisions and efforts may be influenced by politicians' individual backgrounds, including the number of children they have (Baumann et al. 2013). Parenthood might also have an influence on political success (Smith 2017).
Second, because of their prestige, elites could serve as role models for the general population, and thus influence their behaviour (Bohlken 2011: 70;Skirbekk 2008). Therefore, it is interesting to investigate characteristics such as parenthood -especially in Germany, a country known for low fertility rates.
Third, we can learn more about the link between social status and fertility by looking at a subgroup at the top of the societal status distribution. Fully understanding an association means understanding it in all parts of the distribution. As many associations are not linear, it is dangerous to extrapolate associations from the middle of the distribution to the upper or lower end. Regular representative surveys do not capture elite groups -the upper end of the social status distribution -in adequate numbers. High education levels among women, high income, and a heavy workload are often associated with childlessness or low number of children. Exploring the number of children born to elite politicians will enlighten whether this association is also prevalent when high status characteristics are pronounced: Elite politicians are characterized by high education, far-above average incomes, very heavy workloads, societal power, and prestige. So, how many children do they have?
Elite studies in Germany. Although elites are a popular research topic in the German social sciences (cf. Hoffmann-Lange 2001), none of the previous studies have investigated fertility patterns. The main target groups in elite studies have been comprised of individuals from different elite sectors including politics, business, public administration, the judiciary branch, associations, science, culture, and media (compare term "elite pluralism" by Hoffmann-Lange, 2003: 114). Previous elite studies have focused on social background, career paths, connections between the elite sectors and attitudes towards social inequality (Bürklin/Rebenstorf 1997;Gruber 2009;Hartmann 2013;Kaltefleiter/Wildenmann 1972;Wildenmann et al. 1982;Wildenmann 1968). Our goal is to fill the research gap regarding fertility patterns by studying one specific elite group, which Hoffmann-Lange (1992: 403f) describes as one of the most influential and powerful -and therefore most important -elite groups: politicians.
Fertility in eastern and western Germany. Germany has one of the lowest fertility rates in the world (Buhr/Huinink 2015). In the general population, women born in the 1950s in eastern Germany have on average 1.8 children, while their counterparts from the west have around 1.6 children. In the subsequent cohorts, fertility patterns converged as the number of births declined sharply in eastern Germany and only moderately in western Germany. For women born between the mid-1960s and mid-1970s, the average number of children is between 1.5 and 1.6 in both regions. Although the average number of children is similar for this cohort, parity structures differ by region: In the west, women are more likely to remain childless than in the east, but those that have a first child are more likely to also have a second, third, or fourth (childless: 21% vs. 14%, share with 3+ children: 18% vs. 13% for women born in the 1960s, Bujard/Lück 2015; Goldstein/Kreyenfeld 2011). We are not aware of any comprehensive data on cohort fertility patterns of men in Germany.
Findings on social status and fertility: Men. The reported studies use education and/or income as proxy measures for social status. Very few deal with the association between social status and the number of children born to men in Germany (these are: Miettinen et al. 2015;Ruckdeschel/Naderi 2009;Schmitt 2005). Available evidence from Germany and other western societies suggests that men with high social status are less likely to remain childless in most countries, mainly because they are less likely to remain single (Barthold et al. 2012;Hopcroft 2015;Miettinen et al. 2015;Trimarchi/Van Bavel 2017). In the Nordic countrieswhere data quality on this subject is best -men with high social status have a higher average number of children (Jalovaara et al. 2017;Kravdal/Rindfuss 2008;Nisén et al. 2018). Findings on social status and fertility: Women. In Germany, women with higher levels of education and longer working hours tend to postpone motherhood, are more likely to remain childless, and have fewer children on average. These associations are stronger in western Germany than in eastern Germany (Blossfeld/Huinink 1991;Bujard 2015;Dorbritz 2015;Kreyenfeld/Konietzka 2017). Typical explanations for these east-west differences include differences in gender norms and in work-family reconciliation policies between these regions (e.g. Hudde/Engelhardt 2017). In eastern Germany work-family reconciliation is more favourable because the offer of public childcare is higher and the attitudes towards working mothers are more supportive (e.g. Zoch/Hondralis 2017). In international and comparative research, a number of authors argue that the association of social status and fertility is becoming less negative or even vanishing over time, especially in the Nordic countries (Jalovaara et al. 2017;Kravdal/Rindfuss 2008;Skirbekk 2008). Recent changes in German family policy indicate a development towards the Nordic model (Fleckenstein 2011). However, there is no empirical evidence that shows the association between fertility and social status to have vanished. Based on these findings, we expect to observe low fertility rates among female elite politicians in Germany, especially in western Germany.
This study. In this study, we analyse fertility patterns of 537 high-ranking German politicians and compare their number of children to those of the general population.

Data
The data set. We adopt Gruber's (2009) definition of the German political elite, understood in a positional sense (Positionselite): Someone belongs to the elite if she or he is in a position with access to influence and power (Gruber 2009: 41;Hoffmann-Lange 1992: 20). Gruber (2009) defines 369 positions of the German executive and legislative branches as the political elite at both the regional and federal level, including positions such as party leaders, ministers, heads of government, parliamentary party group leaders, committee chairs, and secretaries of state (for the list of all positions, see Gruber 2009: 275ff). As for this study, we analyse politicians that are in the Gruber's list (cut-off date: March 1 st , 2006) and add all politicians that held the same positions on July 1 st , 2017 (some politicians hold an elite position in both years). The overall data set includes 670 politicians.
Data collection. We collected data from all 670 top politicians in July and August 2017. All data (except for position and party) are respective to the year 2017. Our data stem from three sources, listed here in order of priority: (1) the Munzinger biographical online database (www.munzinger.de), (2) numerous volumes of Kürschner's collection of self-written short biographies of members of federal and state parliaments (e.g. Holzapfel 2016; see dataset for full list), and (3) Wikipedia. If none of the above sources reported information about children, we assumed that the politician is childless. To challenge this assumption, we explored various other sources (e.g. media reports; web-pages of the politician, parties or parliaments), which in several cases directly confirmed our assumption with explicit statements of the politician being childless. For the remainder, additional sources did not mention whether the politician did or did not have children. Most importantly, there was not a single case for which the additional sources contradicted our assumption that the politician was childless. We are therefore confident that our information on the number of children is reliable; however, we might still slightly underestimate the actual number of children by using this approach. The data set available online at Datorium shows all politicians and the sources used to gather their information (Friedrich/ Hudde 2018).
Sample selection. As we are interested in the final number of children, we study women aged 45+ and men aged 50+ in 2017 (82%, n=549). As we are also interested in differences by party affiliation, we further exclude politicians from very small parties (n=4) or without party affiliations (n=8). We also calculated mean values including these politicians, and our results were robust. The final sample size for our analysis is 537 German elite politicians.
Reference data: Number of children in the general population. To compare the average number of children born to top politicians with the overall population, we used data from the 2012 Microcensus (RDC of the Federal Statistical Office and Statistical Offices of the Länder 2012). Our reference data refer to the politicians' own birth cohorts, resulting in cohort-weighted averages. We compare our sample of elite politicians to two groups: the general female population and the general female population with a university degree (84% of elite politicians have a university degree). Unfortunately, we have neither reference data for male fertility in the general population nor the general male population with a university degree. Table 1 lists background variables of the sample. Roughly a third of the elite politicians are women, and slightly more than a fifth are from eastern Germany. The vast majority of politicians is married. Singlehood is very low in all groups except for women in the west. Unmarried cohabitation is rare in all groups. Elite politicians have high levels of education: only one in seven does not have a university degree. One in five female and one in three male politicians hold a PhD title.

Comparison to the general population
Average number of children: Female vs. male politicians from the east vs. west Figure 1 presents the average number of children for both elite politicians and the general population by gender and region. Among female politicians, the average number of children is 1.3 in the west and 1.9 in the east. 1 For men, the average number of children is 2.0 and 2.2, respectively. We compare these numbers to two age-standardised reference groups of the general population: all women (by region), and all women that have a university degree (by region). To compare number of children, one-sample t-tests are used (for the sake of simplicity, we treat the reference values from the Microcensus as fixed values, as standard errors are negligible). Female politicians in western Germany have significantly less children than the general population and insignificantly less than the general population with a university degree (1.29 vs. 1.63, p<.01; and 1.29 vs. 1.39, p>.1). Their counterparts in the east have insignificantly more children than the general population and significantly more children than the general population with a university degree (1.86 vs. 1.67, p>.1; and 1.86 vs. 1.49, p<.01).
Unfortunately, no reference fertility data is available for the general German male population. 2 Male politicians appear to have relatively many children: 2.01 in western and 2.21 in eastern Germany.  Tables 2 and 3 show the parity distribution among politicians and the general population. 40% of female politicians from western Germany are childless, compared to "only" 20% in the general population and 29% in the general population with a university degree. Among female politicians in eastern Germany, childlessness is below 10% (note the low number of cases). Among male politicians, childlessness is generally much lower, at 17% in the west and 5% in the east. While for the general population, having 3+ children is more common for women in the west than it is in the east, this association seems to be reversed or absent among elite politicians (difference not significant). No substantial regional differences in progression rates to the second or third child are apparent for male politicians.

Differences by party
Looking at both regions together, differences in the number of children between parties are minor and do not seem to follow any clear pattern. However, this changes as soon as we separate by region, as displayed in Figure 2. Both male and female politicians from more conservative parties in the west tend to have more children (note the low case number of observations and large confidence intervals in some of the groups displayed). The association seems to be stronger in 2017 than in 2006 group (not shown here). Members of the CSU, the Bavarian regional branch of the Christian Democratic parliamentary group, which is associated with the most conservative family values and policies (Fleckenstein 2011), have the most children. We find no clear party-based patterns in eastern Germany.

Multivariate regression analysis
We run multivariate negative binomial regression analyses, by gender, to explore whether substantial, significant associations between the number of children and other variables (including interactions with region) are present (Cameron/Trivedi 2013). Table 4 lists the estimated results. Here are some selected observations. Fertility is substantially and significantly higher in eastern than in western Germany for men and woman across all models. Men from different birth cohorts do not seem to differ in their number of children, whereas younger women (from later cohorts) appear to have slightly fewer children than their older counterparts. Political conservatism is significantly associated with higher fertility among women and men in western, but not in eastern Germany. For men, we find opposite educational differentials in east and west: in the west, holding a PhD title is significantly associated with higher fertility, while in the east the association seems to be the opposite. Are members of the core elite -those who hold the most powerful societal positions within the sample -any different from the "regular elite"? Hartmann (2013: 30f) presents a list of political positions that have the greatest access to power (around 40% of all elite positions in the sample) and terms politicians in such positions the political core elite. We find no statistically significant differences between this core elite and the remainder of the political elite. There are also no evident significant associations between religious denomination and the number of children (note that existing literature suggests that religiosity, rather than religious denomination, is important here; see Zhang 2008). Of course, the absence of significant results might partly be a consequence of the low numbers of observations in some of the groups observed. Note: p-values in parentheses. + p < 0.10, * p < 0.05, ** p < 0.01.

Discussion and conclusion
To our knowledge, this is the first study exploring fertility among contemporary societal elites. We are able to show that fertility patterns among German elite politicians differ substantially by gender and region. The low fertility rate for female elite politicians in western Germany is in line with previous findings. The negative association between social status and fertility might be driven by deficient work-family reconciliation. In eastern Germany, elite politicians have on average more children than the general population as well as the sub-population with a university degree. Results for men in both regions and women in eastern Germany might be a sign of a positive association between social status and fertility at the top of the distribution.
Our study has limitations. First, despite full data coverage of the German political elite, the number of cases is rather low, especially for female politicians in eastern Germany. Second, we do not have information on most politicians' partners, which would allow us to more accurately explore fertility patterns from a couple perspective. In addition, no information on the age of their children is available for most politicians, rendering this analysis unable to examine in which career or life course stage elite politicians had their children. Third, we are not able to uncover causal mechanisms with the analysis of these data.
Future research could study the causal mechanisms behind the observed descriptive associations: Does being part of the elite influence fertility behaviour, for example indirectly through attractiveness on the partner market? Or do certain fertility behaviours increase political career opportunities, such as people being more likely to vote for familyoriented politicians? How do career and fertility interact over the (gendered) life course? Further, future research could study fertility patterns among other elite groups such as top-tier business, culture, and scientific subgroups, as well as in other societal contexts.
All in all, this paper is innovative as it presents first insights into the fertility behaviour of a highly relevant subgroup: those that have great power and societal influence. Further, this investigation serves as a starting point for a more detailed analysis of elite demographic behaviour and the association between fertility and social status at the highest tier of the distribution.