The Motherhood Penalty in Financial Resources for Retirement: A Life Course Perspective on the Accumulation of Public Pension Wealth and Personal Wealth in East and West Germany

2.1 Pensions

While reforms over the last decades have placed greater importance on private pension provision, public pensions remain central to old-age material well-being. In Germany, public pensions posit the greatest source of income for older individuals. Of those 65 years and older, 90 % received public pension benefits with a net average 1,109 Euros per month (BMAS 2020, 86). For women, this net amount is lower at 995 Euros on average (BMAS 2020, 86). Overall, public pensions represent 61 % of the average old-age income compared to 8 % for occupational and 7 % for private pensions (BMAS 2020, 99f).

The German public pension system follows the Bismarck model, emphasizing pension calculation primarily on individual earnings and employment biography. Most employees are insured under the scheme, except for e.g. civil servants and the self-employed, who have separate schemes. Pension entitlements are calculated by collecting ‘earning points’ on a monthly basis through pension contributions. Here, one point represents contributions equivalent to those of the average salary of all insured individuals across one year. The system encompasses various benefits which are awarded for periods of non-employment such as those related to unemployment, education, and child-rearing. In the context of our research question, benefits related to child-rearing are of particular interest. The earning points granted to parents, typically mothers, were significantly reformed and increased in 2014 and 2019. Currently, parents receive 2.5 earning points per child for children born before 1992, and 3 earning points per child for those born from 1992 onwards. These points are generally allocated to the mother^[1] therefore commonly called ‘Mütterrente’ (Mothers’ Pension). Additionally, if the youngest child is under 11 years old, any points earned from part-time work are increased by up to 50 %, with a maximum additional increase of 0.0278 points.

Due to the life-course sensitivity of public pensions, a substantial gender pension gap can be attributed to persistent differences in the life courses of men and women. Women often reduce or give up paid employment to provide unpaid care. In 2020, while 75 % of mothers participated in paid employment, there were significant disparities in part-time work. Among working mothers with minor children, 66 % were employed part-time, compared to 35 % of childless women and only 7 % of fathers (Statistisches Bundesamt 2022). Among women, those with a particularly weak attachment to the labor market over their life course are at a particular risk for low pension entitlements (Frommert et al. 2013).

However, women’s life courses are not uniform across East and West Germany. West Germany has historically promoted a male breadwinner model of labor division. While couples in the immediate post-war cohorts show a tendency towards a ‘pure’ male breadwinner constellation, in younger cohorts, a shift towards a 1.5 earner constellation with women re-entering part-time employment after childbirth can be observed (Möhring and Weiland 2022). In East Germany, in contrast, the German Democratic Republic (GDR) encouraged female labor market participation to remedy labor shortages. The well-developed infrastructure of full-time public childcare allowed mothers to re-enter the labor market quickly after childbirth. While incomes were generally less stratified in the GDR than in West Germany, gender-specific earnings differentials also existed in the former (Rosenfeld et al. 2004). Consequently, women in East Germany show a much higher labor market attachment than their Western German counterparts. In combination with the lower retirement income of East German men in comparison to their West German counterparts, overall gender-specific retirement income differentials are substantially lower in East than in West Germany (Grabka et al. 2017).

In sum, previous research shows that despite these non-employment-related benefits tailored towards mothers and their reduced labor market participation, a motherhood penalty in pensions persists in Germany. Mika and Czaplicki (2017) investigate the motherhood penalty in lifetime earnings in East and West Germany using linked survey and administrative data. Considering child-rearing credits, they still find a significant motherhood penalty for West German women with up to two children. For women with three or more children, the child-rearing benefits seem to outweigh the losses due to reduced employment participation. There is a small motherhood premium in East Germany, reflecting the stronger labor market attachment of East German mothers, but possibly also the lower prevalence of childlessness in the East (Statistisches Bundesamt 2024).

2.2 Personal Net Wealth

Net wealth is typically operationalized as the sum of assets minus liabilities. In Germany, housing wealth posits the greatest source of net wealth (Grabka 2014). Personal net wealth, in addition to retirement income, can provide material security in later life by covering consumption or, in the case of a family home, reducing living costs (Vogel et al. 2022). Scrutinizing personal rather than household wealth, we pay tribute to the couple wealth gap, which indicates that within couples, women typically own less wealth than their partners (Grabka et al. 2015).

Analogous to public pension entitlements, labor market engagement and income are predictors of net wealth (Killewald et al. 2017), contributing to gendered outcomes (e.g. Sierminska et al. 2010). In addition, further mechanisms, other than individuals’ performance on the labor market, may enable the accumulation of wealth. First, transfers such as in vivo transfers and inheritances contribute to accumulating wealth (Killewald et al. 2017). Second, while we look at wealth on a personal level, the sharing or joint ownership of wealth within households, despite one spouse proving a larger initial investment, can lead to to a de facto redistribution of resources such as in the case of male-breadwinner couples equally dividing housing wealth (e.g. Joseph and Rowlingson 2012; Nutz 2022). While not all couples pursue a community property regime or assign joint ownership only to certain components, such as housing wealth, these arrangements have been linked to lower gender-specific wealth differences (e.g. Frémeaux and Leturcq 2022). Lastly, macro-level events such as stock market or housing market crises may have varying and potentially gendered effects on wealth portfolios (e.g. Young 2010).

In Germany, the former division into East and West Germany has had a profound and long-lasting impact upon wealth accumulation. In the German Democratic Republic, private wealth accumulation was inhibited under the socialist regime. Firms as well as land was socialized, leaving only housing wealth and savings deposits as possible routes for saving. Since home ownership was restricted, most savings went into savings deposits (Albers et al. 2020). To this day, home ownership is more widespread in West Germany with about one half of households owning a home compared to only around a third of East German households. Levels of wealth also differ between East and West Germany, with a mean of €67,400 in Eastern compared to more than double €153,200 in the Western Germany (Grabka 2014).

With respect to gender, studies find a substantial gender wealth gap in Germany (Sierminska et al. 2010). Lersch et al. (2017) examine parenthood and personal wealth from a life course perspective using the SOEP. They find that mothers accumulate significantly less personal wealth compared to fathers and childless women, a gap that worsens with early childbearing. In a separate study, Lersch (2017) finds that marriage increases household and personal wealth for both men and women. Conversely, divorce leads to significant short- and long-term wealth losses for both genders, with more pronounced effects for women (Kapelle 2022).

Less research has been done on the joint distribution of pensions and wealth. Utilizing augmented wealth as a combined measure of net wealth and pension wealth, Kuhn (2020) finds a lower extent of general inequality in the combined measure than in pension wealth for Switzerland. In contrast, Bönke et al. (2019) find that the Gini coefficient of an augmented wealth measure is lower than for net wealth in Germany, while Bartels et al. (2023) underscore that pension wealth accounts for the majority of resources of the lower half of the wealth distribution. Using waves 2012 and 2013 of the SOEP, Cordova et al. (2022) investigate the effect of including self-reported pension entitlements in a wealth aggregate, finding that the gender wealth gap is reduced. This could be an indication that compensatory elements within the German Pension Insurance contribute to a more equal distribution of pensions as compared to net wealth. In this study we decide against using a joint measurement of pension and net wealth - such as augmented wealth – in favor of a comparative perspective on pension wealth, with and without compensatory revaluations, and net wealth. This is due to (1) previous research indicating different levels of inequality, which would be obfuscated by a joint measure. Prospects of increased pension entitlements may further disincentive net wealth accumulation, underlining the importance to investigate both measures separately. (2) At the same time, the characteristics of net wealth and pension entitlements are very different in terms of their liquidity and utilization for material well-being in old age (Bönke et al. 2019; Cordova et al. 2022).

3.1 Data

In the joint investigation of public pension wealth and net wealth, we use linked survey and administrative data. The SOEP-RV links data from the German Socio-Economic Panel Survey with retrospective data from the German Pension Insurance (Lüthen et al. 2022). Since 2019, SOEP respondents have been asked to consent to the linkage of their data with administrative records of the German Pension Insurance. In the SOEP release we use here (v38), 12,928 individuals consented to this linkage.

Linked data uniquely facilitates the joint analysis of public pension entitlements and net wealth. Previously, investigations into these areas have relied on survey-based measures (Bönke et al. 2019, 2020; Cordova et al. 2022). The use of linked data offers greater precision in measuring public pension entitlements. Administrative data ensures higher accuracy in recording earning points since it is collected directly from employers, avoiding the inaccuracies of individual self-reporting. Moreover, individuals are highly motivated to self-report pension-related events, such as childbirth, due to the significant benefits they receive from accurate reporting. At the same time, the SOEP offers crucial measurements not found in administrative data, such as personal net wealth and socio-demographic characteristics like educational background and marital status. These additional variables are essential for our investigation, providing a comprehensive view that administrative data alone cannot offer.

3.2 Sample

For our sample, we focus on women born between 1937 and 1989 who lived in East- or West Germany in 1989. Since civil servants and many self-employed individuals are insured under separate retirement schemes in Germany, we excluded them from the analysis (see Appendix A Section 1 ‘Additional Information on Self-Employment and Civil Service’ for our identification strategy in Supplementary Material). For our outcome variables, two further requirements need to be met: Firstly, individuals are required to have participated in at least one of four possible wealth observation in the SOEP (2002, 2007, 2012 or 2017). Secondly, we require that the individual consented to the linkage project, which was first possible in the 2019 wave of the SOEP. Table 1 shows our sample restrictions. Additionally, Appendix A Section 2 in Supplementary Material includes a comparison of characteristics of the original SOEP sample and our final sample. Thus far, proposed weighting strategies taking this selection into account, to our knowledge, only address cases in which data from after 2019 is used (Lüthen et al. 2022). The most notable difference is an overrepresentation of wealth observations in later survey years, which is to be expected given the consent requirement in 2019. Our final sample consists of 3,680 individuals (see Table 1) of which we include 7,275 person-years, with an average of 2.5 observations per individual. Although we have continuous monthly observations of pension entitlements, for consistency reasons, we restrict the observations to those years in which we have a wealth observation available .

3.3 Measurements

3.4 Analytical Approach

We apply a dynamic perspective on the motherhood penalty in pension wealth and net wealth, modeling accumulation trajectories across the life course. Our focus lies on differences between accumulation rates in mothers and childless women, denoted as maternal and childless trajectories in the following. We utilize random effects growth curve models, a multilevel method where individuals’ observations across time are nested in individuals. Random effects growth curve models account for serial correlation and allow for unbalanced panel data. Since models exploit both, between-individual and within-individual information, they allow for data from individuals with one observation point to contribute to the analysis. In comparison to fixed effects models, random effects models can accommodate both time-constant and time-varying variables (Singer and Willett 2003). In its most basic form, our model can be represented as follows:

where Wealth_it represents the (net/public pension) wealth of an individual i at time t. γ₀₀ represents the average intercept. Additionally, we include a random component to the intercept (ζ_0i), which allows for variation of the intercept between individuals due to unobserved characteristics. Our main predictors are age and years since first birth (as well as years since second birth in later models, not represented in the formula). For both age and years since birth, we include three polynomials, which, for the sake of readability, are not included in the formula. For both, age and years since birth, the polynomials allow the effect of this predictor to be s-shaped, meaning that e.g. an additional year since birth may have a different effect immediately after birth and 20 years after birth. For both predictors and their polynomials, we include an interaction effect with East/West Germany, hypothesizing that the accumulation trajectories of mothers and childless women will differ due to the distinct historical and cultural settings. Subsequently, we construct a second set of models including years since second birth as additional predictor. This enables us to examine whether the motherhood penalty differs for maternal trajectories with one and two births. Control variables (represented in the formula as Controls_it with the corresponding coefficient γ₁₅) are added to the models in a stepwise fashion. In our main models we control for East Germany, cohort, survey year, and wealth imputation flag (for net wealth). In subsequent steps, we add education and migration background, marital status at the time of observation, and employment history characteristics (years in full time and years in part-time employment) (see Figures B1–B3 and Tables B1–B6 in Appendix B in Supplementary Material).

For interpretation purposes, we predict outcomes for three different archetypal accumulation trajectories: one childless, one maternal with one birth, and one maternal with two births. In a first set of models, we compare childless and maternal trajectories. This is done as follows: for the childless trajectories, we predict the outcome at different ages, keeping the variable ‘years since birth’ at zero. Maternal trajectories represent ‘median trajectories’. Here, ‘years since birth’ starts at the median age of first birth to model the time elapsed since birth, which corresponds to age 26 in our sample. In the second set of models, we further introduce maternal trajectories with two or more children, utilizing the variable ‘years since second birth’, which starts at age 29. We model accumulation trajectories over a 40-year period, from age 25 to age 65. However, our observation window covers only 15 years, from 2002 to 2017. This discrepancy introduces some diffusion between age and cohort effects, as observations at older ages are influenced by older cohorts, and those at younger ages are influenced by younger cohorts. This is an inherent limitation, encountered by studies utilizing the available German data to explore the accumulation of personal net wealth (e.g. Lersch et al. 2017).

4.1 Pension Wealth

The upper two panels of Figure 1 (see also Table B1 in Appendix B in Supplementary Material) depict the accumulation trajectories of employment-related public pension wealth for mothers and childless women in East and West Germany, not accounting for any redistributive elements of the German pension system such as care credits. Accumulation trajectories of childless women are similar in East and West Germany, with mothers in both regions accumulating significantly lower employment-related public pension wealth across the life course. In East Germany, maternal accumulation trajectories are associated on average with about 99,500 Euros smaller public pension wealth than those of childless women by age 65. In West Germany, the motherhood penalty is much more pronounced, with an average gap of about 240,539 Euros.

Figure 1:

Growth curve models of public pension wealth from employment. Note: Controls for survey year and cohort. Figure by Möhring, K., Overweg, C. and Weiland, A. P., licensed under CC BY 4.0. https://osf.io/gpmej/. Source: Own estimations based on SOEP(v38)-RV.

The upper panels of Figure 2 (see also Table B3 in Appendix B in Supplementary Material) depict results for the comprehensive measure of public pension wealth, accounting for employment-related benefits and revaluations. Accounting for periods of child-rearing, as well as revaluations of part-time employment during child-rearing, reveals shifting dynamics between maternal and childless accumulation trajectories of women in East Germany. Here, accumulation rates of mothers overtake those of childless women between their mid-30s to their late 40s, resulting on average in about 159,408 Euros of public pension wealth at age 40 versus 124,331 Euros for childless trajectories. Subsequently, at age 65 public pension wealth levels have converged and differences are not statistically significant. In West Germany, the penalty in maternal public pension wealth accumulation remains substantial. Trajectories of mothers and childless women diverge later than for employment-related public pension wealth, at around 40 years as opposed to between 30 and 35 years. However, as employment differences persists beyond the main child-rearing years, the motherhood penalty still amounts to about 155,249 Euros by age 65.

Figure 2:

Growth curve models of public pension wealth from employment and revaluations. Note: Controls for survey year and cohort. Figure by Möhring, K., Overweg, C. and Weiland, A. P., licensed under CC BY 4.0. https://osf.io/gpmej/. Source: Own estimations based on SOEP(v38)-RV.

The bottom two panels in Figures 1 and 2 (see also Tables B2 and B4 in Appendix B in Supplementary Material) depict the accumulation trajectories of childless women compared to mothers with respectively one or more than one child. In East Germany, employment-related public pension wealth trajectories of mothers with one child are not significantly lower than those of childless women, while those of mothers with two or more children have accrued about 104,307 Euros less by age 65. This suggests that the motherhood penalty in employment-related public pension wealth observed in the first set of models is predominantly driven by mothers with multiple children. In the comprehensive public pension wealth measurement, differences between trajectories of childless women and mothers are smaller (see Figure 2, bottom panels). Accumulation patterns of mothers with one child and mothers with several children are similar, with both groups exhibiting significantly higher accumulation rates than those of childless women between the ages of 35 and 45 due to care-related revaluations. By the end of the observation window, accumulation trajectories of all three groups have converged. In West Germany, accumulation trajectories of employment-related public pension wealth are clearly ranked along the number of children, with childless women holding about 430,942 Euros at age 65, compared to 252,090 Euros for women with one child and 179,350 Euros for women with two or more children. Accounting for public pension revaluations, substantially closes the gap between mothers with one or more children and narrows that towards childless women. However, differences at age 65 remains considerable at about 463,839 Euros for childless women compared to about 322,821 Euros for mothers with one child and about 306,090 Euros for mothers with more than one child.

Exploring correlates of pension wealth accumulation, we find stratification along respondents’ education, which likely acts as a proxy for earnings. However, controlling for education does not explain the gap between mothers and childless women in West Germany. Furthermore, including familial status at the time of the interview does not greatly alter the gap, indicating that the motherhood penalty in both, employment-related and comprehensive public pension wealth, cannot be explained by familial status (see Figures B1 and B2 in Appendix B in Supplementary Material). Familial status itself, however, has a significant effect. Single women accumulate significantly less public pension wealth compared to married women, likely due to selectivity and a strong correlation with motherhood. In contrast, divorced women benefit from pension splitting regulations following union dissolution. Controlling for employment biographies, the patterns for East and West German mothers become more similar. Both groups now exhibit a motherhood premium in midlife, which is attributable to childcare credits: Vice versa, when employment trajectories are held constant, mothers benefit from the additional entitlements they receive through care-related revaluations in East and West Germany alike.

4.2 Net Wealth

The upper two panels of Figure 3 depict maternal and childless accumulation trajectories of IHS-transformed^[5] personal net wealth for women in East and West Germany (see also Table B5 in Appendix B in Supplementary Material). Generally, East German childless and maternal accumulation trajectories achieve lower wealth levels than their respective West German counterparts. Furthermore, in both regions, maternal wealth accumulation trajectories experience a penalty in comparison to those of childless women. In East Germany, the former exhibit an IHS-transformed wealth score of about 4.8 by age 65, while the latter reach a score of about 6.9. This difference is equivalent to about 87,497 Euros on a non-transformed scale. The difference becomes significant at a slightly later age (in the late 40s) than for West Germany. Here, the wealth trajectories of mothers and childless women begin to diverge significantly at age 40. By age 65, maternal wealth trajectories achieve an IHS-transformed score of about 5.7, compared to 7.9 for childless trajectories, which translates into about 30,119 and 281,037 Euros, respectively. When accounting for the number of children (see bottom panels of Figure 3, Table B6 in Appendix B in Supplementary Material), we find that trajectories between mothers of one or multiple children do not significantly diverge in West Germany. In East Germany, trajectories of childless women and mothers with one child do not significantly differ, indicating that the motherhood penalty predominantly applies to wealth trajectories of mothers with multiple children.

Figure 3:

Growth curve models of personal net wealth (IHS-transformed). Note: Controls for survey year, cohort, and wealth imputation flag. Figure by Möhring, K., Overweg, C. and Weiland, A. P., licensed under CC BY 4.0. https://osf.io/gpmej/. Source: Own estimations based on SOEP(v38)-RV.

Our supplementary analyses show (see Figure B3 in Appendix B in Supplementary Material), that accounting for differences in education and employment mitigates the gap between maternal and childless trajectories, whereas including familial status increases the motherhood penalty. In terms of net wealth, married women are at a clear advantage, even after controlling for years spent in full-time and part-time employment. Including education, familial status and employment trajectories renders the difference between mothers and childless women insignificant in East Germany. In West Germany, the penalty remains significant throughout midlife, with a brief period of insignificance towards the end of the observation window.