3.1 Introduction

In Chapter 2 we demonstrated the complexity of the interdependencies between fertility and women’s labour supply at the macro-level. The data presented suggest that the interrelationship may depend on the incompatibilities between fertility and women’s work which varies across countries, but also on some other country-specific factors. If the mechanism underlying fertility and labour market behaviours is to be understood, the variables governing their relationship, including the intensity of the conflict between the two activities, need to be better explored.

For that purpose, we follow the recommendation of methodological individualism and we move our considerations from the macro- to the micro-level. This does not mean, however, that we ignore the macro-level perspective. On the contrary, we see individuals as rational actors embedded in the macro-context that continuously influences people’s life choices. Furthermore, we look at the decisions they make from the life-course perspective, taking into account not only the fact of the occurrence of the events, but also their sequence and timing. Using this general framework, in this Chapter we provide a theoretical background to fertility and employment choices. In Section 3.2, we introduce a micro-economic model of decision-making. Although its application to explaining fertility behaviours has been widely criticised, it is probably one of the most advanced models of rational choice. In the following sections of the chapter, we introduce certain modifications to the model to make up for its shortcomings. Within this theoretical framework, in Section 3.3 we discuss the impact of the context on fertility and women’s labour supply decisions. Contextual opportunities and restrictions comprise only one determinant of human choice, however, and the other one is preferences. The role of this factor is considered in Section 3.4 while Section 3.5 discusses some of the problems related to measurement and selection effects, which occur due to the inability of researchers to control for women’s plans and preferences. Finally, Section 3.6 summarises the theoretical framework we adopt in this volume and lists certain conditions that should be met if the interdependencies between fertility and women’s labour supply are to be understood.

3.2 Micro-Economic Approach to Fertility and Women’s Labour Supply: A Price-of-Time Model

The foundations for the micro-economic analysis of the fertility–employment nexus were laid by Mincer’s (1963) theory of labour supply and Becker’s (1965) theory of time allocation. Willis (1973) built upon their works and integrated them into a theoretical model of fertility and women’s labour supply (Ermisch, 2003: 114), known also under the label ‘price-of-time’ model. It refers mainly to women in union whose fertility outcomes are likely to be outcomes of conscious decisions and assumes that it is chiefly the mother’s time that is used for childbearing and childrearing. In this model, a couple chooses an optimal allocation of woman’s time between home production (including care) and paid work by maximising household’s utility from a certain combination of parental consumption (parents’ standard of living), number of children and their quality, subject to the household’s lifetime budget constraints. The latter depend on the amount of time a mother spends in employment, her market wage, the father’s lifetime earnings, and other non-labour income.

Due to budget constraint, the demand for children depends inversely on their cost. There are direct costs (financial expenditures) and indirect costs of children (opportunity costs of parent’s time). The latter are measured with a value of time a parent invests in childbearing and childrearing, which are the woman’s market wage as well as the value of depreciated human capital, foregone promotion prospects, etc. As a result, women’s employment has an ambiguous effect on the demand for children. On the one hand, it increases household income, which encourages parents to have more children (income effect). On the other hand, however, employed women may have higher opportunity costs than women who do not work (price effect).

The time a mother supplies in the labour market is also a choice variable that is jointly determined with the time she devotes to childrearing. A woman will take up a job only if her market wage exceeds the value of her time spent at home (her reservation wage). The reservation wage is expected to be higher for mothers of young children than of women with no children or mothers with older children as the necessity to provide care raises the value of mother’s time spent at home (price effect). On the other hand, however, having a(nother) child means additional financial expenditures to the household, which, as a result, may lower the reservation wage of the mother (income effect). In general, it is expected that the income effect starts to outweigh the price effect with a rise in child’s age, making a mother enter the labour market (e.g., Leibowitz et al., 1992; Joesch, 1994; Rønsen & Sundström, 2002).

The price-of-time model illustrates a complexity of the interrelationship between fertility and women’s employment. It is evident that these two activities compete with each other due to restricted time and financial resources of a mother. Nevertheless, the theory does not presuppose that childbearing and employment exclude each other. Whether a woman will stay at home with children, work in the market and remain childless, or combine employment with childrearing depends on a complex interplay between the income effect and the price effect. There are two problems with that approach, however. First, the price-of-time model relies on absolute income, presupposing that it is taken into account in fertility and employment decisions. This may not be sufficient since people differ in their preferences as regards living standards – whereas for some the certain income may be fully satisfactory, for the others it may fall well below their material aspirations. This implies that disregarding the material aspirations may lead to biased conclusions on the interrelationship between women’s employment and childbearing. Second, neither does the model consider women’s preferences as regards the number and timing of children nor as regards her involvement in the labour market. On the contrary, it assumes that these preferences are identically distributed across women. This may not necessarily be true. We will come back to these issues in Section 3.4.

3.3 Macro-Contextual Opportunities and Restrictions

Although the mechanism of decision-making, as described in the previous section, is the same for each individual, its outcomes might be different depending on the macro-context an individual operates in. The macro-context is a multi-dimensional structure that provides individuals with information on opportunities and restrictions. This information defines the choice options people face. These macro-contextual opportunities and restrictions affect the magnitude of the price and income effects, which results in diverse fertility and employment decisions under different contextual settings.

So far, researchers have concentrated mainly on the variation in the price effect across various macro-contexts, referring to two dimensions of the country-context distinguished by Liefbroer and Corijn (1999): cultural and structural. The former refers to ‘broad ideologies, values, and norms concerning the role of women in the society’ (Liefbroer & Corijn, 1999: 52) that define who should take care of children and who should work (see also Muszyńska, 2004, 2007). By contrast, the structural dimension relates to ‘societal opportunities and restrictions [imposed] on the roles of women’ (Liefbroer & Corijn, 1999: 52). It encompasses all institutional and structural arrangements that facilitate or hinder mothers’ employment, i.e., family policies and labour market structures. The impact of family policies on fertility and women’s employment decisions has been widely discussed in the literature. The role of labour market structures, except for an incidence of part-time contracts, was for a long time less recognised. This situation started to change only recently as it has been noticed that strong barriers to employment entry in Southern Europe may constitute an important impediment to fertility and women’s work (Adserá, 2004, 2005; Aaberge et al., 2005: 131–135). In order to turn more attention to labour market structures and disentangle their potential effects from the effects of family policies, we propose to analyse these two dimensions of the macro-context separately. Therefore, for the purpose of this volume we will speak of the structural dimension of the macro-context when referring to labour market structures, while the institutional dimension will relate to family policies.

If the price effect refers to the conflict between fertility and women’s paid work, pre-supposing that the two activities exclude each other, the income effect operates in the opposite direction. Hence, even in traditional societies, where mothers’ work in the market is socially not accepted, institutionally unsupported, and where work arrangements are rigid, women may participate in the labour force to a large extent; paid work may be perceived as an important condition for family formation for economic reasons. This is likely to happen in countries where the husband’s income is not sufficient for maintaining the family or meeting the couple’s aspirations regarding living standards, i.e., in the less affluent societies.

Against this background, four dimensions of the context are considered to be highly relevant for fertility and employment decisions: institutional (family policies), structural (labour market structures), cultural (values and norms for women’s social roles), and economic (living standards). The role of each of these dimensions of the macro-context in shaping fertility and employment decisions is briefly discussed below.

3.3.1 Family Policies

In this study, we use the term ‘family policies’ when referring to policies targeted at parenthood. These policies govern the rules of the absence from the labour market during the childcare period and define the access to social benefits and services related to parenthood. By offering women certain conditions for work and family reconciliation, they define the intensity of the institutional incompatibilities between the two activities. We now discuss potential impacts of the three most important components of family policies: childcare services, maternity and parental leave policies, and income transfers. They constitute the core of welfare policies targeted at childbearing and childrearing (Neyer, 2003). Wherever it is possible, we complement our discussion with the empirical findings on the impact of the policies on fertility and women’s labour supply (or employment). We mainly quote studies based on micro-level data.

3.3.1.1 Childcare Services

Childcare services can be offered by the state, the market, employers, or non-profit institutions (Neyer, 2003). However, what matters mostly for childbearing and employment decisions are childcare supply (in terms of number of places as well as opening hours), childcare costs, and childcare quality. Generally, using private childcare requires financial expenditures. Nevertheless, in some societies parents can be granted childcare subsidies to cover at least part of the childcare costs. Following the price-of-time model one should expect that widely available low-cost and high-quality care increases fertility and women’s labour supply. This is due to two reasons. First, a better access to high quality childcare reduces the opportunity costs of parenting and of women’s work. Second, the decline in childcare costs and an improvement in quality of public care lead to a reduction in mother’s reservation wage.

Generally, the positive impact of childcare on women’s labour supply is widely documented in the literature (e.g., Stolzberg & Waite, 1984; Blau & Robins, 1989; Connelly, 1992; Gustafsson & Stafford, 1992; Leibowitz et al. 1992; Michalopoulos et al. 1992; Ribar, 1992; Kimmel, 1995; Powell, 1998; Del Boca, 2002; Rønsen & Sundström, 2002). The effect of childcare on fertility is, however, more mixed. A positive, albeit weak, impact is found for Italy (Del Boca, 2002), USA (Blau & Robins, 1989) and for Norway (Rindfuss et al., 2010). Similar results were obtained by Aaberge et al. (2005: 150) and Del Boca et al. (2009) on the pooled data for several EU-15 member states. On the other hand, studies for West Germany (Hank & Kreyenfeld, 2003) and Sweden (Andersson et al., 2004) yield insignificant coefficients. Finally, Rønsen (2004) found a negative impact of day care coverage on childbearing in Finland and Norway. Gauthier (2007) explained this diversity in the obtained results by cross-country differences in childcare quality and opening hours, which are hardly accounted for in statistical analyses, the heterogeneity of parents’ needs and preferences as regards childcare as well as the complex relationship between public childcare provision and other social and welfare state institutions. Other researchers attributed the inconsistency of empirical findings to the inability to account for endogeneity of local fertility rates and childcare provision. In fact, in the analysis by Rindfuss et al. (2010) for Norway the positive effects of childcare provision on the transition rates to first, second and third births became stronger after endogeneity of fertility and childcare provision had been controlled.

3.3.1.2 Maternity and Parental Leave Policies

In contrast to childcare, which enables mothers to spend more time at work, maternity and parental leaves give parents the opportunity to withdraw temporarily from economic activity without terminating the employment contract. The main aim of maternity leave is to protect the health of pregnant women as well as mothers right after birth. In EU member states the right to 14-week protection of this kind has been statutory since 1992.Footnote 1 Additionally, parents who want to take care of a young child after maternity leave may often take advantage of a parental leave.

The impact of maternity and parental leaves on women’s employment and fertility depends on the length of the leave, its flexibility, and related benefit entitlements. Generally, well-paid but short leave provisions are considered to increase mothers’ attachment to the labour market and consequently positively affect women’s labour supply. There are several reasons for this state of affairs. First, women may be more likely to enter employment if they know they can make use of a leave in case of birth. Parental leaves may encourage women to take up a job prior to childbirth particularly effectively if parental leave entitlements depend on the work experience gathered before the decision to have a child (Baker & Milligan, 2005). Second, the leaves shield mothers from potential job loss. Third, short career breaks have little negative effect on women’s human capital; mothers returning from such a leave may be still attractive for an employer due to the firm-specific work experience they accumulated prior to the birth. Nonetheless, the longer a mother stays out of work, the greater the loss in her human capital and, consequently, the stronger the wage penalty. Therefore, long leave provisions may in fact have negative impacts on women’s labour supply and employment. In this context, flexible leave schemes, allowing for dividing the leave into several parts and combining it with part-time employment and vocational training, can be very advantageous for parents who want to take care of their children over a longer time. Empirical evidence generally supports these theoretical considerations. Studies conducted for the US, where maternity leaves were introduced only in the 1990s and where it was up to employers whether to grant the right to maternity leave to woman or not, indicate that mothers who were offered such a possibility were returning to work faster than women whose employment contract was terminated due to birth (Berger & Waldfogel, 2004; Klerman & Leibowitz, 1999; Waldfogel et al. 1999). Nevertheless, a study by Rønsen and Sundström (2002) conducted for three Nordic countries (Finland, Norway and Sweden) demonstrated that extensions of parental leaves led to a delay in the return rate to work among mothers. Extended parental leaves were also shown to have markedly negative implications for mothers’ wages (Beblo & Wolf, 2000, 2002). It is notable, however, that empirical studies may be biased by selection effects as women who are not entitled to parental leaves in the US can be more family-oriented than women who are entitled to parental provisions. Likewise, women who take advantage of longer leaves may differ in their family orientation from women who make use of shorter leaves.

Turning to the second variable of our interest, maternity and parental leave provisions are expected to increase fertility. This is because they allow mothers to withdraw temporarily from the labour market but give them the right to return to the same job at the same wage. In that respect, they lower the opportunity costs of parenting. The empirical studies indeed find some positive effects of parental leave mandates on childbearing, albeit these effects are assessed to be relatively weak (Hoem, 1993; Andersson, 1999; Andersson et al. 2006 for Sweden; Rønsen, 2004 for Finland but not for Norway; Lalive and Zweimüller, 2005 for Austria; Adserà, 2004 – cross-country study on aggregate data).

3.3.1.3 Statutory Paternity Leaves

Although parental leaves are usually available for both parents it is mainly women who make use of them. Therefore, with the aim of equalising women and men in the labour market and care some countries introduced paternity leaves or daddy quotas, reserved only for fathers. Nordic countries pioneered in launching such policies, individualising rights to parental leaves already in the 1990s and offering fathers benefits with high replacement rates (80–90% of the remuneration prior to the leave). Other European countries followed only in recent years. Among them a relatively long paternity leave (2 months) was introduced in Slovenia in 2007. In the majority of other countries paternity leaves are granted usually for maximum 2 weeks (O’Brien, 2009; Moss, 2010).

There has been very little empirical evidence on the effects of paternity rights on women’s labour supply or fertility so far as such policies are relatively new. The experience of Nordic countries suggests that fathers are highly interested in making use of their paternity rights, however. In Norway, for instance, around 85% of fathers entitled to daddy quotas, took advantage of it (Lappegard, 2008). Similarly high interest in making use of their paternity rights was observed among fathers in Sweden (Sundström & Duvander, 2002). Take-up rates depend on wage compensation during the leave – in countries where paternity benefits account for high proportion of fathers’ wages take-up rates are higher than in countries where replacement rates are low (O’Brien, 2009). Fathers in Norway and Sweden were also found to be more likely to make use of their paternity rights if their female partners were employed which might have a positive impact on women’s labour supply at the macro level (Sundström & Duvander, 2002; Lappegard, 2008). Duvander et al. (2006) showed additionally that couples in which men made use of a leave decided more quickly for another child. Other researchers found a positive correlation between the use of paternity rights and fathers’ involvement in childcare after the leave has ended (Haas & Hwang, 1999; Brandth & Kvande, 2009). The authors of these studies underline, however, that their results may be biased by selection effects as more family-oriented fathers may be more likely to make use of paternity rights, get involved in care and opt for higher number of children.

3.3.1.4 Income Transfers

There are two types of income transfers that may be directed to parents: child benefits and tax reductions. Both reduce the cost of children, so they are expected to boost fertility. This hypothesis is generally supported by empirical evidence (Ermisch, 1988 for UK; Zhang et al. 1994, Duclos et al. 2001, and Milligan, 2002 for Canada). The magnitude of the effect is, however, small. For instance, based on their analysis of 22 industrialised countries over the period 1970–1990, Gauthier and Hatzius (1997) concluded that an increase in family allowances by 25% would raise fertility by about 0.6% in the short run and by about 4% in the long run. This means an effect of 0.07 children per woman.

The impact of income transfers on employment is less straightforward. Tax reductions are targeted only at those who work. Since they positively affect the income earned by parents, one could expect them to raise women’s labour supply as well. By contrast, child benefits increase household non-labour income and as a result may reduce women’s labour supply. Most negative effects are expected from means-tested benefits. These benefits are withdrawn once family income exceeds a certain threshold. This may happen after a woman takes up a job. In such a situation, the value-added of entering paid employment is her market wage reduced by the amount of the benefit lost. A rate at which benefits are withdrawn and the wage is taxed away as a person takes up a job or increases the number of working hours is called marginal effective tax rate (METR). Apart from means-tested benefits, METRs are higher in countries with joint rather than separate taxation or countries granting special social benefits to the dependent spouse (OECD, 2003, 2004a). A fall in employment probability with a rise in METR is a well-established finding in empirical research (Hoynes & MaCurdy, 1994 for USA; Fortin et al., 2004 for Canada; Gurgand & Margolis, 2005 for France; Schneider & Uhlendorf, 2004 for Germany; Smith et al., 2003).

3.3.2 Labour Market Structures

Labour market structures are another important dimension of the context influencing fertility and labour market behaviours. Given the topic of our study, we are particularly interested in two characteristics of labour markets: the magnitude of the barriers to labour market entry and the flexibility of work arrangements allowing for a relatively smooth adaptation of working hours to family responsibilities. By influencing women’s opportunities to enter the labour market, to maintain employment, and to combine employment with family duties, they affect the intensity of the so-called structural incompatibilities between fertility and market work.

The barriers to the labour market entry are strongest in countries with highly regulated labour markets, where strict rules on hiring and firing of workers apply. Such labour markets are often characterised by a strong insider–outsider divide (OECD, 2004b, Aaberge et al., 2005: 131–132; Adserá, 2005). This means that the employed insiders are strongly protected from a job loss, while the employment opportunities for the unemployed outsiders are severely restricted. This leads to high unemployment, particularly among the youth. Another disadvantaged group is women, who are more likely to become outsiders than men, due to child-related career breaks.

High unemployment pressure leads to lower market wages, which together with low perceived chances of finding employment may discourage women from searching for a job and hence reduce their labour supply. In fact, some empirical studies showed that female labour force participation is indeed lower in regions characterised by higher unemployment (Light & Ureta, 1992; Jaumotte, 2003; Aaberge et al., 2005: 149). On the other hand, an increase in men’s unemployment may increase women’s determination to participate in the labour force. A finding consistent with this hypothesis was established by Jaumotte (2003) who found that labour force participation rates of women are higher in regions displaying higher male unemployment rates.

The effect of unemployment on fertility is ambiguous: By lowering market wages, high regional unemployment lowers the opportunity costs of children. Thus, a temporary spell of women’s unemployment can be perceived as a favourable occasion to have children. Such a relationship was, indeed, found in some empirical studies (Kravdal, 2002 for Norway; Schmitt, 2005 for Finland). However, high and persistent unemployment pressure reduces the expected welfare and increases uncertainty and pessimism about the future. As a consequence, this situation may lead to fertility postponement and finally even to lower family size. In this context, Kravdal (2002) found that short-term unemployment (up to 6 months) among women in Norway tends to facilitate entry to parenthood, while longer unemployment spells exert an opposite effect. Negative impact of the country-specific long-term unemployment on first, second, and higher birth risks was also established by Adserà (2005) on a pooled dataset for thirteen EU-15 member states. In a similar vein, high regional unemployment was found to reduce childbearing in the United States (King, 2005), Spain (Noguera et al. 2005), as well as in a joint analysis for Denmark, the Netherlands, France, Italy, and Spain (Aaberge et al., 2005: 150). Nevertheless, some studies provide also evidence of a rise in the birth risk with an increase in unemployment duration (Kreyenfeld, 2001 for Germany; Vikat, 2004 for Finland; Schmitt, 2005 for Germany and UK). This finding may be explained by deterioration of women’s human capital caused by long-term non-employment, discouragement by the job search, and low assessment of the chances for success in the labour market.

In order to reduce high unemployment, some countries (e.g., Italy, Spain, Poland) liberalised the law on temporary employment with the prospect of encouraging job creation. The underlying idea was that many of these jobs would become permanent as soon as the new labour market entrants gathered the necessary work experience. However, in countries with strong insider–outsider divide, fixed-term contracts were rarely replaced by permanent ones, and the new regulation only resulted in a high rotation of workers (Adserà, 2004). This additionally intensified the uncertainty among the youth and could have further worsened the conditions for family formation. Nevertheless, the empirical studies do not provide consistent evidence on the influence of temporary employment on fertility – whereas some studies yield negative estimates of the effect (e.g., Adserà, 2004; Baizán, 2005 for Italy and Spain; Kurz et al., 2005 for Germany; Liefbroer, 2005 for the Netherlands), some other indicate the effect to be insignificant (e.g., Baizán, 2005 for UK and Denmark; Francesconi & Golsch, 2005 for UK; Róbert & Bukodi, 2005 for Hungary).

Other features of the labour market, apart from barriers to labour market entry, which are in our area of interest are the flexible work arrangements. In recent decades, many European countries experienced a rapid growth in part-time employment. This form of employment is often perceived to facilitate women’s integration into the labour market since it allows for combining market work with parenthood. Empirical studies generally show that the availability of part-time jobs tends to increase female labour force participation (Del Boca, 2002; Jaumotte, 2003; Aaberge et al., 2005). Similarly, women working part-time are generally more likely to give birth (Liefbroer & Corijn, 1999 for the Netherlands and Flanders; Corijn, 2001 for Flanders; Baizán, 2005 for Italy, Spain and UK; King, 2005 for USA; Liefbroer, 2005 for the Netherlands). Nevertheless, some studies find no impact of part-time employment on childbearing (Kreyenfeld, 2001, 2005 for Germany; Olah, 2003 for Sweden and Hungary; Baizán, 2005 for Denmark). According to Del Boca et al. (2009) it is not only the availability of part-time jobs that matters but also their quality, assessed on the degree of job protection, hourly wages, and access to social benefits. By analysing jointly seven EU countries, the authors compared fertility behaviours of part-timers earning hourly wages equivalent to those obtained in full-time positions to fertility behaviours of part-timers receiving lower remuneration. They concluded that only well-paid part-time jobs (i.e., those that pay similar hourly wages to full-time jobs) facilitate childbearing. Poor-quality jobs do not significantly increase the birth risk. Similar conclusion was drawn by Ariza et al. (2005) who compared childbearing behaviours of women employed full-time and part-time in eleven European countries and found that part-time schedules enhance fertility only in Belgium, Germany, Italy, Ireland, and the Netherlands, but not in Denmark, France, Greece, Portugal, Spain nor the United Kingdom.

3.3.3 Social Norms for Women’s Roles

Apart from family policies and labour market structures, individual behaviour is also influenced by values and ideals for the ‘correct’ division of labour between women and men which dominate in a given society (Pfau-Effinger, 1998). By defining who should work and who should take care of children, they determine the intensity of the cultural incompatibilities between fertility and paid employment. Cultural incompatibilities are highest in the societies displaying low acceptance of mothers’ employment, particularly mothers of small children. In these societies, therefore, women’s labour supply is expected to be the lowest. Furthermore, where women are expected to stay at home with children while their earnings could contribute substantially to household income (because of low wages of men or high human capital of women), the opportunity costs of children will be large. As a result, women may eschew childbearing in order to participate in the labour force.

Empirical evidence on the impact of values and norms for working mothers on fertility and women’s employment is scarce. Out of the few researchers who addressed this topic were de Laat and Sevilla Sanz (2006). Using the 1994 International Social Survey Programme, they first constructed an egalitarian index for eleven OECD countries. The index was based on respondents’ responses to ten statements capturing the attitudes towards the division of household chores. Applying regression techniques they found that women living in countries with more egalitarian attitudes tended to have more children and were more likely to participate in the labour force. A similar finding, but only with respect to one of our variables of interest, was established by Algan and Cahuc (2006). The study was conducted for nineteen OECD countries. The attitudes towards working mothers were measured using the World Value Survey. Based on their macro-level regression model, the authors concluded that women’s employment rate was higher in countries where there was a larger percentage of persons opposing the statement ‘When jobs are scarce, men should have job priority over women’.

3.3.4 Living Standards

Living standards comprise the last dimension of the context we find relevant for studying fertility and labour market decisions. They determine the extent to which women are free to decide to have a certain number of children and to allocate their time between paid work, household duties and hobbies without being constrained by their economic position. In less affluent societies or during economic slowdown, women’s work may be perceived as an important source of household income. Hence, low living standards are expected to increase women’s labour supply. The empirical research provides findings largely consistent with this hypothesis (Colombino & Di Tommaso, 1996, Di Tommaso, 1999, Del Boca et al., 2005 for Italy; Shaw, 1994 and Del Boca et al., 2005 for UK; Hotz & Miller, 1988, Moffit, 1984, Macunovich, 1996, Taniguchi & Rosenfeld, 2002, and Budig, 2003 for USA).

The impact of living standards on fertility is less clear as parents may decide to invest their additional income in increasing the quality of the children they have instead of bearing additional children (Becker & Lewis, 1973). On the other hand, however, it is likely that couples living in poorer economic conditions fail to realise their fertility intentions due to economic constraints. If this is the case, one may argue that in poorer societies or countries experiencing economic slowdown, women’s paid work might even be perceived as a pre-condition to family formation. Consistently with this view, Macunovich (1996) showed that, even in the U.S. context, the income effect of women’s paid work on fertility tends to suppress the price effect during the times of economic slowdown, when husband’s earnings are insufficient to satisfy the couple’s material aspirations. Using similar arguments Hoem (2000) and Andersson (2000, 2002) explained the decline in Swedish fertility in the first half of the 1990s when Sweden was hit by economic recession and female unemployment increased markedly. Other researchers referred to the negative effect of economic slowdown or even crisis on fertility while explaining fertility decline in CEE countries right after the collapse of state socialism (Kotowska, 1999; Rychtarikova, 1999, 2000; Macura, 2000; Kotowska et al. 2008). In that context, by applying regression techniques to the macro-level data for 11 post-socialist countries, Macura (2000) showed that the deterioration in economic activity and decline in household income in that part of Europe were partly responsible for a decline in period fertility rates. A more in-depth individual-level analysis conducted by Billingsley (2010) showed that downward occupational mobility and a job loss experienced by Russian women in the 1990s led them to postpone transition to second birth.

3.3.5 Interplay of Contextual Dimensions

On the whole, the impacts of context on childbearing and women’s labour market behaviours are well documented. The empirical studies in this field show that under certain institutional conditions (better childcare provision, short and flexible parental leave schemes, income transfers that do not impose high METRs), structural conditions (low barriers to the labour market entry, high flexibility of work arrangements), and cultural conditions (positive attitudes towards mothers’ work) the price effect can be suppressed. This in fact means lower incompatibilities between fertility and women’s work and hence higher fertility and women’s labour supply. At the same time, living standards affect the income effect. The latter increases with a worsening of the economic situation in a country or the financial situation of a household.

Notwithstanding, some of the reviewed studies report insignificant impacts of the contextual elements on fertility and women’s labour supply; sometimes, they even provide evidence that contradicts our expectations. This does not necessarily mean a failure of the theory to predict individual behaviours. According to Neyer (2006) and Neyer and Andersson (2008), it is possible that family and labour market policies, social norms, and living standards not only influence fertility and employment, but also interact with each other. In support of this argument, Neyer and Anderson (2008) gave as an example the reform of the Swedish parental leave system introduced in the mid-1980s. The so-called speed-premium policy conditioned women’s right to preserving the parental leave benefit on giving birth to a subsequent child within a restricted period of time after a previous delivery. The introduction of the policy resulted in a substantial shortening of the birth intervals and in an increase in second and higher order birth intensities in the 1980s. In the following decade, however, when Sweden was hit by an economic recession, the total fertility rate fell dramatically although the speed-premium was still in force. This did not mean that the policy failed, but its impact was mitigated by worsening employment prospects among women. Since the level of a parental leave benefit in Sweden was tightly linked to women’s earnings prior to birth, a decrease in employment resulted in the shrinkage of the population covered by the speed-premium policy. Interestingly, there was no decline in fertility rate in Finland at that time, although the country also experienced a severe worsening of economic conditions. This phenomenon is explained by an introduction of home-care allowances paid to parents who preferred to take care of a child themselves instead of using public services (Vikat, 2004; Neyer, 2006; Neyer & Andersson, 2008). These allowances could have been used by unemployed women as a partial compensation for the lost earnings in the period of limited employment opportunities. These two examples demonstrate that policies interact with economic developments and may bring different results under different conditions. Based on this finding, Neyer (2006) concluded that a mix of certain welfare-state policies, gender relations and labour-market structures might be much more conducive to fertility and women’s labour supply than explicitly fertility-focused policies.

3.4 The Role of Preferences

Macro-contextual opportunities and restrictions are only one of the two components of decision-making, the one which was well described and operationalised within the price-of-time model. The other component are preferences. This component of human behaviour has so far been largely ignored by economists working in the field of fertility and women’s labour supply. Preferences were assumed to be given and identically distributed across women and all efforts to explain differences in fertility behaviours were made by looking into differences in the contextual opportunities and constraints. As a result, motivational structures that in sociological and psychological research occupy a central place have been collapsed into a one-dimensional concept of utility. On the one hand, this approach has been widely accepted, if not strongly recommended, in the economic community: ‘The economist has little to say about the formation of wants; this is the province of the psychologist. The economist’s task is to trace the consequences of a given set of wants’ (Friedman, 1962: 13). On the other hand, the approach received a strong criticism stemming not only from researchers representing other disciplines, but also from economists (Easterlin, 1976; Meeker, 1980; Robinson & Harbison, 1980; Bagozzi & van Loo, 1980, 1991; Siegers, 1991; Turchi, 1991; Pollak & Watkins, 1993; de Bruijn, 1999). It was argued that a failure to recognise the heterogeneity and nature of human preferences leads to wrong predictions of individual behaviour, particularly as regards fertility.

One of the few economists who recognised the role of preferences was Easterlin. In his work, he concentrated basically on material aspirations (Easterlin, 1976, 1980). His theoretical model of fertility behaviour, known as the ‘relative income’ model, presupposes that a couple’s willingness to have children depends on their outlook with regard to reaching desired living standards. The main advantage of his approach over the price-of-time model is the assertion that material aspirations are not given but formed during the adolescence by experience of certain living standards at the parental home. If the aspirations can be easily met, the couple can afford the desired number of children. By contrast, a failure to reach the demanded standard of living leads to fertility postponement. Nonetheless, Easterlin’s model failed to predict fertility behaviour in empirical tests (Ermisch, 1979; Pampel & Peters, 1995; Soja, 2005). One of the major drawbacks of this approach was its concentration on men’s role in the labour market, neglecting the economic role of women.

Given the advantages and disadvantages of these two apparently competing academic perspectives, the ‘relative income’ and the ‘price of time’, Macunovich (1996) proposed to combine them into one theoretical framework. She assumed that a couple maximises its utility from a certain combination of consumptions goods and number of children of a given quality subject to the household’s relative income. Consequently, in her theoretical model, the demand for children depends on the price of the woman’s time and the household’s material aspirations, proxied by the living standards experienced at the parental home. Macunovich applied this theoretical model to explain the changes in fertility and female labour force participation in the United States in the years 1969–1993. The couple’s material aspirations were measured at the aggregate level as the real average annual income of all families with at least one child under age 18 and with a head of household aged 45–54. The resulting empirical models had a very strong explanatory power.Footnote 2 They provided evidence that the extent to which men’s wages are sufficient to meet the financial aspirations of the partners is largely decisive for women’s fertility and labour market behaviours. A decline in male earnings relative to the desired living standards led females to resume economic activity. At the same time, the income effect of women’s wages on fertility was increasing, outweighing the price effect. On the contrary, when the income earned by men was sufficient to reach the desired standard of living, the income effect of women’s wages was reasonably lower and dominated by the price effect.

The model proposed by Macunovich constitutes an important extension of the classic price-of-time model as it relates childbearing decisions not only to the absolute income of the couple but also to the extent to which the income satisfies couple’s material aspirations. Nevertheless, the shortcoming of the proposed approach is that it focuses solely on financial needs, disregarding the fact that human behaviour can be driven by a whole variety of human motives. Apart from the safety needs (a category which encompasses material aspirations) these motives can be physiological, belongingness and love, self-esteem, and self-actualisation. According to the theory of human motivation proposed by Maslow (1943, 1954) all these needs are organised in a hierarchical order with safety needs at a higher level than physiological needs. The needs that are higher in the hierarchy begin to drive individual behaviour once the lower order needs have been satisfied to a sufficient degree. Thus, once material aspirations are met, women may strive for self-esteem and self-actualisation, and these new motives will determine their childbearing and employment decisions, which may bring results contrary to the predictions of Macunovich’s model. This might be particularly the case in the current Western societies where lower order needs have become largely met and individuals may aim at satisfying postmaterial needs (Inglehart, 1977, 1990). An increase in the importance of post-materialistic values is also to be expected in the former post-socialist countries of Central and Eastern Europe along with an improvement in living standards. At the beginning of the twenty-first century, however, they were still less widespread there than in the West (Kowalska & Wróblewska, 2008).

Against this background, Willekens (1991) developed a concept of career orientations. His underlying idea was that each person has a certain knowledge of the world and certain expectations about the sequence of events throughout life. Each person is also led by certain motives. This constitutes a basis upon which a person’s life goals are formed. A predisposition to engage in certain activities in order to reach the life goals is called a career orientation. The career orientation of an adult can be quite stable since it is partly determined genetically and to some extent formed through experience in childhood and adolescence. Nevertheless, the orientation can also undergo certain modifications as an adult acquires new knowledge and develops new perceptions. The role of career orientation is to build a hierarchy of a person’s life domains. The hierarchy is required for an individual to allocate resources in a proper way. Activities that are less important for achieving the person’s life goals receive fewer resources (time and energy). For example, if the family takes a lower place in the hierarchy than employment, family planning is subordinate to the work schedule. In this study, we use the concept of career orientation when referring to situations in which women’s basic (physiological and safety) needs are met.

Consistently with the concept of career orientation, Hakim divided women into three categories, differing in the level of family and work orientations (Hakim, 2000). Family-centred women give priority to marriage and motherhood. They tend to have big families and try to avoid employment unless they experience financial problems. By contrast, work-centred women are highly oriented to employment, invest continuously in their human capital, and have children only if these do not hamper their professional careers. The last group consists of adaptive women. Women of this type want to combine children with employment, without prioritizing either. They are over-represented in female-dominated occupations, in the public sector, and in part-time jobs. This group of women is most numerous in Hakim’s view, accounting for about 40–80% of all women, whereas the other two groups constitute around 10–30% each.

This theoretical model, known under the term Preference Theory, attracted a great deal of interest but also critics in demographic and sociological literature. In general, empirical studies found the distribution of work-family preferences to resemble the distribution proposed by Hakim (e.g. Vitali et al., 2010). Nevertheless, the causal link between preferences and behaviours was questioned and some researchers even considered the theory as simplistic, arguing that differences in employment choices cannot be simply explained by the fact that the choices are made by different types of women (Crompton & Harris, 1998; McRae, 2003). Instead, one should look into contextual factors that shape these preferences since ‘preferences may shape choices, but they do not […] determine them’ (Crompton & Harris, 1998: 131). We fully agree with these critical views that heterogeneity in preferences does not fully explain heterogeneity in behaviours. Consistently with the rational choice model we see final choices as a product of women’s preferences and contextual opportunities and restrictions. In our view, neglecting one of these components of decision-making weakens the explanatory power of models explaining women’s fertility and employment behaviours. Therefore, throughout this book we will refer to the concept of work and family orientations and will underline their importance for explaining women’s choices additionally to the importance of contextual factors.

3.5 Measurement Problems and Selection Effects

Women of certain work-family preferences and acting in the environment that offers them certain opportunities but also imposes certain restrictions will develop strategies to deal with incompatibilities between fertility and market work in order to maximise their preferences. These strategies imply complete abandonment or temporary termination of one career in favour of the other, reduced involvement in one career as well as changes in the sequence and timing of events pertaining to each career. The choice of a strategy depends on woman’s orientation towards family and paid work. For instance, women who do not plan to have children and who are oriented at work career may choose jobs with steep income profiles and promotion ladders, while women who are strongly family-centred may exit employment before the planned conception. Those who aim at combining the two careers may decide to take up a stable job which offers reasonable income security with a view to have a child in the future. They may also develop strategies that allow them to shorten career breaks between subsequent births, by either returning to work quickly after first birth or by deciding to have the second child right after the first (Ni Bhrolchain, 1986a, 1986b). All these examples tell us that women may adjust their current fertility (employment) behaviour to their future employment (fertility) plans. As a result, the observed sequence of fertility and employment outcomes does not represent a causal order. This poses a major difficulty to modelling fertility and employment behaviours if women’s plans and orientations are unobserved by researchers which is often the case (Bernhardt, 1993; Ni Bhrolchain, 1993). As a result selection effects arise and they lead to a bias of the estimates of the conflict between childbearing and market work.

There might be two types of selection effect: negative (adverse) and positive. Negative selection refers to the situation in which women give up one activity (employment or childbearing) with the prospect of getting involved in the other. For instance, work-oriented women will tend to abandon having children since they plan to be heavily involved in economic activity, whereas the family-oriented will rather withdraw from employment prior to conception with a plan to have a child. By contrast, positive selection implies that women want to combine employment and childbearing and for this reason will undertake one activity (employment or childbearing) with a plan to get involved in the other in the future. Hence, if they expect that having children before entering a professional career will impede their chances in the labour market, they will tend to find a job first and will realise their fertility intentions only after their situation in the labour market is stable enough, i.e., when they have a job they can return to after birth. Positive selection is also likely to occur if material aspirations of women are not accounted for in empirical models.

The inability to establish a causal relationship between fertility and women’s employment has been troubling researchers for the last three decades (Waite & Stolzenberg, 1976; Weller, 1977; Smith-Lovin & Tickamyer, 1978; Cramer, 1980; Lehrer & Nerlove, 1986; Klijzing, Siegers, J., Keilman, N., & Groot, 1988; Felmlee, 1993; Budig, 2003; Engelhardt et al., 2004; Schröder, 2005). Development of dynamic models, improved access to longitudinal data, and advancement in modelling techniques (including event history analysis) raised a hope of understanding the causal mechanism underlying fertility and women’s labour market decisions. Nevertheless, the difficulty still lies in the unavailability of appropriate data (Bernhardt, 1993; Ni Bhrolcháin, 1993). If we had longitudinal data fully describing women’s material aspirations, their orientation towards paid work and childbearing and any other women’s plans regarding fertility and employment, using the simple event history model, we could probably easily measure the price and income effects, net of all other intervening factors. Such data, however, are currently unavailable. Attempts should be undertaken to make up for this deficiency. Interdisciplinary research could probably help in finding the way to a better measurement of human motives, while certain questions could be introduced into questionnaires to collect the necessary data. Such data, however, should be available from panel surveys, and this takes time. For the time being, other solutions need to be implemented. Among them dynamic models in combination with advanced statistical techniques that allow for capturing the unobserved factors have become more and more acceptable and widespread solution in social science.

3.6 Theoretical Framework: Summary and Implications for the Study

The objective of this study is to explore interdependencies between fertility and women’s labour supply. For that purpose, we follow the recommendation of methodological individualism and take our empirical analyses down to the micro-level. Hence, our primary subject of investigation is women and the decisions they make. In order to operationalise the process of decision-making, we adopt the version of price-of-time model modified by Macunovich (1996). The modification lies in replacing the household’s total income against which the household’s utility is maximised with the relative income. According to this model women allocate their time between home production (including care) and market work by maximising the utility of the two activities. Their choice is constrained by the limited amount of financial resources they possess and the extent to which these resources meet their material aspirations. The final decision depends on the complex interplay between two effects: the price effect and the income effect. The former represents a conflict between childrearing and paid work. It is determined by the level of opportunity costs of parenting and hence implies that women’s involvement in the labour market lowers the demand for children. The income effect operates in the opposite direction. Disentangling the impact of the price effect from the income effect is an important step in understanding women’s fertility and employment behaviours.

The magnitude of both effects depends on the context in which fertility and employment decisions are made. Besides the micro-level context (household structure, household’s economic situation) the macro-context constitutes an important determinant of women’s choices. Four dimensions of the macro-context are considered to be particularly relevant for childbearing and employment decisions: institutional (family policies), structural (labour market structures), cultural (gender norms), and economic (living standards). The first three influence the magnitude of the price effect, whereas the latter affects the income effect. On the whole, the impacts of all contextual dimensions on childbearing and women’s labour market behaviours are well documented in the literature. The majority of the reviewed empirical findings suggest that under certain institutional (better childcare provision, short and flexible parental leave schemes, paternity leaves, income transfers that do not impose high METRs), structural (low barriers to the labour market entry, high flexibility of work arrangements), and cultural conditions (positive attitudes towards mothers’ work), the price effect can be suppressed. This gives an opportunity for an increase in fertility and women’s labour supply. The income effect increases with a worsening of the living standards. Furthermore, it is notable that family and labour market policies, social norms, and living standards not only influence fertility and labour supply, but also interact with each other. Identifying the influence of the context and its dimensions on fertility and women’s labour supply is important for understanding the interdependencies between fertility and women’s labour supply and consequently for designing effective public policies.

Apart from the context, fertility and employment decisions are also determined by women’s orientation towards family and paid work. The economic theory has so far accounted little for the variation in preferences (particularly for those referring to higher order needs), which is one of the reasons why it has been subjected to strong criticism in the recent two decades. Three issues need to be raised at that point. First, each woman has certain ideals of her involvement in paid work and the number of her children. Provided she has reached her desired family size, she has no reason to have a(nother) child, even if her income has unexpectedly increased. Second, it is doubtful that women’s behaviours are only driven by material aspirations. It has been well documented in psychological research that once these lower-order needs are satisfied, women start to strive for self-esteem and self-realisation. The impact of the higher order needs on individual behaviour might be particularly important in post-modern societies. Third, women do differ in their orientations towards family and paid work, and these orientations play an important role in assessing the magnitude of the opportunity costs. Hence, the opportunity costs of having children are larger for a work-oriented woman than for a family-oriented one even if they earn the same wages and act in the same environment. Accounting for women’s orientations towards employment and family is another important step in understanding the interdependencies between fertility and women’s work.

Unfortunately, data on material aspirations as well as family and work orientations of women are often unobservable to researchers. This constitutes a serious obstacle to empirical research on the interdependencies between fertility and women’s labour supply. The reason is that the observed sequence of events may not represent a causal order. For instance, the fact that labour force exit precedes birth does not necessarily mean that economic inactivity triggers childbearing. Given that women think prospectively and anticipate consequences of their actions, it is equally likely that the decision to have a child is made prior to the decision to terminate labour force participation. Such behaviour may be typical of family-oriented women for whom economic inactivity or professions with flat earning profiles or low advancement opportunities are most conducive to realising their fertility intentions. By contrast, work-oriented women, who decide not to have children (or to have only one child), may from the beginning of their employment careers choose professions with steep earning profiles and steep promotion ladders. Such inclination, to give up one activity (employment or childbearing) with a prospect of getting involved in the other, is called negative (adverse) selection. In contemporary advanced societies, positive selection is, however, more likely to be in force. This is due to the predominance of the adaptive women and the fact that some of the family-centred women work for pay. These women are interested in combining childbearing with paid work and hence may decide to enter the labour force before the planned conception. Altogether, this implies that women may select themselves into one activity before taking up the other and this effect is often not directly observed by a researcher. Exploring the mechanism that drives this selection effect is another important step in understanding the interdependencies between fertility and women’s labour supply.

Summing up, the theoretical framework presented in this chapter lists four conditions for understanding the interdependencies between fertility and women’s labour supply:

  1. 1.

    to disentangle the price effect from the income effect;

  2. 2.

    to control for the impacts of the context on fertility and women’s labour supply;

  3. 3.

    to control for work- and family-orientations;

  4. 4.

    if data shortcomings make it impossible to account for some of the factors listed in conditions 1–3, then selection effects might occur. Exploring these selection effects may provide valuable information for understanding the interdependencies between childbearing and women’s labour force participation.

In the following chapters, we follow these four conditions. In Chapters 4 and 5 we explore the impacts of the context on fertility and women’s economic activity. Additionally, we review the existing studies on the topic and assess them critically in the light of this theoretical framework. In Chapter 6, using the data for Poland, we try to make up for the shortcomings of the available empirical studies. As expected, we lack data on women’s material aspirations as well as their family and work orientations. As a response to this problem, we propose to use multi-process hazard model to take the selection effects into account.