Religious & Philosophical

Job bust, baby bust: The Spanish case* by Namkee Ahn** Pedro Mira*** DOCUMENTO DE TRABAJO PDF

Published
of 24
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Description
Job bust, baby bust: The Spanish case* by Namkee Ahn** Pedro Mira*** DOCUMENTO DE TRABAJO Abril, 1999 * Namkee Ahn is grateful for financial support from the Bank of Spain and from Spain s Ministerio
Transcript
Job bust, baby bust: The Spanish case* by Namkee Ahn** Pedro Mira*** DOCUMENTO DE TRABAJO Abril, 1999 * Namkee Ahn is grateful for financial support from the Bank of Spain and from Spain s Ministerio de Educación y Cultura, grant SEC ** FEDEA. *** CEMFI. Los Documentos de trabajo se distribuyen gratuitamente a las Universidades e Instituciones de Investigación que lo solicitan. No obstante están disponibles en texto completo a través de Internet: de Trabajo These Working Documents are distributed free of charge to University Department and other Research Centres. They are also available through Internet: de Trabajo FEDEA D.T by Namkee Ahn and Pedro Mira 1 Abstract The unemployment rate in Spain has been exceptionally high for more than two decades by now. During the same period the fertility rate dropped dramatically reaching the lowest level in the world. In this study we look for evidence of a link between the unemployment crisis and the fertility crisis in Spain. We examine the factors that affect individuals ages at marriage and childbirth, focusing on the effects of male employment status. Our results suggest that spells of non-employment have a very strong negative effect on the probability of marriage and childbearing. Part-time or temporal employment also shows negative (but smaller) effects relative to full-time or permanent employment. These effects are strongest on the age at marriage and the age at first birth, while the effects on subsequent births are considerably reduced. Our results suggest that lack of stable jobs among young men is one important factor that has forced many young people to delay their marriage and childbearing, lowering the period fertility rate in Spain to the lowest level in the world. JEL Codes: J12, J13 Key words: unemployment, age at marriage, age at first birth, birth intervals. FEDEA D.T by Namkee Ahn and Pedro Mira 2 1. Introduction During the last several decades, two of the most prominent sociodemographic trends taking place across the developed world have been the decline of fertility and the increase in female labor market participation rates. As a consequence, the relationship between fertility and female labor supply has been one of the most important research topics among social scientists of various disciplines. An example of this is in the New Home Economics initiated by the studies of Becker (1960) and Mincer (1963). By putting female time allocation decisions at the center of economic models of fertility, they explained the above mentioned secular trends in fertility and female labor supply. Increases in female wages were seen as one of the main forces driving these changes. However, the negative relationship between fertility rates and female labor market participation rates does not seem to hold any more as we show in another paper (Ahn and Mira, 1998). Using aggregate data for OECD countries we find that this relationship has become positive and significant since the late 1980s. We note that unemployment rates were increasing during the same period. We suggest that male employment status may have played an important role in the dramatic fertility decline in some low participation countries such as Spain and Italy, helping to bring about the reversal in the correlation between fertility and female employment. In this paper we explore in depth the relationship between fertility, marriage and male employment status using micro data from Spain. Traditionally, studies of family formation have focused on females, looking at female education, female wages and female labor market participation as the main determinants of marriage and fertility. In these studies men played at most a secondary role as an exogenous factor. However, historical observations as well as more recent experiences suggest the importance of male employment in explaining the fluctuations in fertility and marriage. The lack of jobs during the Great Depression and full employment during the 1950s and 1960s were closely matched by corresponding fluctuations in fertility. Southall and Gilbert (1996) find a strong negative correlation between the marriage rate and the unemployment rate during the second half of the 19th century and the early 20th century in England and Wales. 1 Several studies using time series data of the last several decades for developed countries also indicate a significant negative effect of unemployment on fertility (Ahn and Mira, 1998; Gauthier and Hatzius, 1997; Macunovich, 1996). Most recently, the experience of East Germany, Russia and other Eastern European countries also suggest the existence of strong negative correlation between unemployment and family formation (Eberstadt, 1994; Witte and Wagner, 1995). For example, in East 1 See other references in their article for more historical evidence. FEDEA D.T by Namkee Ahn and Pedro Mira 3 Germany the total fertility rate has fallen below one since reunification while the unemployment rate increased to a level close to 20 percent. Among Western European countries, the above mentioned trends in fertility and participation have been specially pronounced in Spain during the last two decades, as shown in Figure 1a and 1b. 2 The total fertility rate dropped from a level around 3 until the mid-1970s which was among the highest in Western Europe to the lowest level in the world, 1.2 children per woman since the mid-1990s. On the other hand, the female participation rate increased slowly but continuously from under 30 percent to more than 45 percent among the working age females, but much more rapidly among younger women (for example, from 30 to 70 percent among those aged 25-34). However, the most striking development in the Spanish labor market has been the evolution of the unemployment rate. The unemployment rate increased from a level below 5 percent through the mid- 1970s to around 20 percent since the mid-1980s (Figure 1c). This change is specially relevant for 2 Also, see Bover and Arellano (1995) and Ahn (1997). FEDEA D.T by Namkee Ahn and Pedro Mira 4 marriage and fertility because the burden of unemployment has fallen disproportionately on young workers. Among those aged 16-29, the unemployment rate has been around 40% during the last decade. One may point out that unemployment declined during the second half of 1980s while fertility continued its downward trend. One thing to notice is that even during this expansion period the unemployment rate was never below 16%. Therefore, over time it was becoming obvious in people s mind that high unemployment is likely to persist for a long time. This change in expectation for the future could have affected even further the decisions of family formation among young people. Another factor to account for the continued fertility decline in Spain in spite of the decreasing or stabilizing unemployment rate is the rapid increase in the proportion of workers holding a temporary contract following a change in the labor market regulations in During the late 1980s and early 1990s most job openings were under temporary contracts, which has greatly increased the proportion of temporary contract holders among young workers. The proportion of males aged with a permanent work contract has declined from 55% during the mid-1980s to less than 40% during the 1990s (Figure 1d). 3 High youth unemployment together with a rising proportion of temporary contract holders have brought enormous uncertainty regarding future careers and income as well as lower current income for many individuals and households. Our conjecture is that the two have combined to inhibit marriage and childbearing, both of which involve long-term commitments. Further insights may be obtained by examining the evolution of age-specific fertility rates in Spain, which is shown in Figure 2. Although fertility rates have fallen for all age groups, the decline was most dramatic among those aged During the last two decades, the fertility rate for this age group dropped by 70 percent. Furthermore, there is no sign of reversal in the downward trend for women aged while fertility of women in their 30s seems to have recovered slightly during the last 10 years. At a purely descriptive level, these trends seem consistent with the difficult conditions faced by the Spanish youth in the labor market. 3 The decline of permanent contract holders was even greater among younger workers. For example, among males aged this proportion fell from 28% to 12 % between 1987 and 1997. FEDEA D.T by Namkee Ahn and Pedro Mira 5 Figure 2: Age-specific Fertility Rates In this paper we estimate discrete time proportional hazard models in order to learn about the relationship between men s labor market experience and their family formation behavior in Spain. We use individual data from the Socio-demographic Survey which contains information on current and past economic and family situations of the members of Spanish households. First, we examine the determinants of the age at marriage and the age at first birth. Second, we examine the intervals from the first to the second birth and from the second to the third birth. Our emphasis is on the effects of individuals' labor market situation but we also consider the impact of family characteristics (parents education, father s line of employment) and other relevant factors (cohort, region, and characteristics of previous children in the analyses of inter-birth intervals). Our results suggest that unemployment was indeed a very important factor contributing to the delay of marriages and childbearing in Spain during the last two decades. FEDEA D.T by Namkee Ahn and Pedro Mira 6 2. Marriage and Childbearing: Theoretical Background A popular theory of marriage by Becker (1974) suggests that there are gains to marriage due to specialization in the production of household goods and the joint production of children and other marriage-specific capital, and that individuals will choose to marry if and when the utility in the married state exceeds the utility when single. The timing of marriage may be influenced not only by the direct benefits of marriage, but also by the costs of finding a suitable mate and the opportunity costs of being married. Since a marriage occurs when the decisions of two individuals agree, attractiveness of own characteristics is as important as that of potential spouses. Therefore, an individuals' labor market situation is likely to be one of the important determinants of the availability and the quality of potential spouses, and in turn, of the age at marriage. Furthermore, marriage usually entails fixed costs in terms of housing and basic household equipment. Due to this basic expenditure the timing of marriage is likely to be affected by one s savings and past employment history as well as current situation. The characteristics of the parental home may also influence the age at marriage since potential spouses may consider it as a factor that determines the desirability of the individual as a marital partner. Similar theoretical implications are obtained from dynamic optimization and search theory applied to the marriage market (Montgomery and Trussell, 1986). In a simplest search theory concept, unmarried individuals would have reservation characteristics (or reservation value) on the one hand, and on the other hand there are offered characteristics (or offered value) of available partners in the marriage market. When offered value is greater than reservation value one will accept the offer and get married. How would unemployment affect the age at marriage? When one becomes unemployed, if offered value in the market for the newly unemployed drops more than the reservation value he considers as a marriageable partner, marriage for this individual is likely to be delayed. Most empirical studies of the age at marriage use hazard models to estimate reduced form equations. Anderson et. al. (1987), using data from Malaysia, found that skilled employment of both husbands and wives delays marriage relative to unskilled employment or non-employment. Keeley (1977) found, using US data, that high wage males marry earlier than low wage males, which is what a theory based on specialization in the household would predict. However, Bergstrom and Schoeni (1996) found, also using US data, first a positive association between male income and age at first marriage under age 30, then a negative association for those who married after age 30. A negative effect of difficulties in FEDEA D.T by Namkee Ahn and Pedro Mira 7 men s career transitions on marriage probability has been shown in recent US data (Oppenheimer et al., 1997). The first New Home Economics models of fertility were formulated within a static framework in which single-period or lifetime decisions were made on the number of children (Willis, 1973). Although it is easier to obtain predictions of the effect of changes in the economic environment on fertility within static models, this is done at the cost of disregarding important aspects of childbearing decisions such as their irreversible nature, uncertainty about future wages and prices, unknown biological capacities and imperfect birth control, etc. This has lead to the development of dynamic, sequential theories of life-cycle fertility (see for a survey Hotz et al. 1997; Arroyo and Zhang, 1997). Hotz et al. (1997) distinguish four basic elements which are necessary in economic theories of fertility in a life-cycle setting: (i) models of optimal life-cycle consumption, (ii) models of life-cycle labor supply decisions, (iii) models of human capital investment and accumulation, and (iv) stochastic models of human reproduction. 4 Due to the difficulties involved in considering all these elements most empirical studies based on dynamic fertility models have employed a strategy of reduced form estimation 5. Most studies focus on the effects of female and male wages on fertility. In a wide range of models a negative effect of female wage and a positive effect of male wage are predicted and empirical results generally have conformed to the prediction. However, there are few studies which examine the impact of male employment status on childbearing. In this paper we adopt a reduced-form estimation strategy and we focus on the implications of male unemployment on the timing of childbearing. In theory, the effect of male unemployment on fertility should be similar to that of a drop in current period household earnings, with an additional impact through expected lifetime income if unemployment is expected to last. As long as children are a normal good, this drop in current and expected household income should decrease the probability of childbearing. Additional negative effects might arise if a housewife decides to enter the labor market or if a working wife delays her exit from the labor market in order to maintain household income. Empirical studies for 4 From a different perspective, Easterlin (1980), in an attempt to explain the postwar baby boom and subsequent baby bust, emphasizes the importance of intergenerational relative income across cohorts in the formation of desired standard of living and preferences for children of young adults. The main prediction is that lower real income of current generation relative to that of their parents will lead to lower fertility. Under a presence of high unemployment Easterlin s theory will argue for the importance of relative job stability in preference formation, predicting lower fertility among the generations faced with lower job stability. 5 Some studies have implemented empirical structural models. The models incorporate many simplifying assumptions in order to make the models empirically tractable. See a survey in Hotz et al. (1997) and FEDEA D.T by Namkee Ahn and Pedro Mira 8 developed countries tend to find positive effects of male income or male employment on fertility (Hotz and Miller, 1988; Heckman and Walker, 1990). 3. Estimation Method In this paper we analyze decisions on the timing of marriage and childbearing within the context of a proportional hazard model. Because our data only provides yearly information on all the events of interest, we use a discrete time hazard estimation method as in Allison (1982). 6 The hazard function at time t is assumed to take a proportional hazard form 1 where h 0 (t) is an unknown baseline hazard for the period t, z i (t) is a vector of explanatory variables, possibly time-varying, and β is the corresponding parameter vector. For the analyses of age at marriage and age at first birth, we construct person-year data for each year since the completion of schooling (from age 20 for college sample) until the time of the event occurrence (completed duration) or until the survey time (censored duration). For the analyses of interbirth duration we construct similar person-year data starting at the time at the birth interval of interest. Using retrospective information about individuals work histories, we construct individuals yearly employment status. Unfortunately, within non-employment periods we cannot further distinguish between unemployment and out of labor force states. However, considering that our working samples contain only prime-aged males, we think it is reasonable to interpret non-employment periods as periods of unemployment. 2. Covariates and Sample Selection The data are drawn from the Spanish Socio-demographic Survey (Encuesta Sociodemográfica) carried out by the Spanish Statistical Institute (INE) during the third quarter of The principal objective of the Survey was to gather information about individuals' history of family situation, residence and housing, economic activities and occupation, and education. The Survey contains information on 159,154 principal interviewees (a representative sample of the Spanish population of ages 10 and over) and their households. Arroyo and Zhang (1997). 6 See Jenkins (1997) for an easy-to-use implementation of this model in STATA. FEDEA D.T by Namkee Ahn and Pedro Mira 9 We limit our analysis to prime-aged males. We do this because our survey (as most other surveys) records labor market histories only for the principal respondents (one person for each household surveyed). Given our interest in the effect of male employment status on marriage and fertility, this limitation is overcome by taking the male respondents as our working sample. An important advantage of using male samples and focusing on the effect of male employment status is that, unlike female employment status, male employment status can be treated more safely as exogenous with respect to the decisions of marriage and childbearing. For example, recent work by Angrist and Evans (1998) shows strong evidence of endogeneity of female labor supply and for exogeneity of male labor supply in childbearing decisions among couples in the United States. In most societies the ages at marriage of male and female spouses are highly correlated. The correlation coefficient for Spain in our data is 0.97 and highly significant. Given that most childbearing occurs within stable unions (births from non-stable unions accounted for less than 5% in our data), the father s and mother s ages at birth are also highly correlated. Therefore we think the results of the analysis of the age at marriage and childbearing for males can be interpreted as applying to both sexes once we adjust for the age gaps between husbands and wives. 7 Our working sample consists of all principal male respondents aged at the time of survey. The reason why we do not include people younger than 26 is that the majority (60 percent) of Spanish males are still unmarried by this age. By excluding those over 40 we reduce recall error about life histories arising from the retrospective nature of the Survey. Since most marriages and births occur after age 20, almost all the decisions recorded in our sample correspond to the 1970 s and 80 s, including the period of rapid f
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks