Fertility, childcare and labour market : dynamics in time and space

Sammanfattning: Paper [I] focuses on the effects of time and space dynamics on the description offertility in Sweden. Fertility is an important determinant of long-term populationgrowth and labour market conditions. The influence of time dynamics inpostponing or accelerating childbearing is assessed by considering two differenteffects of earnings. Firstly, the effect within one generation is considered bycomparing a family’s current earnings with their earnings in the recent past andexpected earnings in the future. The second effect, referred to previously as theEasterlin hypothesis, is examined through the generations by comparing ahousehold’s earnings for a younger generation with the earnings of the parentalgeneration. These effects are expected to be generated by labour mobility acrossmunicipalities. The empirical evidence for the period 1985-2008 involvedestimating space and time dynamics by using a spatial first-order and serialsecond-order panel data model. By comparing different specifications, thehypothesis about a positive spatial autocorrelation of fertility is supported. Currentearnings appeared to have a negative effect on fertility rates within municipalities,and in the long-term, across them. The study makes a theoretical contributionthrough the application of stationarity conditions and evaluation of the long-termeffects in the direct, indirect and total forms of the model.Paper [II] contributes to the study of stationarity conditions for a spatial first-orderand serial second-order model in the presence of time-lagged spatial interaction areconsidered. The stationarity conditions on serial autocorrelation parameters arefound on the basis of the structural vector auto-regression form for the model. Thestationarity in time is a function of the spatial autoregressive parameters. The value of the time-lagged spatial autoregressive parameter defines the shifting of theinterval for first-order serial parameter. However, the sizes of intervals for thevalues of both serial parameters depend only on the value of the simultaneousautoregressive parameter.Paper [III] contributes to an analytical description of the spatial diffusion offertility, in particular, influenced by labour movements of people between places ofresidence and work. It is assumed that the labour market has externality on themarriage market due to commuting, which, in turn, affects fertility. A model ofspatial diffusion of fertility is based on the assumption of global and local spillovereffects. The global spillover effect, as shifts in fertility norms, is motivated by theincreasing variance of the social interactions of an individual, when the places ofwork and residence are different. One local spillover effect is in response to flowsof earnings across space. Another mechanism is related to expected changes inprobabilities to find a partner affected by differences in day and night population.The analytical model, in which the effects on fertility of the cited spillovers aredecomposed, is constructed in the paper on the base of a model of the demand forchildren, the spatial stock-flow model of a market, and a matching model with asex imbalance or spatial mismatch as the probability of matching. Three seximbalances, namely of night-, day-time population and an adjusted to the seximbalance commuters to residents are empirically tested. Empirical evidence onmunicipal Swedish data for the period 1994–2008 does not provide any strongevidence of spatial diffusion of fertility. However, there are externalities of labourmobility on fertility due to the changes of the gender structure of population.In Paper [IV] commuting is linked to fertility through demographic, social andeconomic mechanisms. Average differences in the first-birth rates of young working women are estimated by bivariate model with endogenous commuting.Empirical evidence based on administrative data (Sweden) reveals that commutingwomen have a lower probability of first birth at the age 21-28 years old and higherprobability at the age 29-32 years old. Therefore, commuting women likelypostpone first childbearing. Additional direct and spillover effects of commutingon fertility appear in income cross-municipal flows, the diffusion of fertility normsacross space and changes in the gender structure of population of fertile age. Apositive effect of relative incomes, positive social norms effect and negative sexratio effect are found significant both for commuting women and those who workin the municipality of residence. Marginal effects for commuters are greater inmagnitude.Paper [V] studies earnings and labour force participation (LFP) of native Swedesand recent immigrants in Sweden in response to the childcare reforms of 2001 and2002 using a difference-in-differences approach and register-based data for theperiod of 1995-2009. Immigrant and native Swedish mothers are distinguished inorder to study if increased accessibility to childcare might be particularly beneficialfor groups facing obstacles in entering the labour market. The results show that thereforms had a positive effect on earnings and LFP among native mothers withpreschool children. The group of immigrant mothers studied did not experienceany gain in labour market outcomes as a response to the reform.