INTRODUCTION

The uptake of (in)formal childcarearrangements is associated with a variety of factors related to individual andhousehold characteristics (age, household composition, education and labourmarket position of both partners) which affect the demand for (in)formalarrangements, as well as supply-side characteristics of both formalarrangements (cost, availability, and quality of local childcare arrangements)and informal arrangements (potential for informal care through presence of kin,health status of potential caregivers and educational level/labour marketposition of potential caregivers). Although adequately controlling forpopulation heterogeneity and selectivity in the uptake of formal childcare havebeen identified as major concerns when assessing the effect of family policieson fertility (Gauthier, 2007; Neyer & Andersson, 2008), the existing empirical literature on uptake of (in)formalchildcare arrangements has typically not been able to simultaneously controlfor both demand-side characteristics and supply-side characteristics of (in)formalchildcare at the neighborhood level. The empirical literature, however, showsstrong socio-economic gradients by level of education, household income andmigration background in the uptake of family policies (Ghysels & Van Lancker, 2009;Ghysels & Vercammen, 2012; Neels & De Wachter, 2010; Neels & Theunynck, 2012; Storms, 1995).Controlling for potentially selective uptake is therefore key in evaluating theeffects of family policies on fertility outcomes (Angrist & Pischke, 2009).

DESIGN & METHODS

Using linked Belgian census datafrom the 1991 and 2001 censuses in tandem with longitudinal microdata from thenational register spanning the period 2000-2006, this papers follows n=218.853 womenwho resided in the parental home at the time of the 1991 census (March 1^st,1991) and who had their first child between January 1^st 1999 andDecember 31^st 2000. As these children are aged 9 months to two yearsand 9 months at the time of the 2001 census (October 1^st, 2001) andhave typically not yet enrolled in preschool, information on uptake of (in)formalchildcare arrangements available in the 2001 census is used to draw up atypology of childcare arrangements distinguishing four types (no formal orinformal care arrangements, only informal care arrangements, only formal carearrangements versus a combination of formal and informal care arrangements).Subsequently, late entry discrete-time hazard models were estimated to gauge theeffect of care arrangements on second birth hazards between January 1^st2002 and December 31^st 2005 controlling for i) age of the firstchild (quadratic effect, time-varying), ii) age of the mother at the birth ofher first child (time-constant), iii) educational level of the mother (ISCED in5 levels), iv) migration status of the mother (Belgian, Southern EU, Other EU,Turkish/Moroccan, Other),v) labour market opportunities of the mother(predicted probability of full-time employment given age, educational level andlocal labour market conditions), v) formal childcare availability at themunicipality level (time-varying), vi) potential supply of informal childcare(self-reported health and employment status of the grandparents recorded in the2001 census).

RESULTS

Consistent with the literature theresults show that higher the odds of using formal childcare increase significantlywith the educational level of the mother (Figure 1), even when controlling forage of the mother at the birth of her first child, educational level andmigration background of the mother (Model 1), additionally controlling foravailability of childcare at the municipality level (Model 2), additionallycontrolling for employment opportunities of the mother (Model 3) andadditionally controlling for health and employment status of the grandmother (Model4). With respect to second birth hazards, access to both formal and informalchildcare strategies significantly increase the hazards of having a second child,although the effect of access to formal childcare arrangements on second birthhazards is more articulated (Figure 2). Additional sensitivity analyses (notshown) show that the effect of formal childcare uptake on second birth hazardsvaries significantly by both level of education and migration background of themother, with access to formal childcare arrangements being more important for familyformation of higher educated women and women of Belgian origin.

FIGURES

Figure 1 Uptake offormal childcare arrangements (odds-ratios) by level of education of themother, controlling for socio-demographic characteristics and employmentopportunities of the mother as well as supply of (in)formal care.

Figure 2 Effect ofchildcare strategies (odds-ratios) on second birth hazards, controlling forsocio-demographic characteristics of the mother (Model 1), controlling forsocio-demographic characteristics and employment opportunities of the mother aswell as supply of (in)formal care.