Goldstein, Joshua R.; Koulovatianos, Christos; Li, Jian; Schröder, Carsten Working Paper Evaluating how child allowances and daycare subsidies affect fertility CFS Working Paper Series, No. 568 Provided in Cooperation with: Center for Financial Studies (CFS), Goethe University Frankfurt Suggested Citation: Goldstein, Joshua R.; Koulovatianos, Christos; Li, Jian; Schröder, Carsten (2017) : Evaluating how child allowances and daycare subsidies affect fertility, CFS Working Paper Series, No. 568, Goethe University Frankfurt, Center for Financial Studies (CFS), Frankfurt a. M., https://nbn-resolving.de/urn:nbn:de:hebis:30:3-430067 This Version is available at: http://hdl.handle.net/10419/155334 Standard-Nutzungsbedingungen: Terms of use: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Documents in EconStor may be saved and copied for your Zwecken und zum Privatgebrauch gespeichert und kopiert werden. personal and scholarly purposes. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle You are not to copy documents for public or commercial Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich purposes, to exhibit the documents publicly, to make them machen, vertreiben oder anderweitig nutzen. publicly available on the internet, or to distribute or otherwise use the documents in public. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, If the documents have been made available under an Open gelten abweichend von diesen Nutzungsbedingungen die in der dort Content Licence (especially Creative Commons Licences), you genannten Lizenz gewährten Nutzungsrechte. may exercise further usage rights as specified in the indicated licence. No. 568 Joshua R. Goldstein, Christos Koulovatianos, Jian Li, Carsten Schröder Evaluating How Child Allowances and Daycare Subsidies Affect Fertility Electronic copy available at: https://ssrn.com/abstract=2918877 The CFS Working Paper Series presents ongoing research on selected topics in the fields of money, banking and finance. The papers are circulated to encourage discussion and comment. Any opinions expressed in CFS Working Papers are those of the author(s) and not of the CFS. The Center for Financial Studies, located in Goethe University Frankfurt’s House of Finance, conducts independent and internationally oriented research in important areas of Finance. It serves as a forum for dialogue between academia, policy-making institutions and the financial industry. It offers a platform for top-level fundamental research as well as applied research relevant for the financial sector in Europe. CFS is funded by the non-profit-organization Gesellschaft für Kapitalmarktforschung e.V. (GfK). Established in 1967 and closely affiliated with the University of Frankfurt, it provides a strong link between the financial community and academia. GfK members comprise major players in Germany’s financial industry. The funding institutions do not give prior review to CFS publications, nor do they necessarily share the views expressed therein. Electronic copy available at: https://ssrn.com/abstract=2918877 Evaluating How Child Allowances and Daycare Subsidies A⁄ect Fertility* Joshua R. Goldsteina Christos Koulovatianosb,c, | Jian Lib Carsten Schr(cid:246)derd,e This version: January 31, 2017 a Department of Demography, UC Berkeley b Department of Economics, U Luxembourg c Center for Financial Studies, Goethe U Frankfurt d German Institute for Economic Research (DIW) e Department of Economics, Free U Berlin * We have bene(cid:133)tted from comments and remarks from Alexander Bick, Katharina Wrohlich, participants in seminars at DIW Berlin, LISER, U Luxembourg, U Munich. Koulovatianos thanks theResearchO¢ ceofULuxembourgfor(cid:133)nancialsupport(grantnumberF2R-CRE-PDE-13KOUL). Correspondingauthor. DepartmentofEconomics,UniversityofLuxembourg,162Aavenuede | la Fa(cid:239)encerie, Campus Limpertsberg, BRA 3.05, L-1511, Luxembourg. Tel.: +352-46-66-44-6356. Fax: +352-46-66-44-6341. Email: [email protected] Evaluating How Child Allowances and Daycare Subsidies A⁄ect Fertility Joshua R. Goldstein, Christos Koulovatianos, Jian Li, and Carsten Schr(cid:246)der Abstract Wecomparethecoste⁄ectivenessoftwopronatalistpolicies: (a)childallowances; and(b) daycare subsidies. We pay special attention to estimating how intended fertility (fertility before children are born) responds to these policies. We use two evaluation tools: (i) a dynamic model on fertility, labor supply, outsourced childcare time, parental time, asset accumulation and consumption; and (ii) randomized vignette-survey policy experiments. We implement both tools in the United States and Germany, (cid:133)nding consistent evidence that daycare subsidies are more cost e⁄ective. Nevertheless, the required public expenditure to increase fertility to the replacement level might be viewed as prohibitively high. Keywords: Childcare, fertility, labor supply, vignette survey method, public policy JEL classi(cid:133)cation: J13, J18, J38, D91, C83, D10, C38 1. Introduction For most western economies, fertility is far below population replacement.1 Population ageing hampers the sustainability of publicly-provided health and pension systems. While governments provide generous policy incentives to cope with low fertility, the cost e⁄ective- ness of pronatalist policies is constantly questioned.2 As an example, in January 2016 the Australian government launched the (cid:147)nanny subsidy pilot program(cid:148)in order to shift weight from child allowances to nanny subsidies (subsidies per hour of outsourced childcare), and is currently investigating whether this alternative pronatalist policy is more e⁄ective in terms of raising fertility (Australian Government, 2016). Distinguishing e⁄ective from ine⁄ective policies prior to their implementation and avoiding the implementation of the latter, saves considerable economic resources. Here we develop methods for such an ex-ante assessment, focusing on a speci(cid:133)c comparison: which is more likely to increase intended fertility, per dollar spent: child allowances or daycare subsidies? We focus on understanding how pronatalist policies a⁄ect ex-ante intended total fertil- ity, i.e. intended fertility before children are born.3 We take into account the interactions of fertility plans with other forward-looking decisions of households, most importantly la- bor supply, consumption, savings plans, and plans for devoted parenting time vs outsourced childcaretime. Theinterplayofsuchdynamicdecisionsdeterminesthelong-rune⁄ectiveness of pronatalist policies.4 Permanent and credibly communicated pronatalist-policy reforms 1 About 50% of all countries are below the replacement ratio, with France being the only western economy with total fertility above the 2.03% replacement-ratio threshold (see CIA World Factbook, 2016). 2 Forexample,childallowancesareabout1.15%oftheGermanGDPaccordingtoGermanFederalStatistical O¢ ce Data (2016). 3 Total fertility, the total number of children a woman will bear during her lifetime, should be distinguished from crude birth rates (number of births in a given year). The latter is an imperfect measure of policy e⁄ectiveness. 4 A notable contribution stressing this interplay is Adda et al. (2016). Their model studies how intended fertility interacts with future female work intermittency after child birth. The implied opportunity costs of this interaction shed more light to the complex incentives behind female occupational and career choice. 1 may change the fertility plans of an entire young generation. Morning-after policy assess- ments (e.g., pilot studies) will not capture these changes because only the e⁄ects on crude birth rates can be observed. In order to determine how intended total fertility responds to pronatalist policies, we use and combine two tools: (a) a dynamic model and (b) randomized vignette-survey (policy) experiments. Our dynamic model simultaneously matches age trajectories of fertility, labor supply, consumption/asset-accumulationdecisions,aswellasthedivisionbetweeninternal(parental) and external (outsourced) childcare time (e.g. nannies).5 The later feature is crucial, be- cause outsourced childcare can free time so that parents may work more in the labor market. Yet, it is unclear how outsourced childcare in(cid:135)uences the potential channels through which pronatalist policies may a⁄ect intended fertility.6 To the best of our knowledge, this is the (cid:133)rst dynamic model with asset accumulation to study the age trajectory of childcare-time outsourcing. Inaddition, thedynamicnatureofthemodelallowsustoestimatethelong-run e⁄ects and partial-equilibrium costs of an established pronatalist-policy reform. Our randomized vignette-survey experiments also analyze intended total fertility, and we use this survey in order to validate our model-based policy assessments. The vignettes ask respondents to state their desired number of children and market-labor hours in a number of randomized hypothetical environments.7 These randomized environments di⁄er along two general dimensions: (a) household features: wage, partner income, and partner working 5 We analyze age trajectories from well-established household surveys for the US and Germany. 6 Addaetal. (2016)studyasimilarforward-lookinginterplay,buttheirformulationabstainsfrommodeling childcare outsourcing. Our model extends and complements their study in this respect. Bick (2016) studies publicly provided childcare, modeling the contrast between parenting time and outsourced-childcare time in a model without asset accumulation. Matching household asset-accumulation trajectories is crucial for understanding the planning of future resources available and its role in policy evaluation. In this respect, we complement the insights of Bick (2016) by also examining this dimension. 7 Our vignette survey was embedded in well-established panels: in a module of Understanding America Study (UAS, see https://uasdata.usc.edu/UAS-27) and a satellite of the German Family Panel ((cid:147)Pairfam(cid:148), see http://www.pairfam.de/en/study/satellite-projects/). 2 hours; and (b) family policies: level of child allowances and/or outsourced childcare-time subsidies.8 The data enable us to study the role of family-related policies for stated fertility and labor-supply choices. The underlying assumption for the suitability of the vignette to assesspolicyresponses,forwhichwe(cid:133)ndsupportiveevidence,isthatrespondentsunderstand the environments and provide credible information. Importantly, intended fertility according to the vignette and actual fertility have a substantially highand signi(cid:133)cant positive correlation coe¢ cient. Suchevidencecorroboratestheideathatrandomizedvignette-policyexperiments can complement model-based pronatalist policy evaluation. Our vignette survey adds con(cid:133)dence to estimates of policy responses that are implied by the model. Our model-based policy evaluation uses con(cid:133)dence intervals of household re- sponses to pronatalist policies derived using minimum-distance-bootstrap estimations. Yet, since the derived bootstrap con(cid:133)dence intervals rely on model speci(cid:133)cation, they are self- referential. The vignette-based policy evaluation does not rely on any model, with the direction and the strength of the model(cid:146)s policy responses cross-checked. Estimates from both the dynamic model and the vignette-policy experiments consis- tently indicate that outsourced childcare-time subsidies are more cost-e⁄ective than child allowances in raising total fertility. Further, they indicate that even generous pronatalist policies have small impact on intended total fertility. For example, increasing outsourced- childcare subsidies by 100 US dollars (100 Euros in Germany) per month per household with 8 See Appendix B for a detailed description of the vignette survey. A comprehensive list of vi- gnettes in social sciences and public-health studies is provided on the Web site of Gary King (http://gking.harvard.edu/vign/eg/), applied to a variety of domains, including political corruption, dis- ease risk perceptions and prevention, and spousal in(cid:133)delity, among others. In economics, examples include Kapteyn et al. (2007) who study the role of work disability in the labor market, and Koulovatianos et al. (2005, 2009, 2014) on estimating equivalence scales and the time costs of children. Krueger and Stone (2014)refertovignettesasakeynewadvancetowardtheimprovementofwell-beingevaluationsineconomic policy. For example, along these lines, Bertoni (2015) uses vignettes to link early life experiences and future well-beingevaluations. Alesinaetal. (2016)userandomizedinternetsurveystoanalyzethelinkagebetween intergenerational mobility and preferences for redistribution. 3 children would increase intended total fertility in the United States by 11% and in Germany by 21.8%.9 These (cid:133)ndings are informative in at least two ways. First, policymakers must understand that even costly pronatalist policies can have little impact on fertility. This means that strategies developing alternative pronatalist policies and on (cid:133)nancing those poli- cies must be further developed. Second, pilot schemes may be proven important in deciding about the policy direction. We believe that our proposed tools can facilitate pre-costing policy proposals before actual implementation. At least three recent studies share the policy-evaluation focus of ours. Bick (2016), also using German data, uses a dynamic model to study pronatalist policies similar to the policies investigated in this paper. His (cid:133)ndings on subsidizing childcare time are less en- couraging than ours: they are slightly e⁄ective, but too costly. The model of Bick (2016) is richer than ours regarding the potential labor-market statuses (part time employment, full-time employment, no job). Our model, on the contrary, allows for savings, tightly esti- mating asset-accumulation and other age trajectories. Further, Bick (2016) provides neither a cross-check of the model-based policy evaluation with alternative evaluation techniques (like our vignette) nor any cross-country comparisons. Adda et al. (2016) develop a dy- namic model that they estimate with German data. Adda et al. (2016) focus more on understanding the interplay between intended fertility and occupational choice of females in order to contribute to the understanding of the gender gap. A byproduct of their analysis is the evaluation of how child allowances may in(cid:135)uence fertility. Since Adda et al. (2016) do not explicitly model the substitutability between parenting time and outsourced childcare time, theirmodel is not able tocomparativelyevaluate daycare subsidies vs childallowances. Finally, Laroque and Salanie (2014) focus on pronatalist-policy evaluations in France sug- 9 The total fertility (births per woman) in the US would increase from 1.9 to 2.1, and the total fer- tility in Germany would increase from 1.4 to 1.7. We use the total fertility rate from World Bank (http://data.worldbank.org/indicator/SP.DYN.TFRT.IN?). 4 gesting a novel econometric-identi(cid:133)cation strategy. A key extension of our dynamic model and vignette survey would be to apply it to the case of France, thus allowing for the method- ological approaches to be compared. 2. Overview of data sources The two following subsections brie(cid:135)y describe the data used for matching the dynamic model and those collected in the vignette survey. All details regarding our samples and the construction of variable-speci(cid:133)c age-trajectories appear in Appendix A. 2.1 Data for model estimation Our targets for model estimation are the average age trajectories of: (a) number of children per household; (b) labor supply; (c) parental childcare time; (d) outsourced childcare time; (e)adultconsumption; (f)childconsumption; and(g)householdassets. Becauseourvignette survey randomizes on the attributes (cid:147)wage(cid:148)and (cid:147)income of his/her partner(cid:148), two more key modeling variables that our model uses as exogenous are the average age trajectories of the wage of the household head and partner income. Table 1 summarizes the US and German data sources as well as the sample years used for deriving the above age trajectories. For the United States, we use the Panel Study of Income Dynamics (PSID), the American Time Use Survey (ATUS), and the Survey of Consumer Finances (SCF), while for Germany we use the German Socio-economic Panel (SOEP), the German Family Panel (Pairfam), and the German subset of the Eurosystem HouseholdFinanceandConsumptionSurvey(HFCS).AppendixAprovidesthedetailsabout the construction of the corresponding age trajectories 5