IRPP Investing in Our Children

choicesVol. 14, no. 12, October 2008 ISSN 0711-0677 www.irpp.org New Evidence about Child Care in Use Patterns, Affordability and Quality

Gordon Cleveland, Barry Forer, Douglas Hyatt, Christa Japel and Michael Krashinsky Gordon Cleveland is senior lecturer in economics at the Department of Management, University of Scarborough. Since the 1980s, he has researched, written and published on a wide ounded in 1972, the Institute for Research on range of economic and policy issues related to Public Policy is an independent, national, early childhood education and care. nonprofit organization. F Barry Forer is a Ph.D. candidate in measurement, IRPP seeks to improve public policy in Canada by evaluation and research methodology at the University of . Since 1988, he generating research, providing insight and sparking has been working as a child care consultant debate that will contribute to the public policy specializing in quantitative analysis. decision-making process and strengthen the quality of the public policy decisions made by Canadian Douglas Hyatt is a professor of business economics at the Joseph L. Rotman School of Management, governments, citizens, institutions and organizations. University of Toronto. He has over 20 years of research experience in child care and related issues. IRPP's independence is assured by an endowment fund established in the early 1970s. Christa Japel is a professor in the Faculty of Education, Department of Special Education and Training, at the Université du Québec à Montréal. She is also a member of the team leading the Longitudinal Study of Child Development ondé en 1972, l’Institut de recherche en (QLSCD). politiques publiques (IRPP) est un organisme Michael Krashinsky is a professor of economics F canadien, indépendant et sans but lucratif. at the University of Toronto and chair of the Department of Management at the University of L’IRPP cherche à améliorer les politiques publiques Toronto, Scarborough. He has written five books canadiennes en encourageant la recherche, en mettant and numerous articles about child care policy. de l’avant de nouvelles perspectives et en suscitant des * * * * * débats qui contribueront au processus décisionnel en matière de politiques publiques et qui rehausseront la This publication was produced under the qualité des décisions que prennent les gouvernements, direction of Sarah Fortin, Research Director, IRPP. les citoyens, les institutions et les organismes The manuscript was copy-edited by Jane Broderick, proofreading was by Zofia Laubitz, canadiens. production was by Chantal Létourneau, art direction was by Schumacher Design and L’indépendance de l’IRPP est assurée par un fonds de printing was by AGL Graphiques. dotation établi au début des années 1970. Copyright belongs to IRPP. To order or request permission to reprint, contact:

IRPP 1470 Peel Street, Suite 200 , Quebec H3A 1T1 Telephone: 514-985-2461 Fax: 514-985-2559 E-mail: [email protected]

All IRPP Choices and IRPP Policy Matters are available for download at www.irpp.org

To cite this document: Cleveland, Gordon, Barry Forer, Douglas Hyatt, Christa Japel, and Michael Krashinsky. 2008. “New Evidence about Child Care in Canada: Use The opinions expressed in this paper are those of the authors and Patterns, Affordability and Quality.” IRPP do not necessarily reflect the views of IRPP or its Board of Directors. Choices 14 (12). Investing in our Children/Investir dans nos enfants Research Director / Directrice de recherche Sarah Fortin

his research program examines issues related to family policy from the perspective of lifetime T investment in human capital based on in-depth empirical and analytical evidence of the strengths and weaknesses of current policies as well as evi- dence supporting alternative strategies. The IRPP's research in this area focuses on recent developments across the country in policies that are geared toward children.

e programme de recherche examine les poli- tiques publiques familiales selon une pers- C pective d'investissement à long terme dans le capital humain et sur la base d'études empiriques et analytiques des forces et faiblesses de nos politiques actuelles, et explore des stratégies de rechange. Il met l'accent sur les récents choix des gouvernements fédéral et provinciaux en matière de politiques des- tinées à l'enfance.

Acknowledgements Contents

2 Introduction: Policy Options and History e would like to thank the two anonymous referees, and most especially Sarah Fortin 5 Use Patterns in Nonparental and Parental Child Care W of the IRPP, for working with us to improve exposition of the ideas in this paper. 15 Spending Patterns in Child Care: The Issue of Affordability 22 Quality in Nonparental Child Care 27 Conclusions and Policy Directions 31 Appendix A: Misinterpretations of Child Care Data from the NLSCY 35 Notes 38 References IRPP Choices, Vol. 14, no. 12, October 2008 and Douglas Hyatt,ChristaJapel Gordon Cleveland,BarryForer, and Quality Use Patterns, Affordability Child CareinCanada New Evidenceabout Michael Krashinsky I History Introduction: Policy Optionsand by academicsor advocacyorganizations,or from few questions on childcare,orfromone-off surveys surveys designedforotherpurposes butcontaininga child care.Ourinformationon childcarecomesfrom patterns, pricesandexpenditures, costs,andqualityof surveys dedicatedtocollecting informationontheuse Surprisingly, forinstance,Canada hasnostatistical erated knowledgebaseonchildcarehasbeenmodest. tion-related alternatives. lifetime dividendsthanmanytraining-oreduca- that investmentintheearlyyearscanpaylarger ment inyoungchildren,andbyeconomicfindings on theinfluenceofenvironmentsbraindevelop- early educationonchildren,byemergingresearch longitudinal researchfindingsontheeffectsof opinion hasbeenbuttressedbymuch-discussed have beneficialeffectsonchilddevelopment.This childhood educationandcare(ECEC)servicescan been increasingrecognitionthatgood-qualityearly shortages oflicensedchildcareservices. for alltypesofchildcareandhasledtopersistent years andmorehasproducedhighlevelsofdemand workforce participationofmothersoverthelast40 ty inworkforceaccess.Therapidincreasethe need forchildcarepoliciesthatpromotegenderequi- in Employment,manyCanadianshaveexpresseda Judge RosalieAbella’sRoyalCommissiononEquality Royal CommissionontheStatusofWomen, through public interest.Fromtherecommendationsof Despite allofthispublicinterest,theCanadian-gen- Beyond workforceandgenderissues,therehas care. explosion n thelast25yearsinCanada,therehasbeenan 2 There aremanyreasonsforthisintense 1 2 of policyinterestinnonparentalchild A Glimpse Ahead Box 1 We will show that the use of nonparental care by fami- Terminology lies with preschool children is now commonplace, rather than unusual. A large majority of Canadian children In this paper, we use the term child care very broadly to receive nonparental care regularly, just as a large majori- include all types of parental or nonparental care, ty of Canadian mothers of young children are employed including kindergarten. Inclusion of school-based either full-time or part-time over the course of the year. kindergarten services in the term “child care” is unusu- Our research finds that the number of children using al, but we argue in the paper that these services are nonparental child care is much greater than estimated in often close substitutes, and should be considered the recent literature. Further, close to half of all pre- together. We frequently distinguish between non- school children of employed mothers in Canada primari- parental and parental types of child care, and between ly attend ECEC services (i.e., regulated types of care). regulated and unregulated types of nonparental child Our results suggest that families cannot afford the care. Regulated types of child care include licensed high costs of good-quality nonparental child care. centre-based child care services, regulated family home Average annual child care spending by employed fami- child care (provided by an unrelated person in their lies is about $4,500, which is a burden for these fami- own home), nursery schools, prekindergartens and lies but still not enough to cover good-quality kindergartens. Nonparental child care includes regulat- regulated services unless the cost is subsidized. Some ed types of care and all unregulated types of care families are able to provide all their own child care (by except care provided by a parent (e.g., care by a rela- off-shifting between the two parents or by having the tive, neighbourhood sitter or in-home caregiver). In mother work at home). However, for most families this this paper, early childhood education and care (ECEC) is is too costly, in terms of lost income, to be a viable a synonym for regulated child care (i.e., including option. The mother’s earning power appears to be a kindergartens). The main focus of this paper, because of major constraint on the amount that families are will- data availability, is the types of care used by families in ing to spend on child care. Those families who pur- which the mother is employed or studying. chase child care currently spend, on average, over 20 percent of the mother’s earnings on these services. administrative data collected by municipal, provincial Since the purchase of child care makes these earnings or territorial governments. possible, apparently the mother’s earnings potential The team of academics who worked on this paper limits the amount that families are willing to spend. has unearthed new evidence from these surveys: about The most important finding, given the widespread what types of family uses what types of child care, use of nonparental care, is that the quality of child care New Evidence about Child Care in Canada, by Gordon Cleveland et al. about how much Canadian families spend on child in Canada is too often inadequate to support child care, and about the quality of the child care they use. development. Among the various types of nonparental In the body of the paper, we examine evidence care, nonprofit centre-based regulated care currently from a number of sources: several cycles of the provides the best experiences for children,3 followed by National Longitudinal Survey of Children and Youth regulated family home child care, for-profit centre- (NLSCY), a project of Statistics Canada and Human based regulated care and, lastly, unregulated care. Resources and Skills Development Canada; Statistics In the remainder of this introduction, we outline dif- Canada’s comprehensive (but never repeated) ferent views in the body politic on appropriate early Canadian National Child Care Survey (CNCCS) of childhood policy, the historical policy framework for 1988; Statistics Canada’s annual Survey of Household child care in Canada and a review of recent policy Spending (SHS); the Longitudinal Study of Child developments. In the body of the paper, we review evi- Development in Quebec (begun in 1998 and known by dence and analyze child care use patterns, spending its French acronym, ÉLDEQ, for Étude longitudinale patterns and quality levels in Canada. In a concluding du développement des enfants du Québec); the Quebec section, we draw together the main patterns of this evi- Survey on the Quality of Educational Daycare (con- dence and formulate some recommendations for future ducted in 2003 and known as Grandir en Qualité); and child care policy in Canada. a 1998 survey of child care facilities and quality in six provinces and one territory, known as You Bet I Care!

3 IRPP Choices, Vol. 14, no. 12, October 2008 This viewleads toapolicyrecommendation forasub- financially possible foroneparenttostay at home. with youngchildren,government shouldmake it care, atleastforpreschoolers, andthat,forfamilies and furthermorethatparents preferparent-provided inevitably betterthananyform ofnonparentalcare, subsidies butwouldmake theirownarrangements. income; manyfamilieswouldbeineligibleforECEC incomes wouldreceiveapartialECECsubsidy, basedon developmentally oriented.Familieswithmiddle parental employmentaswelltoprovidecarethatis ECEC targetedtolow-incomefamilies,encourage children. Theyfavourgeneroussubsidiesfortheuseof arrangements madebyparentsofmoreadvantaged ing) butisneitherbetternorworsethanthealternative (whose homelifemaynotbedevelopmentallystimulat- positive effectsonchildrenfromlow-incomefamilies existing ECECservicesandreducewaitinglists. the firstprioritieswouldbetoimprovequalityof larger portionoftheemployedpopulation.InQuebec ble maternity/parentalleaveandbenefitscoveringa charge), togetherwithslightlylongerandmoreflexi- for additionalhoursbeyondacorethatisfreeof of age(userscouldpaysomefeeforservices,perhaps vide affordableservicesforchildrentwotofiveyears first priorityinCanada,outsideofQuebec,istopro- to women’semployment.According tothisview, the cost ofECECcanplayaroleinremovingthebarriers funding. Further, theybelievethatloweringthenet available toallchildrenthroughuniversalpublic arrangements andthatthislevelofqualityshouldbe more beneficialforchilddevelopmentthaninformal including preschoolsandkindergartens,ismuch parental childcare,thatgood-quality believe, inthecontextofalreadywidespreadnon- divergent, positionsonwhatshouldbedone.Some holdpassionate,perhapsirreconcilably early educationandpreschoolarecontroversial; in thisnewenvironment. cial andprogramsupportforfamiliesraisingchildren believe thatgovernmentsneedtostepuptheirfinan- children havehadtochangeaswell.ManyCanadians parents, theestablishedwaysofcaringforpreschool this isadramaticchangefromearliergenerationsof now regularlyemployedinthepaidlabourforce.As A largemajorityofparentswithyoungchildrenare Canadians Policy viewsaboutchildcareamong Some othersbelievethatcarebyparentsis Other peoplebelievethatregulatedchildcarehas The publicpolicyissuessurroundingchildcare, alternative. this paperwewillnotbediscussingfinalpolicy to whetherfamiliesdoornothavechildren.In this view, publicpolicyshouldbeneutralwithrespect not befavouredoverotherconsumptionchoices.In are aconsumptionorlifestylechoice,onethatshould including carebyaparentorrelative. with vouchersredeemableforapprovedtypesofcare, are actuallyspentonchildcare)avoucherscheme Tax Benefit,or(inordertoensurethatpublicfunds family allowance,oramoregenerousCanadaChild status. Supportcouldtake theformofaneven-handed parental choicesofchildcaretypeandemployment financial supportthatinnowaydiscriminatesamong fundamental characteristicofpublicpolicyshouldbe needs oftheirchildren.According tothisview, the the qualityofchildcareofferedandparticular dren, andthatparentsmake thebestjudgesregarding sively parentalcare)areappropriatefordifferentchil- (including regulatedandunregulatedcareexclu- for thefirstthreeyearsofachild’slife. and benefitarrangementsthatprovideincomesupport other familytypes,and/ormaternity/parentalleave higher netbenefitsforsingle-earnerfamiliesthan expanded UniversalChildCareBenefit[UCCB])with home parents,oragenerousfamilyallowance(like an stantial childcare/homecareallowanceforstay-at- CCED allowsthe lowestincomeearnerina familyto Child CareExpense Deduction(CCED;since 1972).The exist inalljurisdictionsexcept Quebec. care subsidiestargetedatlow-income familiesstill ing, waseliminatedin1996, provincial/territorialchild Canada AssistancePlan,whichcost-sharedthisfund- guidelines) andvariedacrossthecountry. Althoughthe lished attheprovincial/territoriallevel(withinfederal or injobtraining,aswellbeingpoor)wereestab- eligibility rules(generally, parents hadtobeemployed taged oratrisk.Theincomelimitsandothersubsidy lies orchildrenotherwiseconsideredtobedisadvan- of thetotalinOntario),butonlyforlow-incomefami- ritories (withmunicipalitiesresponsiblefor20percent between thefederalgovernmentandprovinces/ter- ed childcarewasfundedona50/50cost-sharedbasis policy positions.Fromthe1960sto1990s,regulat- provincial andmunicipallevelsreflectsallofthese The historyofgovernmentchildcarepolicyatfederal, The fourpillarsofchildcarepolicy Finally, therearethosewhobelievethatchildren Others believethatdifferenttypesofchildcare The secondpillarofchildcare policyhasbeenthe 4 4 deduct the cost of nearly any form of nonparental care (but almost universal in take-up) and is provided for (up to a limit of $7,000 per child under seven years of children who are five years of age (or nearly five in age) from her taxable income. The logic behind the September, when school starts). In , half-day CCED is not well appreciated (and is frequently misin- junior kindergarten is universally available for children terpreted).5 In deciding whether to be employed, the four years of age (or nearly four).7 second earner in a family with young children will There is more to child care policy than these four compare the value of her time spent at home to the pillars, of course. Each province and territory has a value of her time spent in the workforce. If one of framework of regulations for child care (staff-child these is taxed while the other is not, the taxation will ratios by child’s age, group sizes, teacher qualifications, distort the employment decision inefficiently, in a way health and safety) and a staff that licenses centres and that discriminates against employment. Even though monitors compliance with regulations. In addition, time at home does produce services of value to the Quebec (as discussed below), and some other family, the value of this “home income” is impossible provinces provide direct operating funding for centres to calculate and is not subject to taxation. To reduce and family homes, to lower the cost borne by parents the resulting tax discrimination against employed and to enhance the quality of services offered. mothers, the CCED shelters child care costs from taxa- tion. These costs are already a barrier to mothers’ The last 10 years of child care policy workforce participation but would be more so if they Four factors have combined to put pressure on the had to come out of after-tax income (i.e., the mother existing “four pillars” framework for early childhood would be paying tax on the income earned to pay for policy. First, there has been a widespread recognition child care, and then paying the child care costs). that the increasing employment of mothers has dramat- The third pillar of child care policy has been mater- ically changed family life, that this trend is continuing nity and parental leave and benefits paid through the and that the great majority of preschool children are federal Employment Insurance Fund, which is a contri- affected. Second, there is recognition of the importance butory fund financed by employers and employees. In to the economy, and to gender equity, of women’s con- effect, eligible parents are paid a proportion of their tinuing attachment to the workforce. Third, there has normal employed income to permit them to stay at been considerable research on child development and home with a newborn child for up to one year. The first the effects on children of positive and negative 15 weeks of benefits (17 weeks of leave) are reserved for parental and nonparental preschool experiences (and the mother of the newborn, but the ensuing 35 weeks the potential negative effects of inaction or lack of pol- can be split between the two parents as desired. Only icy and services). Finally, there has been concern about mothers who have been employed (not self-employed) increased inequality in Canadian society and, in partic- New Evidence about Child Care in Canada, by Gordon Cleveland et al. for at least 600 hours during the preceding year are eli- ular, the burden of child poverty on present and future gible, and the recompense for most is 55 percent of generations; increased subsidization of ECEC could earned income up to $435 per week (or about $22,000 affect both the current incomes of parents and the for 50 weeks). Quebec has recently (2005) negotiated future abilities of children. with the federal government to provide more generous As a result, the last decade has seen a flood of child benefits for Quebec parents, including some flexibility in care initiatives at the federal and provincial levels. Box the amount and timing of benefits as well as coverage 2 presents some of these initiatives in chronological for self-employed mothers (see Phipps 2006 for details). order, with some commentary on the objectives of each. Year-long maternity/parental leave allows parents to spend more time away from work caring for their very young child, but, by providing a job guarantee, simulta- Use Patterns in Nonparental and neously encourages new parents to maintain their work- Parental Child Care force attachment.6 How many families use nonparental care for The fourth pillar of child care policy is the provi- their children? sion of universal, free kindergarten for age-eligible children. Kindergarten is typically provided through he conventional wisdom is that a bare majority of the public school system, for half days in some preschool children in Canada receive nonparental provinces/territories and full school days in others. In T care. This notion has its origins in two recent most provinces/territories, kindergarten is voluntary

5 IRPP Choices, Vol. 14, no. 12, October 2008 group hasaged, additionalyoungerchildren have of childrenrangedinagefrom 0to11 years; asthe large sampleofCanadianchildren. Initially, thegroup ing dataeverytwoyears,starting in1994-95,ona patterns acrossCanada.The NLSCY hasbeencollect- currently theonlysourceof dataonchildcareuse studying only forfamilieswithbothparentsemployedor the dataaredrawncangeneratevalidestimatesofuse tions relatedtochildcareuse,thesurveyfromwhich care inCanada.Indeed,giventhephrasingofitsques- misleading pictureoftheextentnonparentalchild parental care. cent ofCanadianpreschoolchildrenreceivednon- these samesources,asrecently1994-95only42per- Richards andBrzozowski2006).Further, accordingto Baker, Gruber, andMilligan2005;LeRoy2006; This hasbecomeacceptedfactthroughrepetition(see months ofagereceivedanyformnonparentalcare. that in2002-03only54percentofallchildren6to71 reports (Bushnik2006;StatisticsCanada2005)stating tively substitutesfortheelectionpromiseofabusinesstaxcredit. per capitabasistoprovinces/territoriessupportthecreationofregulatedchildcarespaces;thistransfereffec- 2007 2007 the UCCB,taxableinhandsoflower-earningspouse,sothatitismostvaluabletosingle-earnerfamilies. 2006 Conservative government. 2006 and generousmaternityparentalbenefitarrangements,includingcoverageofself-employedparents. 2005 for thispurpose. establish anationalECECsystem,transferring$5billionoverfiveyearsfromOttawatotheprovinces/territories 2005 ed atimprovingthequality, affordabilityandaccessibilityofregulatedcare. 2003 child development. 2001 2001 1997 available for$5perdayinschools. school carewidelyavailablefor$5perdayinnonprofitchildcentresandmakingschool-agedchildren 1997 Chronology ofCanadianChildCareReforms, 1997-2007 Box 2 There isreasontobelievethatthesedatapresenta The sourceofthesedataistheNLSCY, announcement ofanannualsupplementtotheCanadaSocialTransfer of$250million,providedonanequal creation ofataxdeduction($2,000perchild)fortax-payingfamilieswithchildunder18yearsage. federal/provincial/territorial EarlyChildDevelopmentagreement,providingfundsforprojectstargetedat intergovernmental agreementtopromoteaNationalChildren’s Agenda. extension ofmaternity/parentalbenefitsto50weeks. provision of$100 permonthtofamilieswithchildrenundersixyearsofage,intheformafamily allowance, cancellation (asofApril2007) offederal/provincial/territorialchildcarefundingagreements bytheincoming negotiated agreementbetweenthefederalgovernmentandQuebectoallowprovidemoreflexible federal/provincial/territorial MultilateralAgreementonEarlyLearningandChildCare,providingfundsdirect- Quebec familypolicyreforms,includingmakingkindergartenfull-dayforfive-year-oldchildren,pre- individual agreementsbetweenthefederalgovernmentandprovinces/territories,designedto 8 and notforallCanadianfamilies. 9 which is us toseebroad policy-relevantpatternsinthe data.The Reducing thenumberoftypes ofcareinthiswayallows children. ly representativeofCanada’spopulationyoung been recruitedtomake thedatamorecross-sectional- attended bychildrenundersixyearsofage the bestcurrentlyavailable. picture ofchildcareusepatterns.Still,thedataare For thesereasons,theNLSCY providesanincomplete families inwhichparentsareemployedorstudying. tions aboutchildcareusearedirectedonlyatthose one ofitsmanysubjectsinvestigation.Itsques- children’s development,sothatchildcareuseisonly any formofnonparentalcare). care used(inotherwords,if thechilddoesnotattend this categoryonlyifparental careisthesoletypeof case ofparentalcare,however, achild isclassifiedin largest numberofhoursinthereferenceweek.In to thetypeofnonparentalchildcareusedfor are fromCycle6(2004-05).Theterm“primary”refers lies withanemployedorstudyingmother. Thedata In thesetables,weconsiderfour typesofchildcare. Tables 1and2listtheprimarytypesofcare 10 The surveyassessesdifferentaspectsof 6 12 11 in fami- four types are: regulated child care, unregulated care by nursery schools and preschools to make up the regu- a nonrelative, unregulated care by a relative, and care lated care category). This increases the amount of by a parent or immediate family member.13 exclusively parent/family care (i.e., for some families Many parents use kindergarten services as a form of kindergarten is the only type of nonparental care child care for children three, four or five years of age used). If we employ these definitions, over one third (depending on eligibility rules in their province), rather of all preschool children with employed/studying than or in addition to nursery school, pre-kindergarten, mothers attend regulated care services, while just over preschool or centre-based child care (all of which might 30 percent are cared for exclusively by one of their be considered reasonably close substitutes). In table 1, parents and 17 percent, in each case, are cared for by we include kindergarten as a form of regulated child a nonrelative and by a relative. care. If more hours are spent in kindergarten14 than in The answer to our question about how many chil- any other form of care, then regulated child care dren are in particular types of nonparental care becomes the primary type of care used. depends, therefore, on whether kindergarten is consid- Table 1 shows that nearly half of all preschool ered a form of child care. Based on parent behaviour, children of employed or studying mothers attend reg- we believe kindergarten should be included. ulated care as their primary type of child care.15 Just We can look at parent behaviour when Quebec made over 20 percent do not attend any type of non- kindergarten full-day rather than part-day in 1997, parental care (i.e., they are cared for exclusively by a when New Brunswick initiated full-day kindergarten in parent or family member). In 15 percent of cases the the early 1990s and when Ontario opened junior primary type of care is care by a nonrelative and in kindergarten services across the province. In each of another 15 percent it is care by a relative. these cases, kindergarten use increased and other non- One might wonder how this picture of child care parental care (especially regulated child care) de- use would change if the role of kindergarten in pro- creased. When parents see these services as close viding care were ignored. Table 2 omits kinder- substitutes, both for child development and as a form garten as a form of regulated care (leaving child of substitute care while they work or study, it is appro- care centres, regulated family-home child care, priate for researchers and policy-makers to include kindergarten as a type of child care in their analyses of Table 1 use patterns and family decision-making. Primary Type of Child Care among Preschool Moreover, kindergarten is included as a form of Children, Including Kindergarten, Families with ECEC in the Organization for Economic Co-operation Mother Employed or Studying, Canada, 2004-05 and Development reports for nearly all countries (see

Type of child care n % of children OECD 2001, 2007) and is included in regular reports New Evidence about Child Care in Canada, by Gordon Cleveland et al.

Regulated 619,200 48.8 on “Early Childhood Education and Care” by the Unregulated, not relative 186,300 14.7 Childcare Resource and Research Unit of the Unregulated, relative 187,100 14.8 University of Toronto (Friendly and Beach 2005a; Care by parent/family 275,700 21.7 Total 1,268,300 100.0 Friendly et al. 2007). In Canada, most kindergartens are located in schools but attendance is voluntary; Source: Statistics Canada, National Longitudinal Survey of Children and Youth (NLSCY), 2004-5 (cycle 6). except in New Brunswick, they are not part of the Note: Data in this table and the others from the NLSCY typically include children compulsory school system. Kindergartens generally aged 6-71 months (pre-grade 1). offer a play-based early learning experience that is similar to that provided in regulated child care, with Table 2 less advantageous staff-child ratios but with more Primary Type of Care among Preschool Children, Excluding Kindergarten, Families with Mother years of teacher training (although generally with no Employed or Studying, Canada, 2004-05 requirement for special training in early childhood development). Type of child care n % of children If we include kindergarten as a form of nonparental Regulated 435,600 34.3 child care, then 8 out of every 10 children in families with Unregulated, not relative 212,000 16.7 Unregulated, relative 215,800 17.0 an employed or studying mother received nonparental Care by parent/family 405,000 31.9 care on a regular basis. If kindergarten is not included, the Total 1,268,400 100.0 figure is 7 out of 10. In other words, a large majority of Source: Statistics Canada, NLSCY 2004-5 (cycle 6). preschool children whose mothers are employed or study-

7 IRPP Choices, Vol. 14, no. 12, October 2008 employed orstudyingmother. Third,theyreferto dren fromallfamiliesratherthanonlythosewithan ment. Second,theyincludespacesoccupiedbychil- First, theCRRUnumbersmeasurecapacity, notenrol- children inregulatedcare(notincludingkindergarten). pares toourestimate,intable2,of435,600preschool occupied bychildrenundersixyearsofage.Thiscom- it isunknownhowmanyofthesefamilyspacesare another 140,000spacesinregulatedfamilychildcare; spaces forpreschoolchildreninCanada,andabout roughly 390,000centre-based,regulatedchildcare For 2006,thesepublicationsestimatethatthereare use dataprovidedbyprovincialandterritorialofficials. (Friendly andBeach2005a;Friendlyetal.2007), which Resource andResearchUnit(CRRU)inToronto in CanadaisregularpublicationsoftheChildcare percent and25percent,respectively. garten isexcluded, thenthefiguresare50percent,25 remaining 19percentusecarebyarelative.Ifkinder- percent useunregulatedcarebyanonrelativeandthe use regulatedECECastheirprimaryformofcare,19 table 3,wefind(includingkindergarten)over62percent care. Ifwefocusonlyonnonparentalcare,asshownin regulated services.Buttables1and2includeparental and, includingkindergarten,approximately halfattend ing regularlyreceivesomeformofnonparentalcare, yeo hl aeknegre kindergarten kindergarten Type ofchildcare oa 0. 100.0 100.0 25.0 24.6 18.9 18.8 50.4 Source: NLSCY2004-5(Cycle6). Total Unregulated, relative Unregulated, notrelative Regulated 62.4 Studying, Canada,2004-05(percent) Kindergarten, FamilieswithMotherEmployedor by PreschoolChildren,IncludingandExcluding Primary Type ofNonparentalChildCareAttended Table 3 oa 0. 100.0 100.0 21.7 14.8 14.7 31.8 14.8 2004-05 Sources: NLSCY1994-5 (Cycle1)and2004-56). 31.5 48.8 Total 1994-5 Care byparent/family Unregulated, relative Unregulated, notrelative Regulated 21.9 Type ofchildcare Studying, Canada,1994-95and2004-05(percent) Children, FamilieswithMotherEmployedor Primary Type ofChildCareAttendedBy Preschool Table 4 The estimatesaredifferent,forseveralreasons. The mainalternativesourceofinformationonECEC nldn Excluding Including has fallendramaticallyinQuebec. age termsintherestofCanadaoverdecadebut Canada. Carebyarelativeisupslightlyinpercent- parent/family careinQuebecandtherestof tial declineinbothnonrelativecareandexclusively in bothQuebecandtherestofCanadasubstan- been substantialgrowthintheuseofregulatedcare 1994-95 (inparentheses),onecanseethattherehas over 40percent.Still,lookingattheusepatternsin in therestofCanadaitismerelypredominant,at of childrenattendingthistypeservices,whereas mother isnowoverwhelming,withover70 percent garten) byfamilieswithanemployedorstudying Quebec theuseofregulatedcare(includingkinder- corresponding 1994-95figuresinparentheses).In Quebec andtherestofCanadaseparately(with shows the2004-05childcareusepatternsfor unmistakable, andquantitativelylarge.Table 5 reforms inthesechangesoverthelast10 years is more thanathird. exclusive useofparent/familycarehasdroppedby nonrelative hasbeencutbymorethanhalf,andthe percentage terms.Theuseofunregulatedcarebya all children),whilecarebyarelativeisstablein 1994-95 (fromlessthanaquartertonearlyhalfof forms ofchildcarehasgrownsubstantiallysince employed orstudyingmother. Useofregulated years foralltypesofchildcareinfamilieswithan 1994-95. Table 4showsthepercentageuseinboth 05) tothoseavailablefromCycle1oftheNLSCY in We cancomparethenumbersintable1(from2004- 1994-95 and2004-05 Ten yearslater:comparingchildcareusein on theNLSCY survey. appear tobebroadlyconsistentwithourown,based 2006 ratherthan2004-05.However, theseestimates aeb aetfml 652. 3332.3 15.1 23.3 31.4 17.5 29.8 21.2 17.5 13.4 16.5 41.6 31.8 5.8 Care byparent/family 24.9 5.5 Unregulated, relative Unregulated, nonrelative Regulated 72.2 Source: StatisticsCanada, NLSCY2004-5(cycle6). Restof Canada Quebec Type ofchildcare Canada, 1994-95and2004-05(percent) Employed orStudying,QuebecandtheRestof by PreschoolChildren,FamilieswithMother Primary Type ofNonparentalChildCare Attended Table 5 The roleofQuebec’schildcareandkindergarten 8 040 949 040 1994-95 2004-05 1994-95 2004-05 Use of nonparental child care when mothers 2004 for the Vanier Institute of the Family by Professor are not currently employed or studying Reginald Bibby of the University of Lethbridge. The There are no valid current data on use of nonparental study report was subtitled “A Survey of Canadian child care by families with children in which the Hopes and Dreams,” reflecting the design of this survey mother is not employed or studying, although some of more than 2,000 adults. The survey collected data on of these families do use care outside the family (for the aspirations of Canadian families in an ideal world. stimulation of early learning experiences, for peer Survey respondents were asked to imagine that they play or for a parental break). Therefore, table 6, from had preschool children in a two-parent family with the 1988 CNCCS, is valuable in that it shines a light both parents working. They were asked to list their top on this neglected group. five choices among different types of care they could use. Bibby summarizes the results this way: “[I]n an Table 6 ideal world, the number one choice is one’s partner, Primary Type of Child Care Attended by Preschool Children Whose Mothers Are Neither Employed followed by one’s parent, then another relative. Nor Studying, Canada, 1988 Rounding out the top five? Home-based child care, fol- lowed by a child care centre” (Bibby 2004, 55). Type of child care n % To put these responses in context, we should note Mother or father 663,000 73.7 that, despite the phrasing of the question, it is not Regulated 176,100 19.6 Unregulated, nonrelative 20,900 2.3 “employed” Canadian parents of preschool children Relative 40,200 4.5 who are responding to the survey, but a group of adult Total 900,200 100.0 Canadians of all ages, the large majority of whom do Source: Statistics Canada, Canadian National Child Care Survey (CNCCS), 1988. not have preschool children currently and many of Note: Nonparental arrangements are are included if the child attends for eight or more hours per week. whom may not be employed currently. The respondents are answering a hypothetical question about what pref- Given the publicity for positive research findings erences (i.e., unrelated to financial or employment con- about the beneficial effects of preschool child care straints) they would have if they were living in a experiences, it is probable that the use of different family situation that the majority of them are not cur- forms of nonparental care in families in which the rently living in.16 mother is not employed or studying is greater now A study conducted by the Institut de la Statistique du than it was nearly 20 years ago when these data Québec (2006) provides an alternative guide to the child were collected. At that time, there was no substan- care preferences of parents. This study surveyed families tial use of nonparental care in over 70 percent of in Quebec with a child under five years of age in these families (we define substantial use as at least September 2004. The families were asked which type of New Evidence about Child Care in Canada, by Gordon Cleveland et al. eight hours per week). However, nearly 20 percent child care they would choose, based on the age of their of preschool children in these families did attend child, if they needed to use care on a regular basis. They regulated care (often nursery school or kinder- were given a number of alternatives: care provided in garten) on a regular basis and nearly 7 percent their own home, unregulated care in the caregiver’s received substantial amounts of other types of non- home, reduced-contribution care in a family home (co- parental care. ordinated by a CPE — Centre de la Petite Enfance, or nonprofit child care centre), reduced-contribution care What type of care do parents prefer? in a CPE, other type of care. For a child under one year, It has been frequently claimed (e.g., Kozhaya 2006; the majority of families chose care provided in their LeRoy 2006) that Canadian parents do not favour own home. From one year through four years, approxi- regulated child care; it is suggested that this type of mately 65 percent to 80 percent chose care in a CPE or child care is “institutional” and that parents prefer coordinated by a CPE in a family home. The survey did care that is provided by parents or relatives. This not include care by a parent or relative. claim seems to be at odds with the rapid growth in In the final analysis, surveys of preferences do not use of regulated care in both Quebec and the rest of necessarily provide trustworthy evidence about parents’ Canada over the last decade, so the issue of “parent underlying tastes; most respondents are dramatically preferences” warrants some exploration. influenced by the wording of the questions and by the The origin of the view that parents do not favour conditions (child care price and quality; family income) regulated care appears to be a study conducted in that they face while answering them. In the end, the

9 IRPP Choices, Vol. 14, no. 12, October 2008 Canada, touseregulatedcare. increased willingness,inbothQuebecandtherestof decide touse.Thelatestavailabledatasuggestan tions, willbethetypeofcarethatparentsactually best guidetoparentpreferences,undergivencondi- range offactors. main typesofchildcaretouseareinfluencedbya Quebec, weconfinethisanalysistotherestofCanada. Because ofthedominanceregulatedchildcarein mothers forpreschoolchildren(aged6-71 months). types ofcarebyfamilieswithemployedorstudying we canexaminefactorsaffectingtheuseofvarious Using thedatafromCycle6ofNLSCY (2004-05), different typesofcarefortheirpreschoolchildren. Different families,indifferentfamilysituations,use Who useswhattypeofcare? table 8;notethattheexcludes childrenofstu- that themothertypicallyworksinherpaidjob(see age five dren atagesthreeandfourlessthan3percent drops correspondingly, withabout17percentofchil- care services.Theuseofexclusively parent/familycare five, roughlyhalfandthen70 percentattendregulated exclusively byparents/family. Atagesthree,fourand numbers ofchildrenareinregulatedcareandcaredfor from work. family carearecaredforbyaparentonpaidleave proportion ofthesechildreninexclusively parent/ maternity orparentalleaveare surprising, butrememberthatmotherswhoareon other membersoftheimmediatefamily. Thismaybe is employedorstudyingarecaredforbyparents year old(i.e.,age<1)infamilieswhichthemother is theageofchild. yeo hl ae< - Allages 5 3-4 2 1 <1 Type ofchild care oa 0. 0. 0. 0. 0. 100.0 17.5 23.3 17.5 100.0 11.3 2.4 16.0 100.0 16.5 16.5 15.3 41.6 100.0 21.4 29.1 22.5 100.0 70.2 22.0 28.7 21.2 100.0 51.7 15.4 68.0 10.5 27.1 Source: StatisticsCanada, NLSCY2004-5(cycle6). 28.1 Total Parent/family Relative Nonrelative Regulated 6.1 Canada ExcludingQuebec,2004-05(percent) Type ofChildCare,IncludingKindergarten,byChild’s Age,Families with MotherEmployedorStudying, Table 7 The decisionsofparentsaboutwhichthefour A secondfactoristhenumberofhoursperweek At oneandtwoyearsofage,approximately equal The majority(68percent)ofchildrenunderone not using someformofnonparentalcare. 17 As table7suggests,onekey factor employed . Alarge Child’s age by anonrelativeriseswithlevelofeducation. university-educated mothers;useofunregulatedcare els. Useofcarebyarelativeissomewhatlowerfor care diminishesmoderatelyathighereducationlev- choice ofaplurality. Useofexclusively parent/family every levelofeducation,regulatedchildcareisthe moderately relatedtothetypeofcarechosen.At in two-parent(marriedorcommon-law)families. figures comparetonearlyonequarterofthechildren parent/family careisinthelowsingledigits;these whereas inseparated,widowedordivorcedfamilies, mothers carefortheirownchildrenwhileemployed, a never-marriedsingleparent,about15percentof to useexclusively parent/familycare.Infamilieswith families aremuchlesslikely thantwo-parentfamilies more likely tousecarebyarelative.Single-parent tive care,althoughnever-marriedsingleparentsare use ofunregulatednonrelativeandrela- Marital statushaslittleeffectonthecombinedtotal a secondparentwhocouldprovideparent/familycare. gle-parent families,butitalsoreflectstheabsenceof availability oflow-incomechildcaresubsidiestosin- the lowerincomesofsingleparentsandgreater families touseregulatedchildcare.Thismayreflect widowed ordivorced,aremorelikely thantwo-parent whether themotherhasnevermarriedorisseparated, shown intable9.Single-parentfamilies,regardlessof chasing nonparentalchildcareistoohigh. exclusively parentalcareorbecausethecostofpur- work fewerhourspreciselybecauseshepreferstouse parental careastheeffect;amothermaydecideto necessarily regardnumberofhoursasthecauseand of threevs.oneoutfive).Ofcourse,wecannot which themotherworksmorethan20hours(oneout use exclusively parent/familycarethanthosein less than20hoursperweekaremuchmorelikely to dent mothers).Familiesinwhichthemotherworks As table10 shows,themother’seducationisonly A thirdfactorismaritalstatus,whoseeffect 10 Table 8 Type of Child Care Attended by Preschool Children, Including Kindergarten, by Mother’s Usual Number of Working Hours, Canada Excluding Quebec, 2004-05 (percent)

Mother’s usual number of working hours

All employed Type of child care Under 20 20-29 30-39 40 or more mothers

Regulated 37.4 38.3 44.7 42.7 41.6 Nonrelative 14.0 20.1 18.1 18.9 18.0 Relative 16.1 17.2 18.9 17.2 17.5 Parent/family 32.5 24.4 18.2 21.2 22.8 Total 100.0 100.0 100.0 100.0 100.0

Source: Statistics Canada, NLSCY 2004-5 (cycle 6). Note: Does not include children whose mothers are students.

Table 9 Type of Child Care Attended by Preschool Children, Including Kindergarten, by Mother’s Marital Status, Families with Mother Employed or Studying, Canada Excluding Quebec, 2004-05 (percent)

Single parent, Single parent, separated, Type of child care Two-parent family never married widowed or divorced All families

Regulated 40.2 50.4 59.3 41.7 Nonrelative 18.0 12.5 17.3 17.7 Relative 17.1 22.3 17.8 17.4 Parent/family 24.7 14.9 5.61 20.4 Total 100.0 100.0 100.0 100.0

Source: Statistics Canada, NLSCY 2004-5 (cycle 6). 1 Estimate is based on too few cases to be reliable.

Table 10 Type of Child Care Attended by Preschool Children, Including Kindergarten, by Mother’s Highest Level of Education, Families with Mother Employed or Studying, Canada Excluding Quebec, 2004-05 (percent)

Some College certificate University All levels Type of child care High school post-secondary or diploma degree of education

Regulated 41.6 40.1 38.9 44.9 41.8 New Evidence about Child Care in Canada, by Gordon Cleveland et al. Nonrelative 13.1 16.9 18.0 21.0 17.7 Relative 21.5 16.4 20.0 12.8 17.5 Parent/family 23.8 26.5 23.1 21.2 23.0 Total 100.0 100.0 100.0 100.0 100.0

Source: Statistics Canada, NLSCY 2004-5 (cycle 6).

Table 11 Type of Child Care Attended by Preschool Children, Including Kindergarten, by Family Income Quintile, Families with Mother Employed or Studying, Canada Excluding Quebec, 2004-05 (percent)

Second Third Fourth Highest All levels First ($43,001- ($61,001- ($78,001- ($102,001 of family Type of child care $0-$43,000 $61,000) $78,000) $102,000) and over) income

Regulated 39.6 36.7 39.1 40.9 50.5 41.6 Nonrelative 13.8 15.8 16.3 20.6 20.8 17.5 Relative 19.6 20.0 17.8 19.1 11.8 17.5 Parent/family 26.9 27.5 26.8 19.4 17.0 23.3 Total 100.0 100.0 100.0 100.0 100.0 100.0

Source: Statistics Canada, NLSCY, 2004-5 (Cycle 6).

11 IRPP Choices, Vol. 14, no. 12, October 2008 care islesscommonatthetwohighestincomelevels. falls offatthehighestincomelevel.Parent/family relative isnotgreatlyaffectedbyfamilyincomebut care riseswithincome.Useofunregulatedbya dized andmaybeexpensive;theuseofthistype middle incomes. not lesslikely touseregulatedcarethanthosewith As aresult,thoseinthelowestincomequintileare sidies forchildrenofanyage(Friendlyetal.2007). families inCanadamaybeeligibleforchildcaresub- Ontario). Low-income(particularlysingle-parent) is freeforfive-year-olds(andfour-year-oldsin most regulatedchildcareisexpensive,kindergarten rising to50percentinthehighestquintile.Although percent inthefourlowerquintilesoffamilyincome, lies aswellhavingitsownindependenteffect. Family incomereflectsthesedifferencesbetweenfami- mother hashighereducationarelikely toearnmore. are likely tohavelowerincomes.Familiesinwhichthe likely tohavelowerincomes.Single-parentfamilies mother worksforpayfewerhoursperweekare other familycharacteristics.Familiesinwhichthe family incomein2004-05.Incomeisrelatedtomany types ofchildcareandquintilegroupingsannual .Someothermemberofthe immediatefamily(e.g.,an 4. Thechildislessthanoneyearoldandthemother 3. Themotherworksathome(probablyself-employed) 2. Themotherworkspart-time(sometimesfull-time) 1. mother isemployedorstudying. family careforapreschoolchildeventhoughthe which afamilycanendupusingexclusively parent/ spending patternsandaffordability). mother isemployed(seebelow, inthesectionon lies donotpurchasechildcareeventhoughthe the importantquestionofhowitisthatmanyfami- or studying?Thisisanintriguingquestion,linked to their preschoolchildrenwhenthemotherisemployed Which familytypesusecarebyaparentorsiblingfor Who usesparentalcare? Unregulated carebyanonrelativeisnotsubsi- Use ofregulatedchildcareisapproximately 40 Table 11 showstherelationbetweenuseofdifferent older sibling)provides carewhilethemother works. able toprovidecarewhileclassified asemployed. is onmaternity/parentalleave fromherjobsois simultaneously providecare for thechild. and herworkallowssufficientflexibilityforto works (off-shifting). and thefatherlooksafterchildwhileshe It isusefultostartbylistingthedifferentwaysin who areonmaternityorotherleave percent ofparent/familycareisprovidedbymothers employed andworkintheirownhomes).About20 while atwork(mostofthesemothersareself- of cases,themotherprovideschildcarebyherself acts asthefamilybreadwinner. Inabout30percent of cases,thefatherprovidescarewhilemother the mothertocareforchild.Inanother7percent about 40percentofcases,thefatheroff-shiftswith ent/family careisprovidedbythefather. In not showabreakdown forpossibilitynumber4. provided byasiblingistoosmalltoreport,sowedo list abovedoesnotapply. Further, theamountofcare of nonparentalcare,sopossibilitynumber5inthe In thisbreakdown, kindergartenisconsideredaform whose motherswereemployedorstudyingin1988. types ofparent/familycareprovidedtopreschoolers However, the1988CNCCSdoespermitustodoso. looks atarangeoffamilycircumstancesrelatedto with anemployedorstudyingmother. Thistable tors thataffecttheuseofparentalcarebyfamilies course, areabletoprovidecarefortheirownchild). (they arethereforeclassifiedasemployedbut,of (it asksonlyaboutformsof many childrenarecaredforineachofthesesituations (Ifkindergartenisnotconsideredaformofchild 5. oa 9,0 100.0 295,900 19.8 Source: StatisticsCanada, CNCCS1988. excluded duetosmall cellsize. to employment,itisidentifiedas“supplementary” care.Carebysiblingis 59,200 employed orstudying.To distinguishthisfromregularparentalcareunrelated 40.9 parental careformorehoursthananyother formofcarewhilethemotheris Note: Table showsnumber andpercentofchildrenwhousesomeform Total Maternity orotherleave 122,500 Mother providescareatown Care byfatherwhoisnot Care byemployedfather family care Type ofparental/ Mother EmployedorStudying,Canada,1988 Received byPreschoolChildren,Familieswith Type ofSupplementaryParental/Family Care Table 12 mlyd2,0 7.4 30.7 22,200 92,000 workplace (maybeathome) employed As table12indicates,thelargestcategoryofpar- Table 13givesusadditionalinsightintothefac- The NLSCY doesnotpermitustocalculatehow part-day inotherprovinces). Quebec, NovaScotiaandNewBrunswick;typically when thechildattendskindergarten(full-dayin care) themotherisemployedonlyforthosehours Table 12providesabreakdown ofthedifferent 12 nonparental n 18 from ajob child care). % ofchildren Table 13 Factors Related to the Use of Parental Care for Preschool Children, Families with Mother Employed or Studying, Canada, 1988

Explanatory variables Estimated coefficient Standard error Mean value in sample

On maternity leave .489** .023 .07 Works at home .426** .016 .10 Works less than 20 hours/week .028** .012 .23 Regular late-day or non-day shift .326** .016 .10 Student or works varying shifts .113** .011 .38 Predicted hourly wage -.043** .013 10.37 Square of predicted hourly wage .001** .001 115.25 Other income in the family/1000 -.001** .000 28.06 Father’s employment is flexible .050** .011 .26 Single-parent family -.210** .018 .10 Recent immigrant, English or French as mother tongue .107** .027 .03 Recent immigrant, English or French not mother tongue .011 .022 .05 Earlier immigrant, English or French as mother tongue .036 .024 .04 Earlier immigrant, English or French not mother tongue .028 .026 .03 Family lives in municipality less than 100,000 population .006 .010 .41 Number of children 0-5 in family .018* .009 1.31 Age of youngest child -.025** .012 2.56 Square of age of youngest child .004* .002 9.25 Constant .450** .073 1.00

N 6927 Adjusted R-squared .267

Notes: Mean value of parent/family care is .28. Regression is ordinary least squares (linear probability function). In this regression, nonparental child care does not include kindergarten. * statistically significant at p < .10 ** statistically significant at p < .05 Source: Statistics Canada, 1988. the choice of child care arrangements and the effects child regularly works an evening or night shift, the of each, holding the others constant. In other words, probability of using parental care increases by about it analyzes factors related to the use of parent/family .33. If the mother is a student or works varying shifts care in a regression framework. By necessity, it is week to week, parental care is also more likely New Evidence about Child Care in Canada, by Gordon Cleveland et al. based on data from the 1988 CNCCS. (increased probability of about .11). Presumably, these Many of these factors are significantly associated three work situations for the mother would be typical with the decision to use parental care. The probability ones in which a father might off-shift to provide care that any child uses primarily parental care while the evenings or nights. mother is employed or studying varies from 0.00 to The above variables may well be correlated with a 1.00; the estimated coefficients in column two of the family’s income situation; the next variables, howev- table tell us how much, on average, each factor (in er, are explicitly economic. The CNCCS has no data on column one) affects this probability. Naturally enough, the hourly wages of mothers. However, part of the being on maternity leave increases the probability by sample for the CNCCS overlaps with data from the about .49 (i.e. by nearly 50 percent) of using primarily Labour Market Activity Survey, which does have wage parental care for a child. When a mother’s usual data. Lisa Powell (1997) used this unique situation to employment is at her home, this is also strongly asso- estimate wage equations for mothers. Using her ciated with using parental care for a preschool child. results, we can predict wages for every mother in the The probability of using parental care is higher by .43 CNCCS data set. for families in which the mother works at home. Our results show that mothers with higher wages are Part-time work for a mother is also significantly less likely to use parental than nonparental care, but this associated with parental care, but the size of the effect moderates somewhat at higher wage levels.19 For effect is relatively small (approximately a .03 increase example, compared to the effects at a wage of $0 per in probability). However, if a mother with a preschool hour, a mother earning $7 per hour (in 1988) would have

13 IRPP Choices, Vol. 14, no. 12, October 2008 family inwhich themotherisemployed) verymuch shows thatthe decisiontouseparentalcare (fora preted indifferentways.However, thisregression haps becauseoftheavailability ofkindergarten). are somewhatmorelikely toreceiveparentalcare(per- other factorsconstant,four-andfive-year-oldchildren year-old childissimilarlysmall(-.01). Holdingall probability decreasesbyonly.02.Theeffectforatwo- child lessthanone)ofusingparentalcare,butthe year-old, thereisadecreasedprobability(relativeto with theuseofparentalcare.Forinstance,foraone- taken intoaccount. the useofparentalcarewhenallotherfactorsare municipality inwhichafamilylivesdoesnotaffect is bothsmallandinsignificant.Thesizeofthe tial (increasedprobabilityby.11), butinothercasesit ple). Inthisonecase,thesizeofeffectissubstan- tongue isEnglishorFrench—3percentofthesam- one caseoutoffour(recentimmigrantwhosemother sistent enoughtobestatisticallysignificantinonly likely touseparentalchildcare,butthisresultiscon- parental care(decreasedprobabilityof.21). (who isemployedorstudying)lesslikely touse only .05,however).Unsurprisingly, asinglemother care ismorelikely (theprobability isincreasedby flexible hours,parental(presumablypaternal)child probability ofusingparentalchildcarebyonly.01. an increaseof$10,000 inotherincomedecreasesthe dominates. However, thesizeofneteffectissmall; parental childcare.Apparentlythissecondfactor income alsomeansmoreabilitytopurchasenon- parental caremorelikely. However, higherother of hoursworked bythemother. Thiswouldmake income inlaboureconomicsistodecreasethenumber duce thisresult.Theclassiceffectofhigherother There aretwooffsettingfactorsthatlikely topro- lower thelikelihood ofthefamilyusingparentalcare. parental care—thehigherthisotherincome, husband’s earnings)isnegativelyrelatedtotheuseof parental careissomewhatlower, at.39. a wageof$30,thedecreaseinprobabilityusing bility ofusingparentalcare.Theeffectisnon-linear;at a wageof$20,shewouldhave.46decreaseinproba- a .34decreaseinprobabilityofusingparentalcare.At the meanwageinsample($10.37), shewouldhave a decreasedprobabilityofusingparentalcare.25.At No doubttheseregressionresults couldbeinter- Finally, theageof(youngest) childisassociated Immigrant familiesareapparentlysomewhatmore If thefatherinfamilyisnotemployedorhas Other incomeinthefamily(inmostfamilies, Familiesliveindifferent circumstancesandmake • Children whodonothaveanemployedorstudy- • Over thelastdecade,therehasbeenrapidgrowth • Excluding kindergarten,justoveronethirdofall • Includingkindergarten,abouthalfofallCanadian • Summary change ifthosecircumstancesweretochange. ment circumstancesofthefamily, andmightwell ly verymuchrelatedtotheeconomicandemploy- that thedecisiontoprovideparentalcareisapparent- to worktheeveningshift,forinstance.Thepointis to provideparentalcaremaybepartofthedecision ed tothedecisionprovideparentalcare.Thedesire at homeortoworktheeveningshiftarenotunrelat- the variationinlikelihood ofusingparentalcare. work, wageandincomevariablesthatexplainmostof insignificant orhavearelativelysmalleffect.Itisthe However, thesevariableseitherarestatistically the youngestchildbeingatleastoneyearold. in amunicipalitywithfewerthan100,000 peopleand (for parentalcare)wouldbeimmigrantstatus,living that aremostcloselyassociatedwithpurepreferences an economicandwork-relatedone.Thosevariables than two-parent families.Mothers’education is regulated care, andlessexclusively parentalcare, hours perweek.Single-parent familiesusemore sively parentalcareatfewer (lessthan20)working decisions areinterrelated,with moreuseofexclu- considerably. Mothers’workpatternsandcare dren getoldertheuseofregulatedcareincreases different childcaredecisions.Forinstance,aschil- parental care. least onequarterofthesechildrendidreceivenon- available —butveryolddatasuggestthatat ing motheralsoreceivenonparentalcare;thebest and therestofCanada. parental carehavefallensharplyinbothQuebec 40 percent).Carebynonrelativesandexclusively of Canada(whereusehasdoubled,tomorethan (where usehastripled,to72percent)andtherest in theuseofregulatedchildcarebothQuebec parental care. and justunderonethirdreceiveexclusively studying mothersattendregulatedcareservices, Canadian preschoolchildrenwithemployedor care. just overonefifthreceiveexclusively parental mothers attendregulatedchildcareservices,and preschool childrenwithemployedorstudying Of course,thedecisionstoworkpart-time, 14 only moderately related to the type of child care Table 14 chosen. However, families with highest incomes Annual Child Care Costs Borne By Families With are likely to choose regulated over unregulated Preschool Children and Mother Employed, Canada Excluding Quebec, 2005 care. Families with lower incomes are also more likely to use regulated care, but parent/family care Annual child care costs Percent of families

is their second most frequent choice. $0 38.5 • Exclusively parental care by families in which the $1-$999 15.8 mother is employed or studying takes several $1,000-$1,999 6.6 $2,000-$2,999 5.8 forms. Old data, which are nonetheless the best $3,000-$3,999 6.4 available, indicate that about half of this care is $4,000 or more 26.9 Total 100.0 provided by fathers but the other half is provided by mothers while they work (often at home) or Note: All the tables on child care spending in this paper do not include spend- ing on kindergarten services because they are free of charge. while they are on maternity leave from a job. Source: Statistics Canada, Survey of Household Spending, 2005. • For families in which the mother is employed or studying, the evidence suggests that the decision families have no child care expenditures. These are to use parental rather than nonparental care is an families using care provided by parents (some on economic and work-related one. Mothers who maternity/parental leave), using care provided by close work at home, are on maternity/parental leave, or relatives, using only free public kindergarten or receiv- who work on an evening or night shift are quite ing full subsidy for regulated child care. Another 3 out likely to use parental care. Other families are more of 10 spend between $1 and $2,999 per year on child likely to use nonparental care. care. These might be families with parents working part-time or part-year, paying a modest amount to rel- atives or purchasing low-priced informal care, but they Spending Patterns in Child Care: would also include most of the subsidized families in The Issue of Affordability Canada (excluding Quebec) using regulated child care. What do Canadian families spend on child More than 3 out of 10 families spend a considerable care? amount on child care — $3,000 or more annually. Table 15 shows trends in inflation-adjusted annual child care spending in the period 1987 to 2005. These here are a number of sources of information on data are from the SHS (with the 1987 data from the child care spending by families. The most detailed CNCCS). Again, spending by Quebec families is omit- source is the CNCCS, but these data are from

T ted; because of the availability of $7-per-day child New Evidence about Child Care in Canada, by Gordon Cleveland et al. 1988. The Survey of Household Spending (SHS) is an care, average child care expenditures are considerably annual survey (successor to the Family Expenditure lower for Quebec families. In the rest of Canada over Survey) that collects information on the entire range of this period, there has been a modest upward trend in family expenditures, including a small amount of detail child care spending until 2005, when we see a sharp on child care spending. In the SHS, spending is not spike upwards.20 associated with any specific child, so the figure given is Although the total annual expenditure may seem an aggregate for all children in the family. low, it is substantial as a percentage of the mother’s Child care spending by families is remarkable for annual earnings. There is fluctuation from year to year its variation. Table 14, from the 2005 SHS, gives a (apparently mostly in the denominator: the mother’s snapshot of the distribution of annual child care earnings). In nearly every year, the expenditure is over spending by families with an employed mother and at 20 percent of the mother’s earnings on average (and least one preschool child. Because spending in close to 30 percent in 2005). Total family earnings fluc- Quebec is strongly affected by the availability of $7- tuate as well, but child care spending is between about per-day child care, these figures do not include 5 percent and 10 percent of family earnings in different Quebec. Economists expect many variables will be years (and close to 8 percent in 2005). distributed according to a bell-shaped distribution, This points to two further questions. First we need to with most of the observations in the middle and a explore what factors influence the level of spending smaller number at the extremes. The distribution of when families do have child care expenditures. Second, child care spending is the opposite. Nearly 4 out of 10 we need to discern what kinds of families have no child

15 IRPP Choices, Vol. 14, no. 12, October 2008 household income (afteraccountingforchanges in OECD (2005)furtherinvestigates thechangeinnet parent familyearningtwothirds oftheAPW). APW) to42percentofnetfamily income(forasingle- 100 percentoftheaverageproduction worker’s wage— ily income(forafamilyoftwoearnerseachearning children, thesecostsrangefrom18percentofnetfam- ered, thantheOECDaverage.Forafamilywithtwo to childcare,arehigher, forallfamily typesconsid- costs, takingintoaccounteligibilityforbenefitsrelated calculations forAlbertaandBritishColumbia). see Kershaw2007, 25-29,forsimilarwork-incentive incentives initsmembercountries(OECD2005,2007; effect ofchildcarecostsandotherfactorsonwork specific benefits,theOECDhasrecentlyestimated benefits, andamultitudeofgeneralchildcare- software thatsimulatestaxes, tax-relatedchildcare cerning averagefeesforregulatedchildcare,andwith in OECDcountries(OECD2007). Withevidencecon- between workexpectationsandfamilyresponsibilities been investigatingwaysofimprovingthebalance population agingandaboutdecreasedfertility, has The OECD,concernedabouttheworkforceeffectsof Child carecostsandemploymentincentives burden imposedbychildcarecostsinCanada. review somerecentevidencefromtheOECDabout care expenditures.Beforethat,however, itisusefulto 053. 33$,5 4432. 7.8 6.0 7.3 7.0 9.1 10.3 28.4 8.4 23.7 8.6 19.0 7.7 24.4 5.5 22.0 $4,483 23.4 $3,704 21.9 $3,824 22.9 $3,533 24.1 $3,558 $2,757 13.4 $3,331 $2,346 $3,437 $2,377 $3,294 $2,308 $3,116 $2,184 33.3 $2,600 $2,304 30.4 $2,280 30.0 $2,271 29.5 $2,128 29.7 $1,600 30.3 38.5 31.9 36.7 27.6 37.9 28.8 34.7 38.6 23.3 30.8 33.7 31.0 Sources: StatisticsCanada,CanadianNationalChildCareSurvey(1988)forannualdatafrom1987,PublicUserFileof 31.7 figures refertofamilieshavingpositiveexpenditures. those intherestoftable.Figuresforaverageexpenditurecolumnsfourandfivehavebeenadjustedtoreflect2005pri 38.6 Notes: *Figuresfor1987areallofCanadaincludingQuebec.SincethisisfrombeforetheQuebecchildcarereforms1 2005 2004 2003 2002 2001 2000 1999 1998 1997 1987* stant dollars) Families withPreschoolChildrenandEmployedMothers, CanadaExcludingQuebec,1987-2005(con- Average AnnualParental SpendingOnChildCareandasaPercent ofEarnings, Table 15 Year (%) (%) no (%) ex expenditures (%) Year Employment decisionsaremade atthemargin; In Canada(Ontario,specifically),netchildcare pnigzr 300o icuigo aiis s%o spendingas as%of offamilies including $3,000or spending zero ncidcr mr nulyfmle ihwt oiiemte’ %offamily mother’s withpositive familieswith moreannually on childcare aiis pnig hl ae o hl ae pnig Childcare spending onchildcare childcare, spending Families aiis xedtr nepniue Childcare expenditure expenditureon Families favourable still. spendable householdincomebyonly50percent. her spouse(110 percentofAPW),shewould increase hold income,andevenifsheweretoearnmorethan sion tobeemployedwillhavenoneteffectonhouse- minimum wage(about35percentofAPW),herdeci- fees, taxes andreducedbenefits.Infact,ifsheearns her increasedearningswillbeeatenupbychildcare years, andearnstwothirdsoftheAPW, 78percentof lives inOntario,hastwochildrenagedandthree child care.Ifthesecondearnerinamarriedcouple employed tobeingandusingregulated costs) whenanearnermovesfromnotbeing income andpayrolltaxes, benefitsandnetchildcare employment. The universalloweringofchild care reforms havehadasignificant impactonmothers’ Quebec $5-per-day(now$7-per-day) childcare able householdincomewould increaseby50percent. cent; ifshecouldearn110 percentofAPW, her spend- spendable householdincomewouldincreaseby25per- earn atthemuchhigherrateof90percentAPW, her tial earnings.However, ifthissingle parent wereableto financial rewardforsingleparentswithmodestpoten- wall”; employmentmaygeneratevirtuallynonet increased income.Thisisthewell-known“welfare and reducedbenefitswilleatup97percentofher can earntwothirdsoftheAPW, childcarefees, taxes two andthreeyearsolddecidestobeemployed Average The situationofsingleparentsinCanadaisless In thiscontext,itisperhaps nosurprisethatthe 16 21 edtrserig earnings penditures earnings If asingleparentwithtwochildren ces usingtheConsumerPriceIndex.These 997, thefiguresarebroadlycomparableto HouseholdSpending,1997through2004. fees to under $2,000 per year for full-time care has It may be that good-quality child care is prohibitively increased workforce participation, hours worked and expensive for many families and that, as a result, families annual earnings. Employment participation has risen do whatever they can to secure free or low-cost alterna- at least 7 percentage points (Baker, Gruber, and tive care. These alternative arrangements would include Milligan 2005; Lefebvre and Merrigan 2005). having one parent work part-time, parents arranging their Lefebvre and Merrigan (2005) and Lefebvre, Merrigan work shifts so they can share child care responsibilities, and Verstraete (2007) find especially strong employ- having one parent take salaried or self-employed work ment effects for mothers with high school education that can be done at home so that he or she can care for or less. Overall, the $5-a-day policy has moved their child while working, having a relative provide care Quebec from a situation of lower employment rates for little or no remuneration, obtaining a full subsidy for than the rest of Canada amongst mothers with young regulated ECEC, using free public kindergarten as a form children to a situation of consistently higher rates. of ECEC, or some combination of these various strategies. This first interpretation leads to the judgment that Why do some families spend a lot and others good-quality nonparental child care is desirable but nothing? unaffordable for most, and that many families are The puzzle with respect to family expenditure on choosing either nonparental care of inferior quality early learning and child care is that many families (inadequate for optimum child development) or incon- have zero expenditures and many others quite high venient but free family care. Further, this interpretation ones. In the following section, we will see that about suggests that families suffer considerable stress trying 3 out of 10 families with an employed mother and a to provide child care through the family while holding preschool child spend nothing on the purchase of down a job, or else are sacrificing considerable income ECEC services, and among those who do have such to reduce their child care costs to an affordable level. expenditures, the average expenditure, on all their Alternatively, it may be that parents could readily children, is about $4,500. The picture that many ana- afford good-quality nonparental care but are uncomfort- lysts begin with — that families with a preschool able with purchased child care and prefer to spend time child and two employed parents (or an employed sin- with their children themselves when the children are gle parent) must necessarily spend at least $6,000 to young. Even if less expensive non-family services were $9,000 per child per year — is apparently wrong. The available, they would prefer not to use them. This second OECD results discussed above suggest why many interpretation leads to the judgment that families are families do not spend this amount, but several factors doing precisely what they want with their child care influence spending patterns. spending dollar. Many families spend nothing because

they prefer parent, family and relative care. Many others New Evidence about Child Care in Canada, by Gordon Cleveland et al. Explaining spending patterns spend only a small amount because they prefer part-time Most nonparental child care is expensive. In general, to full-time employment. Others spend a lot because the higher the quality of the care, the more expensive they believe the expenditure brings them good-quality, it will be. Although care by a relative may cost noth- enjoyable, healthy and safe experiences for their chil- ing, it is available only to some, and very often only dren, preparing them optimally for school and life. on a part-time or part-week basis. For most types of Let us probe these alternative interpretations by looking nonparental care, quality is closely linked to the care- at the factors that influence the decision to spend nothing giver’s level of training in child development and to on child care and, for those who do purchase child care, the number of children looked after by the caregiver. the decision with respect to how much to spend. Cleveland and Krashinsky (2004) estimate the full The literature includes a handful of economic studies unsubsidized cost of centre-based care for a three- on child care expenditure. Brayfield and Hofferth year-old child to be more than $6,000 per year, even (1995), analyzing data from the US National Child Care at a modest caregiver salary of $26,000, and close to Survey of 1990, find that it is economic factors, partic- $9,000 per year at a salary of $35,000 for staff ularly those associated with the mother, that largely trained in early childhood education. determine who pays for care, how much they pay and There are two possible interpretations of spending the share of earnings spent on care. “[I]t is out-of- patterns. These lead to different judgments about pocket costs of care, mother’s wage and family income family and child well-being and have quite different that will determine what type and quality of care is policy implications. purchased” (p. 175).

17 IRPP Choices, Vol. 14, no. 12, October 2008 wife’s couples ‘netout’thecostsofchildcarefrom full-year. “Anecdotal evidenceindicatesthatmany child careevenwhenbothspousesworkfull-time, income isnotassociatedwithhigherexpenditureon increase onlywithwomen’sincomes—highermale sumption. Expendituresonchildcarearefoundto ing onsomebutnotallcategoriesofhouseholdcon- 1992, findthatfamiliespooltheirincomesindecid- Statistics CanadaFamilyExpenditureSurveyof and Burton1998,609-10) ‘free’ childcarefromaparentorneighbour.” (Phipps only ‘afford’toworkforlowwagesiftheyhave results confirmsuchideas.Apparently, womencan regarded asverymuchawoman’sresponsibility. Our for hertotake apaidjob. Childcareseemsstilltobe aiyhsoecid04yaso g $4,800 $7,700 $6,700 $4,300 authors. $5,200 Source: StatisticsCanada, SHSPublicUserDataFile,2005.Calculations bythe $3,800 $2,800 Family hastwoormorechildren0-4years ofage $1,400 Family hasonechild0-4yearsofage $7,300 $4,200 Mother’s earnings$50,000orover $3,000 Mother’s earnings$40,000-$49,999 $2,000 $2,500 Mother’s earnings$30,000-$39,999 $1,800 $4,700 Mother’s earnings$20,000-$29,999 $1,600 Mother’s earnings$10,000-$19,999 Mother’s earningslessthan$10,000 Family income$100,000 orover $2,200 $3,000 Family income$80,000-$99,999 $3,300 $2,400 Family income$60,000-$79,999 $6,900 Family income$40,000-$59,999 Family income$20,000-$39,999 Family incomelessthan$20,000 $3,100 $4,600 Mother employedpart-time,part-year $3,500 Mother employedfull-time,part-year $6,000 Mother employedpart-time,full-year $3,400 Mother employedfull-time,full-year Rural Urban Separated, widowed,divorced $4,500 Never married Married orcommon-law Youngest childaged4-5 Youngest childaged2-3 Youngest childaged0-1 All familieswithanemployedmother Quebec, 2005 Family andChildCharacteristics, CanadaExcluding with Positive AnnualChildCareExpenditures, by with PreschoolChildrenandEmployedMother Average AnnualChildCareSpendingforFamilies Table 16 aiyo hl hrceitc expenditure Family orchildcharacteristics Phipps andBurton(1998),usingdatafromthe earnings indecidingwhetheritis‘worthwhile’ Average annual able amountonitwhentheyhavethemeanstodoso. value childcare,andtobewillingspendaconsider- level of$50,000ormore.Canadianfamiliesappearto earnings, risingfrom$1,400to$7,700 atanearnings Child carespendingisstronglyrelatedtothemother’s $7,300 forthosewithanincomeof$100,000 ormore. from $1,600forlow-incomefamiliestoanaverageof mother’s earnings.Annualchildcareexpendituresvary expenditure isclearlyrelatedtofamilyincomeandthe the motherisemployedfull-time,full-year).Childcare the family’saveragechildcareexpenditure($6,900if living inurbanareasalsospendmore. ratios). Marriedorcommon-lawfamiliesand ditures forolderchildren,asdohigherchild-staff spend more(kindergartenapparentlyreducesexpen- on childcare.Familieswithatwo-orthree-year-old that differentfamilytypesspendamounts of careprovidedbyparentsandrelatives. by differentincomegroupingsnorthe“cost” burden not abletocontrolforthequalityofchildcareused dicted incomeandchildcarespending,theyare not examinetherelationshipbetweenmother’spre- arrangements theymake. RosenbaumandRuhmdo and paylessthanhigher-incomefamiliesforthecare lower-income familiesuselower-costtypesofcare nomic statusthattheyconstruct,largelybecause not systematicallyrelatedtoameasureofsocio-eco- variation inthecostburdenacrossfamilies,butitis expenses dividedbyafter-taxincome.Thereiswide the “cost burden”ofchildcare,definedascare Survey ofIncomeandProgramParticipation tostudy nonparental care. 1988) inexplaining theamountspenteach weekon ployment andearningssituation is(orperhapswasin results showjusthowimportant themother’sem- what mightbeconsideredits key determinants.The amount ofweeklychildcare spendingin1988on with anemployedorstudyingmotherwas$46.38. expenditure perchildonpurchasedcareforchildren data fromthe1988CNCCS.Inaverageweekly ing arenotavailableintheSHS,sowemustuse allow ustoanalyzefactorsinfluencingchildcarespend- tinct factors.Dataonfamilycharacteristicsthatwould care child on try toseparateouttheindividualinfluencesofthesedis- spending of Many factorsaffectchildcarespending.Itisusefulto analysis Regression The more“fully”employedthemother, thehigher Table 16,usingdatafromthe2004SHS,shows Rosenbaum andRuhm(2005)usethe1996US Table 17showstheresultsofaregression 18 Table 17 Determinants of Weekly Amount of Child Care Spending for Preschool Children, Families with Mother Employed or Studying, Canada, 1988

Estimated Standard Mean value Explanatory variables coefficient error in sample

Works at home -.710** .051 .07 Works less than 20 hours/week -.599** .031 .23 Regular late-day or non-day shift -.522** .045 .09 Student or works varying shifts -.264** .026 .39 (Ln. of) predicted hourly wage .540** .047 10.37 Other income in the family/1000 .002** .001 27.99 Father’s employment is flexible -.011 .028 .26 Single-parent family .107** .044 .10 Recent immigrant, English or French as mother tongue .110 .071 .03 Recent immigrant, English or French not mother tongue -.078 .064 .05 Earlier immigrant, English or French as mother tongue .010 .058 .04 Earlier immigrant, English or French not mother tongue -.010 .071 .04 Family lives in municipality less than 100,000 population -.188** .025 .41 Number of children 0-5 in family .185** .024 1.32 Age of youngest child .057* .030 2.54 Square of age of youngest child -.024** .005 9.17 Constant 2.609** .127 1.00

N 4,574 Adjusted R-squared .249

Source: Statistics Canada, CNCCS, 1988. Note: These results are for families with positive weekly child care spendings. The weekly child care expenditure is transformed into a natural logarithm for the purposes of this regression (often done for dependent variables that are right-skewed). The predicted hourly wage is also in the form of a natural log. The mean log value of weekly child care expenditure is 3.74, or $42.10, which is the mean in the sample used for this regression (i.e., $68/week in mid-2008 $, or over $3,500/year). Regression is ordinary least squares (linear probability function). * statistically significant at p < .10; ** statistically significant at p < .05.

Although an increasing number of families pur- employment situation for the mother, in order to decrease chase ECEC purely for its developmental benefits, the cost burden of nonparental care. most families purchase child care to support the The results shown in table 17 reveal that working at employment and education or training activities of home decreases spending on nonparental care, as does the parents. In the majority of families, the mother working on a part-time basis. Part-time work often means earns less than the father. This, combined with tradi- evening and weekend work or variable shift arrange- New Evidence about Child Care in Canada, by Gordon Cleveland et al. tional role definitions, means that it is most often the ments. All of these employment situations are associated mother’s employment or study situation that triggers with less spending on nonparental child care. the need for nonparental care. As a result, affordability of child care, for most fami- Table 18 Predicted Changes in Weekly Spending (in 1988 $) lies, appears to involve a comparison between the costs on Child Care Associated with Discrete Changes in of care and the mother’s potential earnings. Families Family and Economic Circumstances want the best care they can afford. As table 17 suggests, Predicted change the higher the mother’s potential earnings, the more the Change in in weekly child family will spend on care (for better quality and more explanatory care spending Explanatory factors variable (at mean) hours). Once the mother’s earnings are taken into account, then, “other family income” (generally the Works at home 0 to 1 -$21.39 Works less than 20 hours/week 0 to 1 -$18.98 father’s earnings) plays only a minor role in the amount Regular late-day or non-day shift 0 to 1 -$17.06 spent on care. If the mother’s potential earnings are low, Student or works varying shifts 0 to 1 -$9.63 Predicted hourly wage +$10.00 +$18.67 the family may be forced to patch together arrangements Other income in the family/1,000 +$10,000 +$0.85 of dubious quality (caregivers with no training). The care Single-parent family 0 to 1 +$4.75 Family lives in municipality with chosen may be purely custodial, rather than including a population of under 100,000 0 to 1 -$7.21 developmental elements (see the evidence on the quality Number of children 0-5 in family 1 to 2 +$10.29 of unregulated care in the section on Quality, below). In Age of youngest child 2 to 3 -$1.85 many cases, these adjustments may involve a different Note: Based on results from regression reported in table 17.

19 IRPP Choices, Vol. 14, no. 12, October 2008 .Themotherisemployedonlyforthosehourswhen 4. .Thechildislessthanoneyear oldandthemother 3. Themotherworksathome(probablyself-employed) 2. Themotherworkspart-time(sometimesfull-time) 1. both parentshavejobs. their childrenwithoutspendinganymoneyevenif number ofwaysinwhichfamiliesarrangecarefor also providezero-pricedcare.We canidentifya based orfamilyhomecareandkindergartencan ents, butrelatives,neighbours,regulatedcentre- Much ofzero-pricedchildcareisprovidedbypar- Explaining zero-pricedchildcare on thepredictedvalueofweeklychildcarespending. discrete changesinthestatisticallysignificantvariables them inamoremeaningfulform.Itshowstheeffectof aiyhsoecid04yaso g 36.3 9.4 30.7 30.1 22.3 authors. 46.6 Source: StatisticsCanada, SHSPublicUserDataFile,2004.Calculations bythe 35.5 44.0 34.3 30.3 Family hastwoormorechildren0-4years ofage Family hasonechild0-4yearsofage 18.1 Family income$100,000 andover 24.4 37.3 Family income$80,000-$99,999 Family income$60,000-$79,999 31.8 35.1 Family income$40,000-$59,999 43.8 Family income$20,000-$39,999 Family incomelessthan$20,000 21.1 32.4 Top quintile—mother’s earnings Fourth quintile—mother’s earnings 23.4 Third quintile—mother’s earnings 21.9 Second quintile—mother’s earnings 46.0 Bottom quintile—mother’s earnings Rural Urban Separated, widowed,divorced 31.8Never married Married orcommon-law Youngest expenditures childaged4-5 Youngest childaged2-3 Youngest childaged0-1 All familieswithanemployedmother Family orchildcharacteristics by FamilyandChildCharacteristics, 2004 with PreschoolChildrenandanEmployedMother, Zero AnnualChildCareExpendituresinFamilies Table 19 the childisinkindergarten(fullschoolday provide carewhileclassifiedasemployed. is onmaternityleavefromherjob and thereforecanprovidethechildcarebyherself. and thefatherlooksafterchildren(off-shifting). Table 18takes theresultsfromtable17andexpresses 22 with nochildcare so isableto % offamilies hours, thenzero-priced (oftenpaternal) care ismore priced care,buttheeffectappears tobenegligiblysmall. band’s earnings)isnegatively relatedtotheuseofzero- er wouldhaveadecreasedprobability of.58 care. Atthemeanwageinsample($10.37), amoth- have adecreasedprobabilityof.46usingzero-priced hour, amotherwhoearns$7perhour (in1988)would culate that,comparedtotheeffectsatawageof$0per wages tousezero-pricedcare.Forexample,wecancal- are, ingeneral,muchlesslikely thanthosewithlower working fewerthan20hoursperweek. mother beingonmaternityleave,workingathomeor increase thelikelihood ofparentalcare,suchasthe zero. Manyofthesefactorsarethesameonesthat impact ofeachfactorontheprobabilitypaying starred coefficients;thesecoefficientsmeasurethe 20 indicatesthestatisticallysignificantfactorswith employed motherspayingzeroforchildcare.Table hood offamilieswithpreschoolchildrenand rest werechargedsomeamount. mothers whousednonparentalcarepaidzeroforit;the 11.7 percentofchildrenwithemployed orstudying reason oranother, paidazeroprice.Inotherwords,only only about90,000usednonparentalcareand,forone studying motherusedparentalcare(andpaidzero),and Canada, about390,000childrenwithanemployedor this representsaminorityofchildren.Indeed,in1988 the child’sownparents. care. Themajorityofzero-pricedcareisprovidedby parental careinsteadofsomeformnonparental table 13,explainingthedecisionoffamiliestouse surprisingly, theseresultsareverysimilartothosein affect thedecisiontousezero-pricedchildcare.Not Thefamilyiseligible,duetolowincome,forafull 7. Aneighbourwithchildrenprovidescarefreeof 6. Agrandmotherorotherrelativeprovidescareand 5. If thefatheris notemployedorhasflexible Other incomeinthefamily(in mostfamilies,thehus- Our resultsshowthatmotherswithhigherwages There aremanyfactorsthatwillaffectthelikeli- Some nonparentalcareisavailableatazeroprice,but Table 19considersthedifferentfactorsthatmay small parentalcontributionisthenorm). rare evenforthoseon“full”subsidy, becausea child caresubsidy(infact,zero-pricedis other daysoftheweekorhoursday. day inexchange forcarebythemotheronsome charge oncertaindaysoftheweekorhours does notchargeforit. part-day inotherprovinces). Quebec, NovaScotiaandNewBrunswick;typically 20 Table 20 Factors Related to the Use of Zero-Priced Care for Preschool Children, Families with Mother Employed or Studying, Canada, 1988

Estimated Standard Mean value Explanatory variables coefficient error in sample

On maternity leave .473** .025 .07 Works at home .304** .017 .10 Works less than 20 hours/week .132** .013 .23 Regular late-day or non-day shift .139** .018 .09 Student or works varying shifts .061** .012 .39 Predicted hourly wage -.087** .014 10.37 Square of predicted hourly wage .003** .001 115.24 Other income in the family/1,000 -.000* .000 27.99 Father’s employment is flexible .085** .012 .26 Single-parent family -.044** .019 .10 Recent immigrant, English or French as mother tongue .123** .029 .03 Recent immigrant, English or French not mother tongue .140** .024 .05 Earlier immigrant, English or French as mother tongue -.014 .026 .04 Earlier immigrant, English or French not mother tongue .126** .028 .04 Family lives in municipality of under 100,000 population -.010 .011 .41 Number of children 0-5 in family -.064** .010 1.32 Age of youngest child -.136** .013 2.54 Square of age of youngest child .022** .002 9.17 Constant .006 .081 1.00

N 7,062 Adjusted R-squared .200

Source: Statistics Canada, CNCCS, 1988. Note: The weekly child care expenditure is transformed into a natural logarithm for the purposes of this regression (often done for dependent variables that are right- skewed). The predicted hourly wage is also in the form of a natural log. The mean log value of weekly child care expenditure is 3.74, or $42.10, which is the mean in the sample used for this regression (i.e., about $65/week in 2006 $, or close to $3,400/year). Regression is ordinary least squares (linear probability function). * statistically significant at p < .10 ** statistically significant at p < .05. likely (increasing the probability by .09). A single Summary mother (who is employed or studying) is slightly • Families with an employed or studying mother and less likely to use zero-priced care (decreased proba- preschool children may spend nothing on child care, bility of .04). or a little, or a substantial amount. In fact, 4 out of Immigrant families, according to the results dis- 10 such families in Canada outside Quebec do not cussed earlier, are generally not more likely to use have annual child care expenditures. Another 3 out New Evidence about Child Care in Canada, by Gordon Cleveland et al. care by a parent while the mother is employed. of 10 spend less than $3,000, while a final 3 out of However, they apparently have other sources of zero- 10 spend more than $3,000. priced care (e.g., live-in relatives). Immigrant families • Accounting for inflation, those who do spend on who have been in the country more than 15 years child care are now spending much more than they and whose mother tongue is English or French are did 20 years ago; average spending has risen from similar to native-born families. However, both of the $2,600 to nearly $4,500 annually. In a typical year, other immigrant categories (in the country more than this is between 20 percent and 30 percent of the 15 years but with a different mother tongue; in the mother’s gross earnings, or about 8 percent of the country less than 15 years with any mother tongue) family’s total earnings. are more likely to use zero-priced care (increased • Looking at cross-tabulations or using regression probability of .12 to .14). analysis, it is clear that economic and work situa- This regression gives evidence that the decision to tions dramatically influence family decisions about use zero-priced rather than paid nonparental care is child care spending. In particular, working at home, very much an economic and work-related one, as working part-time and working shifts are associated well as being influenced by cultural and preference with lower child care expenditures, while the moth- factors. While immigrant status and the age of the er’s ability to earn an income has a strong positive child affect the use of zero-priced care, employment association with child care spending. This is consis- and the mother’s wage variables are central explana- tent with work by Phipps and Burton (1998), who tory factors related to the use of zero-priced care.

21 IRPP Choices, Vol. 14, no. 12, October 2008 best evidence we haveisthatqualitythe key fea- Thestrongassociationsbetweeneconomicvari- • Careprovidedbyaparent(andoftenrelative) • when qualityispoor. ments, particularlyforvery youngchildrenand behavioural effectsofgroup childcarearrange- Vandell 2004). Gormley etal.2005;NICHDandDuncan2003; learning environment(Doherty2007, 30-36; dren’s development,particularlyinahigh-quality The resultssuggestthatchildcarematterstochil- R Quality inNonparentalChildCare matter forfindingsonbehaviour decisions constrainedbytheaffordabilityofcare. represent difficultemploymentandchildcare Canada arenotdrivenpurelybypreferencebut child careuseandspendingpatternsintherestof 1997-2001 (seetable5)alsostronglysuggestthat lowing thechildcareandfamilypolicyreformsof patterns (andemploymentpatterns)inQuebecfol- care costs.Thedramaticchangesinchilduse er’s potentialearningsandout-of-pocket child ability involvesacomparisonbetweenthemoth- suggests that,inmanyfamilies,childcareafford- mothers’ earningsandworkvariables of employedparents.Further, theimportanceof ability ofcareconstrainsthechilddecisions ables andchildcarespendingsuggestthatafford- zero-priced care. a singleparentarenegativelyassociatedwith a higherwage(inthecaseofmother)andbeing priced care.Ontheotherhand,beingabletoearn English orFrenchisalsoassociatedwithzero- grant oranimmigrantwhosemothertongueisnot the useofzero-pricedcare.Beingarecentimmi- and workingshiftsareallstronglypredictiveof nity leave,workingathome,part-time children fourorfiveyearsofage.Beingonmater- is freeofcharge,askindergartenforeligible mother’s earnings. child careexpendituresdependlargelyonthe pool theirincomestopayforchildcareandthat found that,ineffect,Canadianfamiliesdonot There issomeconcernabout possiblenegative child developmentisnowwidelyavailable. child carequalityondifferentaspectsof esearch evidenceconcerningtheeffectsof 23 However, contextseemsto 24 and, overall,the some studiesintheUnitedStateshavedoneso. provided byparents.Suchdata kindergartens, nurseryschoolsandpreschools,or quality ofcareprovidedbyrelatives,in care. Noneofthestudiesprovidesgooddataon ments familieswereusing,includingunregulated inspections ofqualityinwhateverchildcarearrange- inspection. TheÉLDEQisdistinctinthatitentailed of childdevelopmentbasedonin-depthon-site the period2000to2003. dren twoandahalftofiveyearsoldinQuebec 1,540 ECECarrangementsusedbyasampleofchil- and theÉLDEQ,whichcollectedinformationon lated familychildcarehomesinQuebec2003; preschool roomsinchildcarecentresand200regu- 1998; theGrandirenqualitéstudyof705 infant and care homesinsixprovincesandoneterritory 223 childcarecentresand231 regulatedfamily child from threedatasources:theYou BetICare! study of Canada’s informationonchildcarequalitycomes services. Ignoringafewsmall,localizedstudies, statistical informationonthequalityofitsECEC influence ofchildcareonchildren,Canadahaslittle fied qualityasthemostconsistentfactorin and childdevelopment: ings basedontheliteraturechildcarequality Shonkoff andPhillips(2000)summarizetheirfind- The ScienceofEarlyChildhoodDevelopment development. In ture ofchildcareprogramsthatpromote profit centre-based settings(CPEs—Centres dela of twoandahalfyears.Itevaluated notonlynon- attended bythesechildrenafter theyreachedtheage an evaluationofthequality ofchildcaresettings October 1997throughJuly1998. Thestudyincluded sentative sampleofchildren borninQuebecfrom The ÉLDEQbegan,in1998,tocollectdataonarepre- Evidence fromtheÉLDEQ Each ofthesestudiesprovidesdataonthequality Although childdevelopmentresearchhasidenti- reading… (p.313-14) peers, toearlycompetenceinmathand initiate andsustainpositiveexchanges with co-operation withadultstotheability want toseeintheirchildren,rangingfrom is associatedwithoutcomesthatallparents developmental outcomes,high-qualitycare poor qualityisassociatedwithpoorer developmental science.Whilechildcareof ied isoneofthemostconsistentfindingsin children’s developmentthathasbeenstud- care qualityandvirtuallyeveryfacetof In sum,thepositiverelationbetweenchild- 22 From NeuronstoNeighbourhoods: can be collectedand , 25 Petite Enfance), for-profit centre-based settings centre-based and home-based CPEs. However, just over (garderies) and regulated family child care (home- one out of every four arrangements in a for-profit cen- based CPEs), but also unregulated home-based set- tre or an unregulated family home is inadequate by this tings. The evaluation of the quality of unregulated measure.26 care is virtually unprecedented in Canada, making Similarly, while about 27 percent of all arrange- this data source particularly valuable for the study of ments are of good quality, this is true of less than 15 the quality of child care settings. percent of for-profit centres and about 10 percent of The sample included 2,223 children but only 1,540 unregulated family homes. About 35 percent of cen- on-site evaluations were conducted: 728 in centre- tre-based CPEs and about 30 percent of home-based based CPEs, 337 in home-based CPEs, 296 in for-profit CPEs score in the good-quality range. This distribu- centres and 179 in unregulated home-based settings. tion is also reflected in the average scores for each The instruments used to evaluate quality (the type of arrangement. The average score is 4.53 for Early Childhood Environments Rating Scale — centre-based CPEs and 4.32 for home-based CPEs. Revised, or ECERS-R [Harms, Clifford, and Cryer For-profit centre-based arrangements have an aver- 1998], and the Family Day Care Rating Scale, or age quality of 3.65, while unregulated home care FDCRS [Harms and Clifford, 1989]) have frequently averages an almost identical 3.64. been used in studies of this type and provide a Chart 1 shows how much difference there is in the global assessment of the quality of the service pro- underlying distributions of quality for care of differ- vided. As Japel, Tremblay, and Coté (2005) put it, ent types. Very few unregulated home care arrange- “These scales, whose validity and reliability are well ments are above the midpoint on the quality scale; established, measure the parameters elaborated by the vast majority score below 4. The majority of cen- the NAEYC (National Association for the Education tre-based CPEs are in the top half of the scale. of Young Children in the U.S.) to ascertain whether Home-based CPE care is oriented towards lower a particular setting is conducive to child develop- quality, with both low- and high-quality facilities. ment” (p. 11). For-profit care is more strongly distributed towards The ECERS-R, for example, consists of 470 indica- lower quality. tors grouped into 43 items in 7 subscales (Space and Furnishings, Personal Care Routines, Language and Reasoning, Activities, Interactions, Program Struc- Chart 1 Distribution of Quality Levels by Type of Child ture, Parents and Staff). Each centre gets a quality Care in Quebec, Actual Plus Predicted score from 1 to 7. A score of 1 is “inadequate,” 3 is 70

“minimal,” 5 is “good” and 7 is “excellent.” New Evidence about Child Care in Canada, by Gordon Cleveland et al.

The ÉLDEQ findings show that the majority 60 (about 70 percent) of all child care arrangements score below the level of good developmental quali- 50 ty. While the majority of these are not harmful to children, neither do they stimulate development in 40 a way that would be considered adequate. Just over Percent 10 percent of arrangements are problematic in 30 terms of quality (below minimal), another 25 per- cent are just over minimal quality and therefore 20 well below the level needed for desired child devel- 10 opment, and a further 35 percent are just below good quality on average. Only about 5 percent of 0 arrangements are truly exceptional (between 6 and 1 to 1.99 2 to 2.99 3 to 3.99 4 to 4.99 5 to 5.99 6 to 6.99 Good 7 on average). Inadequate Minimum The adequate and inadequate child care arrange- CPE CPE-Home- For-profit Unregulated ments are not randomly distributed across types of based daycare home-based care. While just over 10 percent of all measured Source: Authors’ calculations based on ÉLDEQ. Chart shows distribution of qual- ity scores (1 to 7) based on actual, but also predicted, scores. Scores for non- arrangements are considered inadequate (i.e., below respondents are predicted by regression of quality on family characteristics. minimal quality), this is true of just over 5 percent of Results available from authors.

23 IRPP Choices, Vol. 14, no. 12, October 2008 months); centre-based careforpreschoolchildren(18-59 children, andinteractionbetweeneducatorspar- zation ofactivities,interactionbetweeneducatorsand four dimensions:physicallayoutoffacilities,organi- for childreninfamilychildcare.Thescalescomprise centres, oneforpreschoolchildrenincentresand ment instrumentweredeveloped—oneforinfantsin care inQuebec.Threeversionsofthequalitymeasure- behind theeducationalprogramofregulatedchild cally tomeasureconformitywiththeprinciples (Bourgon andLavallée2003)wasdevelopedspecifi- not includeunregulatedsettings. trast totheÉLDEQ,GrandirenQualitéstudydid with CPEsandservingchildrenofallages.Incon- ( viding centre-basedcareforpreschoolchildren;CPEs from alargenumberofCPEs( child carearrangements(coordinatedbyCPEs). care providedinCPEs, in 2003toconstructasnapshotofthequalitychild The GrandirenQualitédatawerecollectedinQuebec Evidence fromthegrandirenqualitésurvey and familychildcareproviders( 33percentverygoodquality. 83.3 Note: Inthischart,less than33.3percentrepresentsinadequate quality, 33.3to50percentrepresentsminimum quality, 66.6 Distribution ofLevelsQualityinfor-ProfitDaycaresandNonprofitCentresdelapetiteenfance(CPE), Chart 2 N The This study(Drouinetal.2004) = 128)and Proportion Échelle d’observationdelaqualitééducative 20 40 60 0 garderies 04 08 04 080 60 40 20 80 60 40 20 garderies (for-profit daycare garderies ( N = 124)servinginfants; For-profit N N and regulatedfamily N = 228)providing 27 ga = 200)affiliated collected data r d eries ) ( N = 225)pro- Levels ofq average scoreof2.93, Measured onascaleof1to4,nonprofitCPEsratean average, acrossnonprofitandfor-profitcentres. we findsubstantialgrossdifferencesinquality, on set, preschoolisdefinedascovering18-59months), quality andthefinalthirdtoverygoodquality. fair quality, thenextthird(i.e.,3.00to3.49)good tory, withthelowestthirdofthisrangecorrespondingto ty. Average scoresof2.50to4.00areconsidered satisfac- the finalthird(i.e.,2.00to2.49)somewhatpoorquali- to verypoorquality, thenextthirdtopoorqualityand the firstthirdofthisunsatisfactoryrangecorresponding are consideredtorepresentunsatisfactoryquality, with scores uptothemidpointofthisrange(i.e.,1.02.49) 4.0, sothattherangeisessentiallythreepoints.Average makes upthesubscalesandscalesisratedfrom1.0to score foreachgroupofchildrenchosen.Eachitemthat intervention andpositivecommunication. placed onlearningthroughplay, democraticmodesof ties. Theinteractionscalehasthreesubscales:value planning, observation,dailyscheduleandplayactivi- ond scale,measuringactivities,hasfoursubscales: arrangement andmaterialsequipment.Thesec- ents. Thefirstscalehastwosubscales:physical Looking atpreschoolclassroomsonly(inthisdata The fourdimensionsareaveragedindeterminingthe ua lity 24 non-profit CPE garderies percent representsgood qualityand an averagescoreof Quebec, 2003 2.58. If we transfer these scores onto a percentage The ITERS (Harms and Clifford 1990) is a similar meas- scale, for-profit centres score, on average, 52.5 per- ure that assesses process quality in classrooms for children cent, whereas nonprofit CPEs score 64.2 percent. In under two and a half. Its 35 items assess the same seven other words, there is a 12-percent difference between aspects of quality. We treat the seven-point ITERS and these two types of centre (or the quality of CPEs is, ECERS scales as equivalent measures of quality, so that we on average, more than 22 percent superior to that of can pool together infant/toddler and preschool classrooms garderies). This is a substantial difference, greater (and for convenience we express this score on a scale from than one standard deviation (s.d. of index quality is 0 to 100, rather than from 1 to 7).30 .26 for CPEs and .34 for garderies). Table 21 shows, in the bottom row, that 28.6 percent We have depicted the sample distributions of qual- of all classrooms observed in this study are in for-profit ity on a graph. Chart 2 shows two histograms of centres, while 71.4 percent are in nonprofits. In other quality scores (in percentage terms). The histogram words, about one out of every three classrooms is for- on the left shows scores for preschool classrooms in profit and two out of every three are nonprofit. If quality garderies, the one on the right for CPEs. Somewhat were randomly distributed across classrooms, we would less than half of the mass of the left-hand distribu- anticipate that about one out of every three classrooms tion is below the 50-percent mark on the quality at each quality level (1 to 7) will be for-profit and about scale (i.e., unsatisfactory). In contrast, only a small two out of every three will be nonprofit. However, even part of the right-hand distribution lies below 50 per- though for-profits span the range of quality levels from cent. In these graphs, 66.6 percent represents good 1 to 6, and there are few in the two lowest categories, quality and 83.3 percent very good quality. Only a they are overrepresented in quality levels 1 through 4. small number of garderies are in the good or very Nonprofits span the range of quality levels, too, from 1 good range, compared to a substantial number of through 6. However, the bulk of nonprofits are in the CPEs. higher ranges, being overrepresented at levels 5 and 6. Three statements are simultaneously true. First, the Evidence from the You Bet I Care! survey distribution of nonprofits shows a mean shift towards You Bet I Care! is the collective name for a group of higher quality. Second, a high percentage of both for- linked data sets investigating earnings, working con- profits and nonprofits provide good-quality care (5 or ditions, and observed quality in daycare centres and better) — about 26 percent of for-profits provide care at licensed family homes in Canada in 1998.28 this level, compared to about 43 percent of nonprofits. The two main measures of classroom quality in Third, a high percentage of both for-profit and non- this data set are the Infant-Toddler Environment profit centres provide relatively low-quality care (3 or

Rating Scale (ITERS) and the ECERS-R, which are less) — about 39 percent of for-profit and 27 percent of New Evidence about Child Care in Canada, by Gordon Cleveland et al. global measures of the developmental potential in the nonprofit centres. classroom (the environment fostering interactions). The updated version of the ECERS-R (Harms, Clifford, What factors determine child care quality? and Cryer 1998) comprises 43 items that evaluate This is not an easy question to answer. We do not yet seven aspects of centre-based care for children from have all the tools or the information to fully answer it. two and a half to five years.29 The point of this section is to say that we do, however, know quite a lot. In this section, we provide evidence Table 21 from the You Bet I Care! study with respect to factors that Number and Proportion of Classrooms in For- influence the quality of care offered in child care centres. Profit and Nonprofit Centres at Each Quality Researchers in child care and developmental psychol- Level, Canada, 1998 ogy describe a two-stage process through which ECEC Quality For-profit (%) Nonprofit (%) Total (%) services affect children: first, a set of structural and

1 (inadequate) 2 (40) 3 (60) 5 (100) other inputs produces “process quality” in the child care 2 9 (47.4) 10 (52.6) 19 (100) facility (e.g., classroom); second, the process quality of 3 (minimal) 25 (33.3) 50 (66,6) 75 (100) ECEC services, in combination with family, neighbour- 4 33 (32) 70 (68) 103 (100) 5 (good) 20 (21.5) 73 (78.5) 93 (100) hood and other factors, determines the child outcomes 6 4 (13.3) 26 (86.7) 30 (100) (cognitive, behavioural, emotional, etc.). Total 93 232 325 It seems logical to economists to consider the prob- Source: Cleveland et al. (2007, 36) based on the You Bet I Care! data set. lem of producing quality in ECEC facilities in the con-

25 IRPP Choices, Vol. 14, no. 12, October 2008 output; ECERS-RandITERScanbeusedtomeasure between inputsandoutput.Classroomqualityisthe therefore reliableandreproducible,relationship cost perunitofqualitychosen. purchasers arewillingtobuy, whileminimizingthe inputs toproducethechosenlevelofqualitythat Inputs areavailableatdifferentprices;centresselect other inputstoprovidequalityservicesforchildren. with variouscharacteristicsandcombinethem text ofproductiontheory. Centreshirehumaninputs xlntr aibe 40 325 * Statisticallysignificant at (1.08) authors. (1.07) ables controllingformarketlocation.Fullregressions areavailablefromthe Note: FromtheYou Bet ICare!dataset.Regressionalsoincludesdummyvari- (1.04) 0.03 Adjusted -0.05 (3.36) Explanatory variables (1.97) Number ofobservations (2.37) 0.13 Constant 26.96** (1.90) 0.09** 0.018** % ofchildrensubsidized % ofchildrenwithspecialneeds % ofinfants/toddlers (1.91) 4.20* Number ofFTEchildrenincentre (0.28) Child variables 6.82* Revenue perFTEchild (1.77) (1.93) 1.26 Financial resources Centre isnonprofit 5.77* 3.92* (2.97) Ownership ofcentre (1.40) University degree Post-college certificate 12.35** College diplomaorcertificate (3.47) (1.86) 5.34 (4.05) Education levelofdirector (1.39) (2.48) in last12months 9.22** Professional developmenttraining -0.06* 10.32** ECE —universitydegree (0.01) 1.14 ECE —post-collegecertificate 1.83** (2.83) ECE —collegediploma(2or3years) ECE —collegecertificate(1year) 0.04 Training inearlychildhoodeducation -3.04** Square ofgroupsize Group size(3-5years) Group size(0-2years) Child-staff ratio(3-5years) Child-staff ratio(0-2years) Coefficients Child-staff ratio andgroup size Explanatory variables Canada, 1998 Determinants ofQualityinChildCareCentres, Table 22 A productionfunctionisatechnological,and R surd.463 -squared p < .10 **statisticallysignificantat and statistical significance p t < .05. -statistics of acentremakes apositivecontributiontoquality. time-equivalent childheldconstant,thenonprofitstatus characteristics heldconstant,andwithrevenuesperfull- classroom caregivers. effects, ontopoftheeffectsduetoeducation education levelofthecentre’sdirectorhaspositive development traininghasadditionaleffects).The and substantialinmagnitude(andrecentprofessional effect onqualitythatisbothstatisticallysignificant teacher orcaregiverintheclassroomhasapositive The earlychildhoodeducationortrainingofthemain over arange,thoughnegativeatlargergroupsizes). larger groupsizehasacurvilineareffect—positive nificant onlyforyoungerchildreninthesedata,and classroom matterforquality(althoughtheyaresig- actually doso.Child-staffratioandgroupsizeinthe factors thathavebeenhypothesizedtoaffectquality is notanimpenetrablepuzzle.Infact,mostofthe start. quality inCanadianchildcareservices,theyarea definitive answertothequestionofwhatdetermines and ofchildrenattendingthecentre. overall teachingstaff,ofmanagementandownership, include characteristicsofthedirector, ofthecentre’s especially theleadteacher. Thecentre-specificinputs ronment, childreninthatclassroom,andtheteachers, inputs includecharacteristicsoftheclassroomenvi- and centre-specificinputs.Theclassroom-specific or moreclassrooms,therewillbeclassroom-specific room. Sinceatypicalchildcarecentrewillhavetwo room environmentandofinteractionsintheclass- this quality—thedevelopmentalnatureofclass- factors inthis remaining unexplainedvariation). Cleveland and Hyatt2008forevidenceon possible ty, althoughabouthalfremainsunexplained(see a largeproportionoftheobserved variationinquali- value of.463suggeststhatwe areabletoaccountfor children inchildcarecentres. TheadjustedR-squared together totheproductionofaqualityexperiencefor economies ofscaleintheproductionquality. in thisCanadiansample,suggestingthatthereare centre alsoaffectsqualitywithintherangeofsizes inputs listed.Thenumberofchildrenattendingthe even holdingconstantthevalueofallother significant inimprovingqualitybyabout2percent, Charging anextra$100 permonthisstatistically child isalittleover$500permonthinthesample. Even withalloftheseclassroom,teacherandcentre While theresultsreportedintable22arenota All ofthesefactorscontributeindependentlyand The averagerevenueperfull-time-equivalent 31 They tellusthatproducingqualitychildcare 26 32 Summary We also know, from 1988 data, that a considerable • The child-development quality of nonparental child number of children whose mothers are at home go to care is a central determinant of its effects on children. nursery school, pre-kindergarten, or kindergarten or The evidence from Canadian studies is that quality of attend another form of regulated child care service to care varies considerably across types of care, across provide stimulation, peer-play and parental relief. nonprofit and for-profit care, across regulated and More Canadian children use regulated care than unregulated care, and across larger and smaller either unregulated care provided by a nonrelative or urban/rural areas (i.e., thick and thin markets). nonparental care provided by a relative. There is little Although a sizable proportion of early learning and evidence, in the actual choices of Canadian families, of child care arrangements appear to be of developmen- a rejection of regulated child care. If we include kinder- tal quality, the greater proportion (especially of unreg- garten, of those preschool children with employed ulated arrangements) is probably not of sufficient mothers who receive some nonparental care, 62 percent quality to optimally promote child development. attend primarily regulated care services. Ignoring • Many factors affect the quality of child care. Some kindergarten, the total is still 50 percent. of these are currently regulated (for some types of However, the child development quality of much of care): child-staff ratios, group size, training level Canada’s unregulated and regulated child care does not of teachers/caregivers. Other factors that affect provide the stimulation in supportive and safe condi- quality could also be influenced by policy: regular tions that is recommended by developmental psycholo- professional development training, financial sup- gists. We have too little information on the quality of port to child care programs, status of services Canada’s child care services to be definitive. However, (whether for-profit, nonprofit or public). A first the evidence we do have suggests that although the step would be to sponsor regular on-site evalua- vast majority of regulated care is supportive and safe, it tions of quality in licensed centres and follow-up does not optimally develop children’s cognitive, lan- monitoring of quality improvements. Equally guage, behavioural, social and motor skills. Further, the important, perhaps, would be to lower the cost to small amount of evidence we have about unregulated parents of accessing higher-quality child care. care suggests that its average quality is much lower • One problem with existing research on quality is than the quality of most regulated care. insufficient attention to the cost of quality improve- The cost of full-day regulated care in Canada (outside ments. The National Day Care Study (Ruopp et al. Quebec) is considerably higher than the average amount 1979) examined quality in the context of cost; we spent annually by families with preschool children, and need to recapture that policy-oriented approach to is even higher than the amount spent by families with a

studying quality. Further, at present there are no mother working full-time, full-year. The cost of regulat- New Evidence about Child Care in Canada, by Gordon Cleveland et al. Canadian data that allow us to study the effect of vari- ed care of developmental quality would be higher yet. ations in quality on current and future child outcomes. Average annual expenditures by Canadians living outside Quebec on care for preschool children are $4,500 for all families in which the mother is employed and Conclusions and Policy Directions $6,900 for families in which the mother is employed full-time, full-year. The expenditure is, on average, more he evidence presented in this paper indicates than 20 percent of the mother’s annual earnings. These that nonparental early learning and child care is figures, combined with the evidence on the factors T widely used in Canada. Approximately 80 per- affecting the level of child care spending, many of which cent of preschool children with employed or studying are economic, tell us that affordability of good-quality mothers use nonparental care regularly (Table 1). child care is an issue for Canadian families and therefore Further, the large majority of Canadian mothers of a key policy problem for Canadians interested in chil- young children are employed. At any point in time, dren’s optimal development. over 70 percent of these mothers participate in the Although an increasing number of families purchase labour force to earn income for their families (Roy, early childhood education and care purely for its devel- 2006, Table 1). With entries and exits to the labour opmental benefits, most parents purchase child care to force over the course of the year, the annual participa- support their own employment and study activities. In tion rate is closer to 80 percent (Phipps, 2006). the majority of families, the mother earns less than the father. This, combined with traditional role definitions,

27 IRPP Choices, Vol. 14, no. 12, October 2008 cial burdenofnonparentalcare. time work,shiftwork)inordertodecreasethefinan- ferent employmentsituationforthemother(part- mental elements.Inmanycases,itwillinvolveadif- be purelycustodial,ratherthanincludingdevelop- with notraining)atalowcost.Thecarechosenmay together arrangementsofdubiousquality(caregivers mother’s earningsarelow, thefamilymaypatch greater thefamily’sspendingonchildcare.If that thehighermother’spotentialearnings, choose thebestchildcaretheycanafford.Itappears of carewiththemother’spotentialearnings.Families families, appearstoinvolveacomparisonofthecosts parental childcare. or studysituationthattriggerstheneedfornon- means thatitismostoftenthemother’semployment not employed or studying.Itisalsotruefor 20per- dren. Thisistrue formostchildrenwhose mothersare families donotusenonparental carefortheirchil- of theservices. regarding theaccessibility, affordabilityandquality of supporthasitsadvantagesanddisadvantages care providersandkindergartens.Eachofthesetypes nursery schools,preschools,regulatedfamilychild families; or(4)directfundingofchildcarecentres, existing childcaresubsidysystemsforlow-income the useofnonparentalchildcare;(3)enhancement preschool children;(2)vouchersforfamiliestiedto ber offorms:(1)cashortaxcreditsforparents able forCanadians.Suchsupportcouldtake anum- that wouldmake high-qualitychildcaremoreafford- ing andqualityprovidessupportforpublicpolicies ginal publiccosts. of twoandfivewouldbetwiceasgreatthemar- high-quality childcareforchildrenbetweentheages marginal benefitsofusingpublicmoneytosubsidize study (ClevelandandKrashinsky1998)foundthatthe make itawiseuseoftaxpayerdollars.ACanadian the publicorsocialbenefitsfromthisinvestment investment inthedevelopmentofchildren,andthat lends supporttothenotionthatearlyeducationisan be significantlygreaterthanthecosts.Thisevidence programs, thebenefitsofthishigh-qualitycarecan mounting economicevidencethat,forwell-designed child careisgoodforchildren.Further, thereis There isnowconsiderableevidencethathigh-quality Policies toimproveaffordabilityandquality As aresult,affordabilityofchildcare,formost 1. Cashortaxcreditsforparents. The evidencewehavepresentedaboutuse,spend- Currently, many families. makes itaninefficientandinequitablewaytoassist 2006. However, theparticulardesignofUCCB was theapparentlogicbehindUCCB,createdin refundable taxcreditisanappropriatesolution.This children isthekey objective,thencashpaymentsora sacrifices inordertoprovidecarefortheirchildren. statistical information)aretypicallymakingfinancial employed orstudying(andonwhomwehavesome receive exclusively parentalcarewhilethemotheris in thiscategory. Thosefamilieswhosechildren not needfinancialassistance,butthemajorityare home tolookaftertheirchildrenareaffluentanddo employed orstudying.Someoftheparentswhostay cent to30percentofchildrenwhosemothers finance arestricted rangeofchildcarechoices. the scopeofparental choice,rarelywillavoucher plan ernment subsidizationofservices) isthattheyenhance the key arguments for vouchers(insteadofdirectgov- the rangeofeligiblechildcare services.Becauseoneof details ofthepolicymatterenormously. for childcare.However, withvoucher programs,the care options,thisisknownasprovidinga“voucher” the decisiontouseoneofarangeapprovedchild care. quality childcareservices. necessarily increasetheavailabilityanduseofhigh Child Tax Benefithasnostringsattached, itdoesnot provided throughafamilyallowance,UCCBorCanada child carewillprovideincentivestouseit.Sincemoney preschool years.Moneytiedtotheuseofhigh-quality more childrentoattendsuchqualityservicesintheir and educationalquality, itisinthesociety’sinterestfor 17). Inotherwords,ifchildcareisofdevelopmental not theirparentsarewillingorabletopayforit”(p. “something thatallchildrenareentitledto,whetheror for youngchildrenmaybeviewedasameritgood— and Merrigan(2003)suggest,high-qualitychildcare way tomake childcaremoreaffordableAsLefebvre will bewelcomedbythem,itisnotnecessarilyagood incomes —seeBattle2008). financial needratherthantothosewithhigher single-earner families,butespeciallytothosein Benefit (whichcurrentlygoesdisproportionatelyto enhancement oftheexistingCanadaChildTax of thelowerearner)wouldbepreferable,as with risingfamilyincome(andnottheearnings If directingfinancialassistancetoallfamilieswith Vouchers canbeeitherbroad ornarrowintermsof 2. Vouchers tiedtotheuseofnonparentalchild However, eventhoughfinancialsupporttoparents When paymentstoparentsareconditionalon 33 An untaxed benefit,orabenefitdecreasing 28 34 are Typically, as with the Australian Child Care Benefit Richards and Brzozowski (2006) and Riddell (2007) plan, the voucher can be spent on a wide range of suggest that making existing subsidy systems more gen- approved or registered child care services, and the erous is the best way to increase affordability. This seems requirements for approval are relatively easy to meet. to be the direction Ontario is taking with its newly Therefore the voucher is likely, especially initial- revised child care subsidy system, and it does tend to ly, to allow more parents to afford child care. With ensure that the most needy (and, according to Richards this type of voucher scheme, however, parents are and Brzozswki, those who stand to benefit the most) are the primary gatekeepers, determining the quality of first in line. This is a feasible way of increasing afford- nonparental care that their children will receive. ability for some families. However, if early education While parents are generally very good judges of the and care is beneficial for children (i.e., a merit good), this needs and characteristics of their children, it is diffi- type of reform does not provide any incentives to cult to judge the child-development characteristics encourage attendance by a large number of children of care situations. And, because quality is difficult who could and should benefit. McCain, Mustard, and to judge, child care providers, particularly those Shanker (2007) cite evidence that 70 percent of vulnera- offering care on a commercial basis, generally have ble children in Canada live in families that are not poor. incentives to claim higher quality than they in fact Vulnerability is measured on scales measuring different deliver (Walker 1992). So, while vouchers will aspects of child development. In a recent IRPP publica- encourage parents to spend more on child care, and tion, Doherty (2007) reviews the evidence with respect to while they do offer parental choice, they are not a targeted versus universal early childhood programming, particularly good way to ensure that high-quality and concludes that universal programs, if quality is suf- child care is both more affordable and more widely ficiently high, are the better alternative. used (for opposing perspectives, see Krashinsky The other difficulty with targeted child care subsi- 1986; Levin and Schwartz 2007). dies is that they often strongly discourage parents from In Canada, we do not entrust the quality of medical earning additional income (Cleveland and Hyatt 1998). care to the medical knowledge of parents furnished This is particularly true for subsidies with high benefit- with a voucher. Similarly, we do not entrust the quali- reduction rates — in most provinces, child care subsidy ty of primary, secondary or tertiary education to the is reduced at a rate of $1 for every additional $2 ability of parents to judge that quality and punish earned, whereas in the new Ontario scheme the benefit- low-quality providers via their market choices. Since reduction rate starts at about $1 in $10 and rises to the quality of early education and care has such a about $3 in $10. vital effect on children, it is not at all obvious that 4. Direct operating funding provided (on the supply parent regulation of quality through a market mecha- side) to child care centres, nursery schools, preschools, New Evidence about Child Care in Canada, by Gordon Cleveland et al. nism will achieve social policy objectives in the early regulated family child care providers and any other reg- childhood development field.35 Further, government ulated child care service. In an introductory economics regulation of quality combined with a voucher system course, one learns, as received wisdom, that there is no has its own problems; unless all forms of care are difference in the effect of subsidies provided on the equally regulated, government regulation will drive up demand side of the market (i.e., to consumers) and the prices of the best-regulated types of care, encour- those provided on the supply side (i.e., to producers). aging parents to spend their vouchers elsewhere. This may be true in a perfectly competitive market for a 3. Enhancement of existing child care subsidy sys- good whose characteristics are either uniform across tems. Canada’s child care subsidy systems (which different producers or easily judged and evaluated by have their origins in the Canada Assistance Plan of consumers, and in which there are no institutional lim- the 1960s) are modified voucher programs targeted at itations on the supply response of service providers. lower-income families (typically, with an employment But when these conditions are not in place, subsidizing or training requirement for parents). All provinces demand and subsidizing supply do not necessarily have and territories36 have a child care subsidy system that the same effects (Cleveland and Hyatt 1997). For is largely restricted to the use of regulated and instance, Romer (2000) provides evidence that demand- licensed services. In most jurisdictions the subsidy side research and development subsidies will not rules are not very generous, so only small numbers of increase the numbers of scientists and engineers avail- two-parent families, and not all single-parent fami- able to work in industry, whereas conditional subsidies lies, have incomes low enough to qualify.37 on the supply side will do so.

29 IRPP Choices, Vol. 14, no. 12, October 2008 side ofthemarket. fied ways,ifthefundingisprovidedonsupply or improvethequalityofservice,andinspeci- that theoperatinggrantbeusedtolowerpriceof conditions ontheproducerofservice,suchas given totheconsumer. Anditiseasiertoimpose the subsidy(oratleastrighttoavoucher)is generally simplerwhenthedifferentialamountof income oraccordingtoachild’sspecialneedsis er side.So,subsidizingfamiliesaccordingtofamily easier toimpose,monitorandenforceonthesuppli- the consumersideofmarket whileothersare monitor andenforcewhensubsidiesareprovidedon or limited,andsomeconditionsareeasiertoimpose, Human Development [NICHD]2005;NICHD and Merrigan 2003; NationalInstituteofChild Healthand Zaslow 1990;Lambetal.1998; Lefebvreand important factorinachild’s life(Hayes,Palmer, and of studieshavefoundthatparents arethemost parents arethemainsource of thiscare;awiderange pensable tochilddevelopment.Formostchildren, Warm, loving,stimulating, empathetic careisindis- Complements, notsubstitutes isons, relyheavilyonsupply-sidefunding. the highest-qualitychildcareininternationalcompar- The Nordiccountries,whichareacknowledgedtohave child careservicestake theformofoperatinggrants. may insistthatasignificantportionoftheirfunding providers topositivelyinfluencequality, governments and sincesupply-sidefundingincreasestheabilityof child careservices,intermsoftheireffectsonchildren, op. Sincequalityisthefundamentalcharacteristicof strongly influencethenatureofservicesastheydevel- side andthoseprovidedonthesupplyside. difference betweensubsidiesprovidedonthedemand better buthard-to-observequality, therewouldbea favour supplierswithlowerpricesover If, intheirpurchasingbehaviour, parentswereto by government,affordsthegovernmentlessleverage. on thedemandside,allocatedbyparentsratherthan and reportingstandards)moderatecosts.Funding improve quality(meetspecificoperating,monitoring use theleveragegainedtocompelprovidersboth to regulatedprovidersisthatwisegovernmentscan to parents.However, therealvirtueofdirectassistance care providerscanservetolowertheamountcharged assess. Providingoperatingfundingdirectlytochild In truth,subsidiesorgrantsareoftenconditional Direct fundingtoproviderspermitsgovernments The qualityofchildcareservicesisdifficultto 38 39 affordable forfamilies. care thatarestimulatingtochilddevelopmentmore ways tomake thosetypesofearlylearningandchild for high-qualitynonparentalcare,weneedtofind either/or than assubstitutes.Theissueisincreasinglynot parental careandchildascomplementsrather cant portionsofeachweek.We neednowtoview are overwhelminglyinnonparentalcareforsignifi- mothers areinthelabourforcetostay, andchildren The geniecannotbepushedbackintothebottle— uncovered, isthatthesesentimentsaremisdirected. ernments toinsteadencouragechildcarebyparents. care willtipthebalanceevenfurther;theywantgov- ernment supportofregulatedearlylearningandchild ment inthenextgenerations.Theyworrythatgov- effects oncognitive,socialandbehaviouraldevelop- reduced parentaltimewithchildrenwillhavenegative fear thatthegrowthofnonparentalchildcareand amount ofparentaltimespentwithchildren.They nonparental childcare,theapparentdeclinein employment ofmothers,theincreaseinuse the changesinCanadianfamilies,increase Barr, andPeters 2004).SomeCanadiansworryabout Duncan 2003;Shonkoff andPhillips2000;Tremblay, with theirchildren. realistic waystohelpallparentsspendmoretime quality familylife,andviceversa.We needtofind ways tomake paidworkmoreconsistentwithhigh- Our view, basedpartlyonthenewfactswehave but both/and 30 40 And, giventhewidespreadneed . We needtohelpfamiliesfind Appendix A In fact, different parents probably have different Misinterpretations of Child Care interpretations of how the NLSCY questions apply to their situation.41 According to the above reports, a child Data from the NLSCY is in child care (i.e., nonparental care) if the parent has answered yes to the first question above. This means he National Longitudinal Study of Children and that, if a family with a stay-at-home parent and using Youth is the only survey that provides current preschool for their child has answered no (because the data on child care use; it collects information on T preschool care was not work-related), the child is child care used to support employment or study. The recorded as not using child care. NLSCY asks the reporting parent While you (and your Unfortunately, many families with a parent at home that spouse/partner) are at work/studying, do you currently regularly use care by a relative, a nanny, a nursery school use child care such as daycare, babysitting, care by a or some other form of child care for their preschool children relative or other caregiver or a nursery school? The fol- may not have indicated this in the NLSCY survey, because low-up question is While you (and your spouse/part- of how the questions are phrased. And, because the NLSCY ner) are at work or studying, which of the following asks about child care used while parents are working or methods of child care do you currently use? A list of studying, a significant proportion of families with all par- types of nonparental child care follows this question. ents working or studying may not have indicated use of One might think, from the phrasing, that only child care outside of parental working hours. families in which all parents are employed or study- Yet some researchers (Bushnik 2006) have concluded ing would answer the child care questions in the that only a bare majority (54 percent) of all children 6 NLSCY. This is not the case. All parents are eligible to to 71 months of age used nonparental care in 2002-03 answer them; in fact, a large number of parents in and that as recently as 1994-95 only 42 percent of families where one of two parents is not employed them used nonparental care. have answered these child care questions. As a result, The true percentage of all preschool children using Statistics Canada (Statistics Canada 2005; Bushnik child care is almost certainly much higher than these 2006) has assumed that data from these questions estimates. We would argue that the figure must be at apply to, and have been correctly supplied by, all least 70 percent of all children 6 to 71 months of age, Canadian families, not just those families with all but the NLSCY does not collect data that would allow parents working or studying. us to calculate the exact current percentage.

Table A1

Number of Preschool Children Using Nonparental and Parental Care, Excluding Kindergarten, All Families New Evidence about Child Care in Canada, by Gordon Cleveland et al. Regardless of Employment Status, Canada, 1988

% of children in Type of child care N % of all children nonparental care

Nonparental child care

Nonrelative, outside the child’s home 344,100 17.4 26.9

Nonrelative, in the child’s home 177,100 8.9 13.8

Relative, outside the child’s home 255,000 12.9 19.9

Relative, in the child’s home 182,600 9.2 14.3

Daycare centre 166,300 8.4 13.0

Other, including nursery school and preschool 154,800 7.8 12.1

Total nonparental care 1,279,900 64.6 100

Parental care 701,400 35.4

Total 1,981,300 100

Source: Statistics Canada, CNCCS, 1988. Note: Only primary (i.e., most number of hours per week) type of nonparental care used is shown. Variables are defined so as to be comparable to NLSCY data set.

31 IRPP Choices, Vol. 14, no. 12, October 2008 did 82.9percent ofchildreninsingle-parent 03 (upalmost 5 percentagepointsfrom1994-95), as employed orstudyingused nonparental carein2002- percent ofchildreninfamilies withbothparents the useofkindergartenasa formofchildcare,70.9 parental careonaregularbasis. Evenifweignore employed orstudyingoverwhelminglyreceivenon- data: childreninfamilieswhichbothparentsare conclusion that,inouropinion,isindicatedbythe tracted publicandresearchattentionawayfromthe NLSCY), 65percentofchildrenusednonparentalcare. of nonparentalcarearrangementsasusedinthe the parents(andbasedonexactlysamedefinitions families, regardlessoftheemployment/studystatus by children6to71 monthsoldin1988.Across all mum estimateofusein1994-95orlater. estimate ofchildcareusein1988mustbeamini- child carehasgrowninCanadasince1988,soany is reasonabletoassumethatuseofearlylearningand assessing therecentestimatesbasedonNLSCY. It nearly 20yearsold,theyprovideabenchmarkfor by StatisticsCanada.Althoughthedataarenow 25,000 children0to12yearsofage,werecollected tary childcare.Thedata,basedonasampleofover designed tomeasureuseofallformssupplemen- 2002-03 arelow. TheCNCCSwasspecifically using nonparentalcarein1994-95and54percent Statistics Canada’sestimatesof42percentchildren hlrn( f(0)(4 3)(0)(7 (25) (47) 13 13,706 22 25,534 27 (100) 27 34 11 53,707 85 41 71 15 22 17 26 16 (36) 28 $40,000+ 26 $30,000-$40,000 122,522 $20,000-$30,000 30 $15,000-$20,000 13 subsidized $10,000-$15,000 11 (64) Childrenin 82 14 4 categoryin 216,217 2 income n.s. 76 1 income 66 n.s.: Too fewobservationstobestatisticallysignificant. (100) 340,339 49 Source: ReproducedfromLefebvre(2004,55).Author’s calculationsbasedonNLSCYmicro-data,cycle4. Childrenin 22 49 40 total) n.s. children (%of income 27 Number of Family 11 subsidized 14 Childrenin $80,000+ 16 $60,000-$80,000 categoryin 8 $50,000-$60,000 2 income $40,000-$50,000 categorychild $30,000-$40,000 income $20,000-$30,000 Childrenin $10,000-$20,000 care category income Family Children Aged1-4Years, byFamilyType andIncomebyUseType ofCare,Quebec,2000-01 (percent) Table A2 Regulated family-basedandcentre-basedchildcare. Family currentlyusingchild-careservicesforworkorstudy. This interpretationoftheNLSCY datahasalsodis- Table A1showsestimatesfortheuseofchildcare The 1988CNCCSgivesusreasontobelievethat ntoprn aiisInsingle-parentfamilies In two-parentfamilies hlrni Childrenin Children in 1 child care 2 age pointsfrom1994-95)(Bushnik2006,49). employed orstudentfamilies(upjustover5percent- is carefultorestrict one’sanalysistofamilies inwhich sions aboutissues relatedtochildcare.But unlessone “crunched” thenumbersanddrawnanalyticalconclu- social scientistshavetaken theNLSCY datasets, lic debateoverthelastseveral years. NLSCY findingsonchildcaremayhaveskewed pub- one parentstaysathome).Inotherwords,the (and wouldalsoprovidesupporttofamiliesinwhich and affordabilityofchildcareisgoodpublicpolicy may believethatimprovingthequality, accessibility form ofnonparentalchildcare,thengovernments ent isathomehavealreadydecidedtousesome substantial percentageoffamiliesinwhichonepar- families inwhichallparentsworkorstudy, anda majority ofallparents,anoverwhelming parent-only care.If,ontheotherhand,alarge the largenumberofparentswhoobviouslyprefer improve nonparentalcarewoulddiscriminateagainst could perhapsarguethatgovernmentactionto sibility andaffordabilityofnon-parentalcare.One by providingprogramstoimprovethequality, acces- argue thatgovernmentsshouldnottipthebalance decided tousenonparentalchildcare,onecould is political.IfonlyhalfofCanada’sparentshave pretation oftheNLSCY data onchildcare.Thefirst There areatleasttwoproblemscausedbymisinter- Misinterpretations basedontheNLSCYdata aeoyctgr cidcr childcare childcare category category The secondproblemisanalytical. Anumberof 32 2 all parents are employed or studying, any conclusions ing columns 2 and 4 and columns 6 and 8. For two-par- drawn from the NLSCY data could be misleading. ent families, we can see that while 8 percent of families This apparently is what happened when Professor have an income between $20,000 and $30,000, only 4 Pierre Lefebvre (2004) used the NLSCY to determine percent of the users of subsidized child care are in this whether child care spaces under the Quebec child care range. On the other hand, while 22 percent of families reforms of 1997-2000 were used equally by families in have an income over $80,000, 28 percent of children in different income classes. Lefebvre’s main conclusion subsidized care are in this range. (shown in table A2), since widely repeated as a major A similar pattern appears for single-parent families. criticism of Quebec’s child care policy, was that afflu- While 26 percent of all children in these families are in ent Quebecers were disproportionately likely to benefit the $10,000-$15,000 income range, only 11 percent of from the $5-per-day policy because they used child those in subsidized care are in this range. On the other care services in greater proportion than their demo- hand, 16 percent of children are in the $15,000- graphic weight would warrant; those in lower income $20,000 income range, whereas 27 percent of children classes were disproportionately likely either to not use in subsidized care are in this class. child care at all or to use non-subsidized forms of care. Thus in Lefebvre (2004) the evidence on unequal The data in table A2 are from Cycle 4 of the access among children in single-parent families is not NLSCY. The table provides information on children in very strong, except for single-parent families with two-parent and one-parent families, showing in exceptionally low incomes. There is no consistent pat- columns 2 and 6 the percentage of children from fam- tern of increased access for children in more affluent ilies in each income category. So, for instance, we see single-parent families. However, the evidence on access that 2 percent of all the children in two-parent fami- by children in more affluent two-parent families is clear. lies in Quebec have an annual family income between Lefebvre’s conclusions are, inadvertently, based on $10,000 and $20,000 (column 2), and 26 percent of all faulty data. The information on children using non- children in single-parent families have a family parental and $5-per-day care comes from the NLSCY. income between $10,000 and $15,000 (column 6). Since the NLSCY data on child care use are valid only Not all of these children receive nonparental care, for families in which all parents are employed or study- of course. Columns 3 and 7 show us what percentage ing, we must restrict our attention to these families. In of children in each income category receive non- table A3, we recalculate the evidence from the NLSCY parental care, according to the NLSCY. Generally doing exactly that. speaking, the percentage using any kind of non- This recalculated table provides no clear evidence parental child care rises as family income rises. that subsidized child care spaces are accessed in a dis-

Columns 4 and 8 show us what percentage of chil- criminatory fashion. This conclusion was a statistical New Evidence about Child Care in Canada, by Gordon Cleveland et al. dren in each income class have been able to gain artifact of using inappropriate data from the NLSCY (in access to Quebec’s $5-per-day child care (i.e., regulat- a way facilitated by Statistics Canada). Table A3 shows ed family-based and centre-based care). that the percentage of two-parent families in each The “proof” that the allocation of Quebec’s subsi- income class is closely matched by the percentage dized child care has been unequal comes from compar- using subsidized child care. The percentages among

Table A3 Children 0-4 Years with Employed and Studying Parents, by Income Category and Participation in $5-per- Day Care, Quebec, 2000-01 (percent)

Two-parent families Single-parent families

Family income Children in Children in Children in Children in category income category subsidized care Income category income category subsidized care

$10,000-$30,000 4.9 4.8 $10,000-$15,000 16.2 12.8 $30,000-$40,000 10.9 11.8 $15,000-$20,000 20.4 28.2 $40,000-$50,000 14.8 11.8 $20,000-$30,000 24.1 27.0 $50,000-$60,000 14.0 12.7 $30,000-$40,000 25.0 19.9 $60,000-$80,000 26.3 27.5 $40,000+ 14.3 12.1 $80,000+ 29.1 31.3 Number of children 190,474 88,455 23,578 11,811

Source: Statistics Canada, NLSCY Cycle 4 (2000-01).

33 IRPP Choices, Vol. 14, no. 12, October 2008 trate thepossibility ofdrawinginappropriate conclu- higher-income families. child careuserstobeatleast slightlytiltedtowards than part-time,wewouldexpect thedistributionof work full-timeandearn incomesrather Quebec childcareprogramistopermitmothers for two-parentfamilies.Sinceanobjectiveofthe single-parent familiesandnotdramaticallydifferent access toregulatedchildcare,butitisveryclosefor of allfamilieswithanemployedmotherhaving mother is,therefore,notidenticaltothedistribution have anincomeatthislevel. all childrenintwo-parentfamiliesregulatedcare family incomeabove$80,000,whereas41.3 percentof two-parent familieswithanemployedmotherhavea a similarincome.Andonly37.4percentofchildrenin children intwo-parentfamiliessubsidizedcarehave a familyincomebelow$40,000,and10 percentofall in two-parentfamilieswithanemployedmotherhave are notdramatic.Forinstance,12percentofchildren and 3alsocolumns56),butthedifferences lies inhigherincomebrackets (comparingcolumns2 evidence intableA4oftiltingaccesstowardsfami- NLSCY cycle—Cycle6,from2004-05.Thereissome 01. We canalsolookatdatafromthemostrecent NLSCY data.Lefebvre’soriginalfiguresarefrom2000- icum ofsupportforhisconcernsusingmorerecent above whoare(slightly)disadvantaged. dized childcareandfamiliesearning$30,000 are disproportionatelyabletogainaccesssubsi- families earningbetween$15,000and$30,000who single-parent familiesarenotquitesoclose,butitis ubro hlrn204214512,2 17,650 28.9* ** ** 27.6* 15.1* 25,420 28.3* 15.8* 13.8* 22.8* 19.2* $40,000+ 41.3 $30,000-$40,000 $10,000-$15,000 $20,000-$30,000 $15,000-$20,000 25.6 134,511 12.2 2.9* 10.9 7.1 37.4 220,412 ** Estimatecannotbereported(cellcountunder5). 23.8 * Estimateisconsideredunreliablebecauseoflownumbers(cellcountunder30). 14.1 Source: StatisticsCanada,NLSCYCycle6(2004-05). 4.2 12.8 7.8 Number ofchildren $80,000+ $60,000-$80,000 $50,000-$60,000 $40,000-$50,000 $30,000-$40,000 $0-$30,000 Day Care,2004-05(percent) Children 0-4Years withEmployedandStudyingParents, byIncomeCategoryandParticipation in$7-per- Table A4 aeoyicm aeoysbiie aectgr noectgr subsidizedcare Childrenin incomecategory Childrenin category Familyincome subsidizedcare Children in incomecategory Childrenin category Family income In anycase,the pointofthisexercise istoillus- The distributionofallfamilieswithanemployed To befairtoProfessorLefebvre,wecanfindamod- w-aetfmle Single-parentfamilies Two-parent families analyzing childcareusebyallfamilies. studying, buttheyare,wewouldargue,invalidfor ful forfamiliesinwhichallparentsareemployedor NLSCY’s childcaredata.Thedataarevalidanduse- sions byusinganinappropriateinterpretationofthe 34 Notes 8 See Appendix A for an explanation of why NLSCY data 1 Other milestones include the Special Parliamentary cannot be used to generate estimates of the use of child Committee on Child Care of 1987, the Mulroney gov- care by all preschool children in Canada. ernment’s failed child care bill of 1988, the Quebec 9 The NLSCY is a longitudinal and cross-sectional probabil- child care reforms of 1997-2001, the Multilateral ity survey designed to gather information on children Framework Agreement on Early Learning and Child and youths in Canada. It is a joint project of Statistics Care of 2003, the 2004 OECD Country Note on Early Canada and Human Resources and Skills Development Childhood Education and Care in Canada and the bilat- Canada. The survey covers a range of topics, including eral early learning and child care agreements between child care for employed or studying parents, children’s the Martin government and all 10 provinces in 2005. physical development, learning and behaviour, and chil- 2 See box 1 for a discussion of the terminology used in dren’s social environment (family, friends, schools and this paper. communities). The first cycle of data collection took place 3 There is insufficient evidence from direct measure- in 1994-95 and data collection continues biennially. The ments to form a judgment about Canadian kinder- unit of analysis for the NLSCY is the child or youth. gartens (Cleveland et al. 2006; see also Johnson and 10 The data do not include children in the northern territo- Mathien 1998), but there is a widespread perception ries nor children in institutions or army camps. Children that kindergarten positively affects child development. of recent immigrants are underrepresented because they 4 In Quebec, families receiving social assistance have cannot be part of the cohort recruited in 1994-95. access to 23 and a half hours per week of free regulated 11 To be precise, children from 6 to 71 months inclusive child care; in addition, full-time care is available (since and not yet in Grade 1. This age range mimics that used 2002) free of charge for children of families in disad- in Statistics Canada’s publications (Bushnik 2006; vantaged neighbourhoods who are referred by the local Statistics Canada 2005) and reflects the poor coverage in CLSC (community health and social service centre). the NLSCY sample of children younger than six months. 5 Some commentators argue that the CCED is bad policy 12 This practice in the NLSCY is likely to cause some under- either because it makes child care cheaper for families estimation of the role of parental/family care, and it does with higher earnings than for those with lower earn- not provide breakdown into different forms of ings, or because it gives child care assistance to two- parental/family care. An alternative method, used in the earner families that is denied to single-earner couples. Canadian National Child Care Survey (CNCCS) of 1988, In our opinion, neither of these arguments is valid. If is to consider all types of care used while the usual care- the annual price of child care is $7,000, both low- and giving parent is otherwise occupied, and take the one high-earning families will still have to pay $7,000 for it used for the most hours to be the primary one, whether after the CCED. By recognizing child care as a neces- this is parental or nonparental care. According to this sary expense incurred by earning income (for the sec- method, care provided by one spouse while the other is ond earner in the family), the CCED simply means that employed, and care provided by the usual caregiving this child care can be purchased with untaxed dollars. parent at her workplace, may turn out to be the primary Similarly, the claim that the CCED is unfair to single- form of care used when some form of nonparental care

earner families has no grounds. The potential “second” is needed for only a few hours; the questions asked in New Evidence about Child Care in Canada, by Gordon Cleveland et al. earner in a single-earner family does not, of course, the NLSCY ignore this possibility. See below for CNCCS pay tax on the value of her home production, including data on parental child care. on the value of the child care she produces by looking 13 Regulated care includes centre-based care, nursery schools, after her children during the day. The CCED places the preschools, regulated family-home child care, and (except second earner in a two-earner family in a similar situa- when noted) junior and senior kindergarten. Unregulated tion; as a result of the CCED, she will not pay tax on nonrelative care is care provided by a nonrelative of the the child care she uses during the working day. In tax child, whether within or outside the child’s home. Unreg- terms, the CCED re-establishes horizontal equity in the ulated relative care is care provided by a relative of the treatment of single-earner versus two-earner families, child (but outside the nuclear family), whether within or and in the treatment of work-related child care for low- outside the child’s home. Parent/family care is care provid- and high-income earners (see Vincent and Wooley ed by any member of the child’s immediate family (when 2000, 12-17, for a full discussion). no form of nonparental care is used). In order to group 6 Almost 90 percent of Canadian women who were similar programs together, we include nursery school pro- employed when pregnant return to work within a year grams in all jurisdictions as part of regulated care. In fact, after the child’s birth (Statistics Canada 2005). Longer nursery schools are not currently regulated in Quebec, absences from the workforce can have negative conse- or Territory but are regulated in other quences for future employability and earnings (OECD provinces and territories. 2007; Ruhm 1998). 14 The NLSCY records whether the child attends kinder- 7 The current Ontario government has announced a garten but not the number of hours attended. We use the several-year transition to full-day junior and senior typical number of hours for kindergarten in the province kindergarten (perhaps combined with child care to determine which is the primary form of care for the services). child.

35 IRPP Choices, Vol. 14, no. 12, October 2008 3Forinstance,theNICHD(2001, 2003a)foundevidence 23 Since1990,therehasalsobeenpaidparentalleave 22 Presumably, theseOECDcalculationsassumethatthis 21 Theyear2005wasthelastfullofPaul Martin’s 20 We includetwovariablesrelatedtothepredicted 19 Thepercentagewould belargertoday. In1988,there 18 Regressionsdesignedtoexplaintheuseofregulated 17 Thesamesurveyfoundthat80percentofemployed 16 Forparentsusingnonrelativecareinthefamilyhome, 15 more externalizing behaviouralproblemsand more behaviours. Inkindergarten,they areassociatedwith play, lesssocialcompetenceandmoreexternalizing hours inchildcareareassociated withmorenegative various behaviouralproblems. At 54months,more ciated, at24and54months inkindergarten,with that morehoursinachildcare arrangementareasso- providing zero-pricedcare. available toeligiblefamilies,sothefathercouldbe more generousonebasedonincomeassessment. transformed fromonebasedonneedsassessmenttoa 2001, the childcaresubsidysysteminOntariohasbeen single parentisnoteligibleforchildcaresubsidy. Since parental decisionsabouttheappropriateexpenditure. child developmentaffectedeithersupplierpricesor discussion ontheimportanceofqualitycareinfostering regulated childcare.Perhaps thegreatamountofpublic ments todramaticallyexpandtheavailabilityof ritorial governmentswerenegotiatingindividualagree- prime ministership,whenthefederal,provincialandter- decision touseparentalcare. mothers’ wageshaveanon-linearrelationshiptothe wage. Thisformulationallowsustodeterminewhether hourly wage—theitselfandsquareof was limitedto15weeks. ployment InsuranceProgramandthematernitybenefit was noparentalbenefitavailableundertheUnem- Quebec, becauseofQuebec’suniqueprograms. care. Theseregressionsexcluded preschoolchildrenin important factorsindeterminingtheuseofregulated child inthefamily(negative)allareindependently children inthefamily(negative)andbeingyoungest income, locationinalargeurbancentre,numberof sity education,mother’sincome,otherhousehold number ofworkinghours,single-parentstatus,univer- 6 oftheNLSCY, confirmthatthechild’sage,mother’s forms ofchildcare(includingkindergarten),usingCycle public policy. Dreams” butthatitisnotnecessarilyausefulguideto ensure thatthesurveydoesindeeddeliver“Hopesand such variancewiththecurrentCanadianrealityasto vey. Theseresponsesabout“idealarrangements”areat includes 85percentofthemenrespondingtosur- own childreniftheycouldaffordto.Thatfigure parents withchildrenwouldstayhomeandraisetheir estimate thenumberofchildreninregulatedcare. result, itislikely thattheseNLSCY dataslightlyover- regulated andmaygiveanincorrectresponse;asa Some parentsdonotknowwhethertheircaregiveris the NLSCY surveyaskswhetherthiscareisregulated. 8FortheYou BetICare!study, researchersdrewastrat- 28 Groupsofchildrenfrom some650establishmentswere 27 26 For-profit Forinstance,theNICHD-ECCRN(EarlyChildCare 25 Forinstance,Magnuson,Ruhm,andWaldfogel (2004) 24 plus information onthe“observed”staffmember for variables. Dataonthequalityin alltheseclassrooms, data onthedependentvariable andkey explanatory Yukon Territory). Atotalof325classroomshadusable Saskatchewan, Ontario,Quebec, NewBrunswickand provinces andoneterritory(British Columbia,Alberta, includes datafrom234childcare centresinsix ed) andnotlocatedonaNativereserve.PhaseII ated byamunicipality(municipalcentresareexclud- per day, in operationforatleast12months,notoper- providing careforpreschoolchildrenatleastsixhours areas insixprovincesandoneterritory, ofallcentres ified randomsample,inseveralurbanandsuburban cussed inthefollowingsection). ilar indesigntotheYou BetICare!studyof1998(dis- local population.Broadlyspeaking,thestudywassim- centre, thecentreitself(includingitsfinances)and classroom, thesupervisorofchildcareservicesat tion, datawerecollectedonthemainteacherin internal consistencyandinterraterreliability. Inaddi- intensive 10-day training,withappropriatetestsof the InstitutdelaStatistiqueduQuébecandunderwent reforms inQuebec.Qualityobserverswerechosenby data onqualityasdefinedbythenewchildcare using aninstrumentspecificallydesignedtocollect site inspectionsofthequalitycareinthesefacilities, providers. Attheheartofstudywereday-longon- chosen atrandomfromalistofchildcareservice CPEs withregardtotrainedstaff. profits werenotsubjecttothesameregulationsas subsidized servicesat$5(now$7)perday. Thesefor- ties operatingoutsidetheCPEsystembutproviding and unregulatedfamilycaresettings. Galinsky etal.(1994)evaluatedcareinbothregulated including carebyrelatives(NICHD,variousyears). different formsofregulatedandunregulatedcare, Research Network)studiesusedaninstrumentvalidfor 2006). Hausfather etal.1997;Wylie, Hodgen,andThompson ence inearlychildhoodeducation(Andersson1992; associations betweenchildoutcomesandlengthofexperi- qualifications. Anumberofstudieshavefoundpositive ratings whenchildrenattendedcentreswithhigh-level (Sammons etal.2003)foundmorepositivebehavioural that childrenattended.AstudyconductedinEngland garten programswerelocatedinthesamepublicschools found nonegativebehaviouraleffectswhenpre-kinder- report similarnegativebehaviouraleffects. Milligan 2005;Magnuson,Ruhm,andWaldfogel 2004) these influences.Otherstudies(Baker, Gruber, and cant andquantitativelyimportantaftercontrollingfor received bythechild,butarestillstatisticallysignifi- the qualityofchildcareandparenting teacher-child conflict.Theseeffectsaremoderatedby 36 garderies in Quebecarecentre-basedfacili- each classroom, other staff members in the centre, the ments. Except for voucher schemes targeted at low- director and centre characteristics, are available in the income families, these quality controls are perfunctory — data set. For more details about You Bet I Care! study, they exclude only those facilities that are dangerous for see Doherty et al. (2000a and 2000b); Goelman et al. children, instead of including only those that provide (2000). care of educational/developmental quality. See Cleveland 29 The seven assessed areas are: space and furnishings and Hyatt (1997) for a general discussion. (including the physical setup of the room; personal care 36 Quebec is an exception here. Quebec makes $7-per-day routines (such as greeting, meals, snacks, diapering, regulated child care available to families independent of health practices and departure); language-reasoning their work/study situation and their income status. There experiences; activities (such as fine motor activities, art, is special fee relief for families on social assistance and music, movement, blocks, sand, water, dramatic play, for children recommended by the local CLSC (community nature and science, mathematics and numbers, comput- health and social service centre). ers, and diversity); interaction (including gross motor 37 Historically, child care subsidy systems have given full activities, supervision of children, discipline, staff-child subsidies to families with a total income of about interactions, and interactions amongst children); pro- $15,000 annually who are employed or in training. If the gram structure (including scheduled time, free play annual price of child care were $8,000, eligibility for a time, group time, and provisions for children with dis- subsidy would fall to half at a family income of $23,000 abilities); and parent and staff needs (communication and disappear entirely at $31,000 (see Cleveland and with and involvement of parents, personal and profes- Hyatt 1997; Kershaw 2007). sional needs of staff and opportunities for professional 38 Demand-side voucher systems typically involve some growth, staff interaction and cooperation, and supervi- quality assurance measures too, but generally these do sion and evaluation of staff). little more than identify facilities that might be danger- 30 On a scale of 0 to 100, 33 is minimal quality, 67 is ous to children. For instance, Australia’s Quality good and 100 is excellent, corresponding to scores of Assurance system (a form of accreditation) finds that 3, 5 and 7 on the original ITERS or ECERS scales. The only about 7 percent of facilities are below acceptable ITERS and ECERS (and their subscales) are, by con- quality levels. Empirical measures of child development struction, treated as cardinal scales. The distances quality in the United States and Canada find that more between inadequate and minimal, between minimal than half of facilities fail to provide services that stimu- and good, and between good and excellent are consid- late child development. Visits by one of the present ered equivalent distances. authors to child care centres in Australia give us little 31 Other publications in Canada (Beach and Friendly reason to believe that the distribution of quality levels in 2005; Cleveland and Hyatt 2008; Cleveland et al. 2007; Australian “long day care” is very different. Doherty 2007; Friendly and Beach 2005b; Friendly, 39 The alternative point of view, that demand-side funding Doherty, and Beach 2006; Goelman et al. 2006) and fosters competition and efficiency in public services, can elsewhere (Blau 1997, 2000; European Commission be found in Lundsgaard (2002). Network 2004; Helburn 1995; NICHD 1996a, 2000; 40 Sweden’s flexible parental benefits plan provides a

Ruopp et al. 1979) provide additional insights and model. Parents have the option, with cooperation from New Evidence about Child Care in Canada, by Gordon Cleveland et al. qualifications. their employer, to use a portion of parental benefits to 32 See Cleveland et al. (2007) for a discussion of differ- reduce hours of work over the first eight years of the ences in the effects of nonprofit status in thin and child’s life. Parents can decide to work six hours a day thick child care markets — child care markets with dif- instead of eight or four days a week instead of five, in ferent concentrations of population and potential order to spend more time with their young child. Parental demand. benefit is used to compensate for lost salary. 33 The UCCB puts money ($1,200 per child annually) in 41 This is discussed in Bushnik (2006, 39). the pockets of families with children under six years of age. Because this benefit is taxable in the hands of the lower-earning spouse, the actual amount of money that parents get to keep is less than $1,200, and sin- gle-earner families (especially those with higher incomes) tend to benefit more than two-earner or sin- gle-parent families. See Battle, Torjman, and Mendelson (2006) for an analysis. 34 The exceptions tend to occur when the voucher is tar- geted at low-income families (presumably because of the belief that, for these families, quality of services is more important than parental choice). 35 Sensitive to this argument about the ability of parents to regulate quality, some voucher schemes include quality assessment as part of their eligibility require-

37 IRPP Choices, Vol. 14, no. 12, October 2008 Blau, D.1997.“TheProductionofQualityinChildCare Battle, K.,S.Torjman, andM.Mendelson.2006. Battle, K.2008 Baker, M.,J.Gruber, andK.Milligan.2005.“Universal Andersson, B-E. References Bibby, R.W. 2004. Beach, J.,andM.Friendly. 2005. _____. _____. _____. Cleveland, G.,andD.Hyatt.1997. Cleveland, G.,B. Forer, D.Hyatt,C.Japel,andM. Cleveland, G.,C.Corter, J.Pelletier, S.Colley, J.Bertrand, Brayfield, A.,andS.Hofferth.1995.“BalancingtheFamily Bourgon, L.,andC.Lavallée.2003. _____. Bushnik, T. 2006. Centres.” Institute oftheFamily. Canadian HopesandDreams. Institute ofSocialPolicy. $5,000 CanadaChildTax Benefit National BureauofEconomicResearch. Being.” Working Paper 11832. Cambridge,MA: Childcare, MaternalLaborSupplyandFamilyWell- Swedish Schoolchildren.” Socioemotional CompetenceofThirteen-Year-Old Research Unit,UniversityofToronto. Environments Ottawa: CaledonInstituteofSocialPolicy. a NameChange:TheUniversalChildCareBenefit Mimeo. Toronto: University ofToronto atScarborough. Care andEducation: DoWe KnowtheKeyIngredients? for ResearchonPublicPolicy. Governments?” Families: AGoodBargainforCanadian Resources DevelopmentCanada. 7E, AppliedResearchBranch.Ottawa:Human Provide ChildCareAssistance? or SubsidiestoProviders:HowShouldGovernments Final ProjectReport. Early LearningandChildCareServicesinCanada: Current andFutureRoleofNonprofitProvision Krashinsky. 2007. Programs. Child Care,KindergartenandFamilySupport Field ofEarlyChildhoodLearningandDevelopmentin and J.Jamieson.2006. Series. Ottawa:StatisticsCanada. 599MIE-003, ChildrenandYouth ResearchPaper Structure.” Race/Ethnicity, EconomicStatusandFamily Budget: DifferencesinChildCareExpendituresby Gouvernement duQuébec. 2003 Version utiliséedansl’enquêteGrandirenqualité la qualitéeducative:leservicedegardepréscolaire. Science Centres: AnotherLook.” 2008. 1998. “SubsidizingChildCarefor Low-Income 2000. “TheProductionofQualityinChildCare . Quebec:MinistèredelaFamilleetl’Enfance, 4 (3):136-48. The RecipeforGoodQualityEarly Childhood Journal ofHumanResources . ABiggerandBetterChildBenefit: Ottawa: CanadianCouncilonLearning. 1992. “EffectsofDay-CareonCognitiveand Social ScienceQuarterly Child CareinCanada. The FutureFamiliesProject:ASurveyof . Toronto: ChildcareResourceand IRPP Choices An EconomicPerspective onthe Toronto. A ReviewoftheState Applied Developmental Child Development Child CareCentrePhysical 4 (2).Montreal:Institute Subsidies toConsumers Ottawa: TheVanier Échelle d’observationde Research Paper R-97- . Ottawa:Caledon Catalogue 89- 76 (1):158-77. 32: 354-87. More than 63: 20-36. . Goelman, H.,B. Forer, P. Kershaw, G.Doherty, D.Lero,and Goelman, H.,G.Doherty, D.S.Lero,A.LaGrange, andJ. Galinsky, E.,C.Howes,S.Kontos,andM.Shinn.1994.“The Friendly, M.,G.Doherty, andJ.Beach.2006. Friendly, M.,J.Beach,C.Ferns,andM.Turiano.2007. _____. Friendly, M.,andJ.Beach.2005a. European CommissionNetworkonChildcareandOther Drouin, C.,N.Bigras,C.Fournier, H.Descroisiers,andS. Doherty, G.,D.S.Lero,H.Goelman,J.Tougas, andA. Doherty, G.,D.S.Lero,H.Goelman,A.LaGrange,andJ. Doherty, G.2007. “EnsuringtheBestStartinLife:Targeting _____. Cleveland, G.,andM.Krashinsky. 1998. A. LaGrange.2006. “Towards aPredictiveModelof Well-Being, UniversityofGuelph. Care! Quality inChildCareCentresacross Canada.You BetI Tougas. 2000. Young Children Care —KeyFindingsandPolicy Recommendations.” Study ofChildreninFamilyChildCareandRelative Toronto. Childcare ResourceandResearchUnit,Universityof and WhatDoWe Think?ALiteratureReview Know AboutQualityinEarlyLearningandChildCare, University ofToronto. Toronto: ChildcareResourceandResearchUnit, Childhood EducationandCareinCanada2006. and ResearchUnit,UniversityofToronto. and ChildCareSystem and ResearchUnit,UniversityofToronto. and CareinCanada2004 and ResearchUnit.(Originallypublishedin1996.) and ChildCareServices. of MenandWomen. 2004. Measures toReconciletheEmploymentResponsibilities ifs québecoise surlaqualitédesservicesdegardeéducat- Bernard. 2004. and Well-Being, UniversityofGuelph. You BetICare! Quality inRegulatedFamilyChildCareacrossCanada. LaGrange. 2000b. and Well-Being, UniversityofGuelph. Care Centres on Wages, Working Conditions,andPracticesinChild Tougas. 2000a. on PublicPolicy. IRPP Choices versus UniversalityinEarlyChildhoodDevelopment.” Winnipeg. on ChildCareinCanada,November12-14, Council onSocialDevelopmentNationalConference Canada.” DiscussionpaperpreparedforCanadian for UrbanandCommunityStudies,UniversityofToronto. Toronto: ChildcareResourceandResearchUnit,Centre Investment inYoung Children. of GoodChildCare:TheEconomicRationaleforPublic 2005b. 2004. “FinancingEarlyLearningandChildCarein . Quebec:InstitutdelaStatistiqueQuébec. . Guelph,ON:CentreforFamilies, Work and Elements ofaHighQualityEarlyLearning 38 . Guelph,ON:CentreforFamilies,Work 13 (8).Montreal:InstituteforResearch Caring andLearningEnvironments: Grandir enqualité2003:Enquête You BetICare!ACanada-WideStudy Guelph, ON:CentreforFamilies,Work 50 (1):58-61. Caring andLearningEnvironments: . Toronto: ChildcareResource . Toronto: ChildcareResource Toronto: ChildcareResource Quality inEarlyLearning Early ChildhoodEducation Monograph 1,March. The BenefitsandCosts What DoWe . Toronto: Early Child Care Quality in Canada.” Early Childhood Children.” IRPP Choices 9 (5). Montreal: Institute for Research Quarterly, 21: 280-95. Research on Public Policy. Gormley, W.T. Jr., T. Gayer, D. Phillips, and B. Dawson. _____. 2005. Low-Fee ($5 per Day) Regulated Child Care 2005. “The Effects of Universal Pre-kindergarten on Policy and the Labour Supply of Mothers with Young Cognitive Development.” Developmental Psychology 41 Children: A Natural Experiment from Canada. (6): 872-84. Working paper 05-08. Montreal: Centre interuniversi- Harms, T., and R.M. Clifford. 1989. Family Day Care Rating taire sur le risque, les politiques economiques, et Scale. New York: Teachers College Press. l’emploi. _____. 1990. Infant/Toddler Environment Rating Scale. New Lefebvre, Pierre, Phillip Merrigan and Matthieu Verstaete. York: Teachers College Press. 2007. Dynamic Labour Supply Effects of Childcare Harms, T., R.M. Clifford, and D. Cryer. 1998. Early Subsidies: Evidence from the SLID and a Canadian Childhood Environment Rating Scale, Revised Edition. Natural Experiment. mimeo. New York: Teachers College Press. Leibowitz, A., L. Waite, and C. Witsberger. 1988. “Child Care Hausfather, A., A. Toharia, C. La Roche, and F. Engelmann. for Preschoolers: Differences by Child’s Age.” 1997. “Effects of Age of Entry, Day Care Quality, and Demography 25 (May): 205-20. Family Characteristics on Preschool Behaviour.” Journal LeRoy, S. 2006. “Child Care Choices.” Fraser Forum February: of Child Psychology and Psychiatry 38 (4): 441-48. 6-8. : Fraser Institute. Hayes, C.D., J.L. Palmer, and M.J. Zaslow. 1990. Who Cares Levin, H.M., and H.L. Schwartz. 2007. “Educational Vouchers for America’s Children? Child Care Policy for the for Universal Preschools.” Economics of Education 1990s. Washington, DC: National Academy Press. Review 26 (July): 3-16. Helburn, S.W. (ed.). 1995. Cost, Quality and Child Outcomes Lundsgaard, J. 2002. Competition and Efficiency in Publicly in Child Care Centers. Technical Report, Public Report, Funded Services. Working Paper 331. Paris: Organization and Executive Summary. Denver, CO: Center for for Economic Co-operation and Development. Research in Economics and Social Policy, University of Magnuson, K., C. Ruhm, and J. Waldfogel. 2004. Does Colorado at Denver. Prekindergarten Improve School Preparation and Institut de la Statistique du Québec. 2006. Rapport descrip- Performance? Madison, WI: School of Social Work, tif et méthodologique : Enquête sur les Besoins et les University of Wisconsin-Madison. Préférences des Familles en Matière de Services de Mathien, J., and L.C. Johnson. 1998. Early Childhood Services Garde, 2004, Tome 1. Quebec: Institute de la for Kindergarten Children in Four Canadian Provinces: Statistique du Québec. Scope, Nature and Models for the Future. Ottawa: Japel, C., R.E. Tremblay, and S. Côté. 2005. La qualité, ça Caledon Institute of Social Policy. compte ! Résultats de l’Étude longitudinale du McCain, M., F. Mustard, and S. Shanker. 2007. Early Years développement des enfants du Québec (ÉLDEQ) concer- Study 2: Putting Science into Action. Toronto: Council nant la qualité des services de garde. Montreal: for Early Child Development. Institute for Research on Public Policy. National Institute of Child Health and Human Development — Johnson, L.C., and J. Mathien. 1998. Early Childhood Early Child Care Research Network. 1996a.

Services for Kindergarten-Age Children in Four “Characteristics of Infant Child Care: Factors New Evidence about Child Care in Canada, by Gordon Cleveland et al. Canadian Provinces: Scope, Nature and Models for the Contributing to Positive Caregiving.” Early Childhood Future. Ottawa: Caledon Institute of Social Policy. Research Quarterly 11 (3): 269-306. Kershaw, P. 2007. “Measuring Up: Family Benefits in British _____. 1996b. “Infant Child Care and Attachment Security: Columbia and Alberta in International Perspective.” Results of the NICHD Study of Early Child Care.” Infant IRPP Choices 13 (2). Montreal: Institute for Research Behavior and Development 19 (Suppl 1): 171. on Public Policy. _____. 1998. “Early Child Care and Self-Control, Compliance Kozhaya, N. 2006. “$7 a Day Childcare: Are Parents Getting and Problem Behavior at 24 and 36 Months.” Child What They Need?” Economic Note, October. Montreal: Development 69: 1145-70. Montreal Economic Institute. _____. 1999. “Child Outcomes When Child Care Center Krashinsky, M. 1986. “Why Educational Vouchers May Be Classes Meet Recommended Standards for Quality.” Bad Economics.” Teachers College Record 88 (2): American Journal of Public Health 89: 1072-77. 139-51. _____. 2000. “The Relation of Child Care to Cognitive and Lamb, M.E., W. Damon, I.E. Sigel, and K.A. Renninger. Language Development.” Child Development 71 (4): 960-80. 1998. “Nonparental Child Care: Context, Quality, _____. 2001. “Overview of Early Child Care Effects at 4.5 Correlates and Consequences.” In Handbook of Child Years.” Paper prepared for symposium on Early Child Psychology. Vol 4: Child Psychology in Practice, 5th ed. Care and Children’s Development Prior to School Entry, (73-133). New York: John Wiley. biennial meeting of Society for Research in Child Lefebvre, P. 2004. “Quebec’s Innovative Early Childhood Development. Education and Care Policy and Its Weaknesses.” Policy _____. 2002a. “Child Care Structure? Process? Outcome: Options 25 (3): 52-57. Direct and Indirect Effects of Child-Care Quality on Lefebvre, P., and P. Merrigan. 2003. “Assessing Family Young Children’s Development.” Psychological Science Policy in Canada: A New Deal for Families and 13 (3): 199-206.

39 IRPP Choices, Vol. 14, no. 12, October 2008 Riddell, W.C. 2007. “InvestinginHuman Capital:Policy Richards, J.,andM.Brzozowski. 2006.“Let’sWalk Before Powell, L.M.1997.“TheImpactofChildCare Costsonthe _____. _____. _____. _____. Organization forEconomicCo-operationandDevelopment Phipps, S.,andP. Burton.1998.“What’sMine IsYours? The Phipps, S.2006.“Working forWorking Parents: The 2007. _____. National InstituteofChildHealthandHumanDevelopment, (ed _____. _____. _____. _____. _____. _____. Priorities forCanada.” In Toronto: C.D.HoweInstitute. Institute We Run:CautionaryAdvice onChildcare.”C.D.Howe Canada.” Labour SupplyofMarriedMothers:Evidencefrom (260): 599-613. Author. Employment, LabourandSocialAffairs.Paris: Paper 31. H.ImmervollandD.Barber, Directoratefor OECD Social,EmploymentandMigrationWorking Costs, Tax-Benefit Policies andWork Incentives Ottawa: HumanResourcesandSocialDevelopmentCanada. Canada. CountryNote Education andCare. (OECD). 2001. Development Children’s PreschoolCognitiveDevelopment.” 2003. “ModelingtheImpactsofChildCareQualityon Household Expenditure.” Influence ofMaleandFemaleIncomesonPatterns of Research onPublicPolicy. Canada.” Evolution ofMaternityandParental Benefits in Countries Family Life:ASynthesisofFindingsforOECD Education Division. Early ChildCareResearchNetwork,andG.J.Duncan. Youth Development Results fromtheNICHDStudyofEarlyChildCareand Quarterly Development at54Months.” Infant BehaviorandDevelopment Interaction from36MonthsthroughFirstGrade.” 451-69. Outcomes atAge4?” 976-1005. Transition toKindergarten?” Predict SocioemotionalAdjustment duringthe Educational ResearchJournal Development PriortoSchoolEntry.” and Care. and 2002b. “EarlyChildCareandChildren’s 2006. 2004. “TypeofChildCareandChildren’s 2003c. “EarlyChildCareandMother-Child 2003a. “DoesAmountofTimeSpentinChildCare 2005. 2004. 2003b. “DoesQualityofChildCareAffect .). 2005. Early ChildhoodEducationandCarePolicy: Babies andBosses:ReconcilingWork and Starting Strong II: Early Childhood Education Childhood Early II: Strong Starting Can Parents AffordtoWork? Childcare Commentary . Paris: Author. Canadian JournalofEconomics IRPP Choices 19 (2):203-30. Paris. EducationandTrainingPolicy, 74 (5):1454-75. Child CareandDevelopment: Starting Strong:EarlyChildhood . NewYork: Guildford. . OECDDirectorateforEducation. Developmental Psychology Paris: Author. 237. TheEducationPapers. 12 (2).Montreal:Institutefor A CanadianPriorities Agenda: Economica Child Development Early ChildhoodResearch 39 (1):133-64. 26 (3):345-70. New Series65 American 30 (3):577-94. 39 (3): Child 74: . Statistics Canada.2005.“ChildCare:1994/95and Special Parliamentary CommitteeonChildCare.1987. Shonkoff, J.P., andD.A.Phillips.2000. Sammons, P., K.Sylva,E.Melhuish,I.Siraj-Blatchford,B. Ruopp, R.,etal.andYouth andFamilies,Washington D.C. Ruhm, C.1998.“TheEconomicConsequencesofParental Roy, F. 2006.“FromShetoShe:ChangingPatterns of Rosenbaum, D.T., andC.J.Ruhm.2005. Romer, P.M. 2000. Wylie, C.,E.Hodgen,andJ.Thompson.2006. Walker, J.R.1992.“NewEvidenceontheSupplyofChild Vincent, C.,andF. Wooley. 2000.“Taxing Canadian Vandell, D.L.2004.“EarlyChildCare:TheKnownandthe Tremblay, R.E.,R.G.Barr, andR.D.Peters (eds.).2004. 2000/01.” Committee onChildCare Sharing theResponsibility:ReportofSpecial Development Neighborhoods: TheScienceofEarlyChildhood Institute ofEducation,UniversityLondon. Development oOverthePre-schoolPeriod. Preschool onChildren’sSocial/Behavioural Taggart, andK.Elliot.2003. Final ReportoftheNationalDayCareStudy, Vol. 1. Center: SummaryFindingsandTheirImplications. Administration forChildren.1979. Journal ofEconomics Leave Mandates:LessonsfromEurope.” 010 June,Ottawa:StatisticsCanada. Canada CanadianEconomicObserver Women intheCanadianLabourForce” Research. Cambridge, MA:NationalBureauofEconomic for Young Children. Bureau ofEconomicResearch. NBER Working Paper 7723.Cambridge,MA:National or DemandintheMarket forScientistsandEngineers? (13-55). Montreal:InstituteforResearchonPublicPolicy. Being, Policy ChoicestoImproveEconomicandSocialWell- Research. Wellington: NewZealandCouncilforEducational of EarlyChildhoodEducationto Age-14Performance. 27 (1):40-69. Analysis ofTheirFees.” Care: AStatisticalPortrait ofFamilyProviders andan Montreal: InstituteforResearchonPublicPolicy. Families: What’sFair, What’sNot?” Developmental Psychology Unknown.” http://www.excellence-earlychildhood.ca. Childhood Development.Available at line). Montreal,Quebec:CentreofExcellence forEarly Encyclopedia onEarlyChildhoodDevelopment 001XIE: 6-9.Ottawa:Author. edited byJ.Leonard,C.Ragan,andF. St-Hilaire 40 The Daily, Merrill Palmer QuarterlyJournalof . Washington, DC:NationalAcademy Press. Should theGovernmentSubsidizeSupply NBER Working Paper 11837. Feb. 7,2005.Catalogue11- 113: 285-317. Journal ofHumanResources . Ottawa:Queen’sPrinter. 50 (3):387-414. Measuring theImpactof From Neuronsto Children atthe The CostofCaring IRPP Choices . Catalogue11- Quarterly Contributions Statistics London: (on- 6 (5). New Evidence about Child Care in Canada: Use Patterns, Affordability and Quality Gordon Cleveland et al. Résumé

u cours des dernières décennies, la garde non frais de garde. Pourquoi ces frais sont si élevés pour cer- parentale a suscité un intérêt croissant au Canada taines familles et si faibles pour d’autres ? Et d’expliquer A et ce pour au moins deux raisons : l’augmentation l’existence de services de garde à coût nul. importante du taux de participation des mères au marché Ils constatent que les frais de garde annuels des du travail et une meilleure connaissance des effets béné- familles canadiennes ayant des enfants d’âge préscolaire fiques des services de garde éducatifs de qualité sur le (à l’exception du Québec) s’élèvent en moyenne à développement des jeunes enfants. 4 500 dollars pour toutes les familles dont la mère tra- Mais en dépit de cet intérêt, il est étonnamment difficile vaille et à 6 900 dollars pour celles dont la mère travaille de trouver des informations précises et actualisées sur la toute l’année à temps plein. Soit, généralement, près de question. Il n’existe par exemple aucune enquête statistique 20 p. 100 du revenu de la mère. Et pour la plupart des visant à recueillir des données sur l’utilisation, le prix et les familles, constatent-ils, les dépenses en frais de garde dépenses, le coût ainsi que la qualité des services de garde. sont associées au niveau de revenu de travail de la mère. La présente étude tente de combler cette lacune en Du point de vue des politiques publiques, ces observations puisant à plusieurs sources de données comme l’Enquête suggèrent que les frais de garde sont un obstacle à la partici- longitudinale nationale sur les enfants et les jeunes pation au marché du travail pour de nombreuses mères. Dans (ELNEJ), l’Enquête nationale sur la garde d’enfants certains cas, cela leur imposera de modifier leur situation (ENGE), l’Enquête sur les dépenses des ménages (EDM) et d’emploi ou le type de garde non parentale qu’elles utilisent quelques enquêtes québécoises et canadiennes sur la qua- afin de réduire le fardeau financier de services de garde. lité des services. Elle apporte ainsi un éclairage nouveau Étant donné le recours très répandu à la garde non sur la garde non-parentale au Canada, en termes d’utilisa- parentale, ces conclusions soulèvent aussi la question de tion, de dépenses et de qualité des services utilisées. leur qualité. En troisième partie, les auteurs examinent donc Après avoir présenté les façons de penser les plus les données disponibles et concluent à une qualité trop sou- répandues en matière de garde des enfants et décrit le vent insuffisante. Même si la grande majorité des milieux contexte politique dans lequel se tient le débat actuel réglementés offrent un environnement sain et sécuritaire autour de cette question, les auteurs examinent en détail aux enfants, beaucoup ne développent pas pleinement les comment les familles canadiennes gardent ou font garder aptitudes cognitives, linguistiques, comportementales, leurs enfants, en soulevant les questions suivantes : sociales et motrices des enfants. L’information dont on dis- Combien de familles utilisent des services de garde non pose en ce qui concerne les milieux non réglementés semble parentale ? Comment ces habitudes ont-elles évolué dénoter une qualité plus faible encore. depuis 1995 ? Les familles dont la mère ne travaille ou Puisque les mères sur le marché du travail sont là n’étudie pas ont-elles recours aux services de garde non pour y rester et que la grande majorité de leurs enfants parentale ? Qui au juste utilise quels types de services ? fréquentent des services de garde non parentale, les Les auteurs examinent toutes les catégories de familles, auteurs font valoir qu’il est temps de changer notre mais les données disponibles les ont amenés à privilégier façon de voir la situation et de considérer la garde celles dont la mère travaille ou étudie. Par ailleurs, ils étu- parentale et la garde non parentale comme des dient la question en adoptant une définition étendue des approches complémentaires et non comme des substituts. modes de garde, incluant la garde parentale, la garde non Ils préconisent donc une stratégie qui rendrait plus abor- parentale réglementée ou non, ainsi que la maternelle. dables pour toutes les familles des services de garde édu- Ils constatent que près de 80 p. 100 des enfants d’âge catifs de qualité. Ils évaluent à cette fin les avantages et préscolaire dont la mère travaille ou étudie sont régulière- les limites des diverses formes d’aide publique : 1) les ment en garde non parentale. Et la moitié le sont dans des crédits d’impôt ou les transferts en espèces aux parents ; milieux réglementés, une proportion qui atteint près des 2) les coupons liés à l’utilisation de services de garde deux tiers si l’on inclut la maternelle. Par ailleurs, un non parentale ; 3) les subventions ciblées pour les nombre appréciable d’enfants dont la mère est au foyer familles à faible revenu ; 4) le financement direct des fréquenteraient aussi une garderie, une prématernelle, une fournisseurs de services de garde réglementés. C’est avec maternelle ou un autre type de garde réglementé. ces deux dernières formes d’aide que l’incidence serait la En ce qui concerne les dépenses, les auteurs tentent de plus positive, estiment-ils, surtout le financement direct savoir combien les familles canadiennes dépensent en pour améliorer l’abordabilité et la qualité. 41 New Evidence about Child Care in Canada: Use Patterns, Affordability and Quality Summary Gordon Cleveland et al.

ver the past 25 years there has been an explosion Turning to spending patterns, the authors provide of interest in nonparental child care in Canada. answers to the following questions: What do Canadian O The rapid increase in the workforce participation families spend on child care? Why do some families of mothers has produced high levels of demand for these spend a lot and others nothing? How can we explain services. There has also been increasing recognition that zero-priced child care? good-quality early childhood education and care can The authors show that the average annual spending on have beneficial effects on child development. child care by Canadians with preschool children living Despite all of the public interest, it is surprisingly diffi- outside Quebec is $4,500 for all families in which the cult to find accurate and current information on child care mother is employed and $6,900 for families in which the in Canada. For instance, there are no statistical surveys mother is employed full-time, full-year. Typically, child dedicated to collecting information on the use patterns, care costs are close to 20 percent of the mother’s earn- prices and expenditures, costs and quality of child care. ings. Decisions on child care spending, for most families, The authors attempt to fill this knowledge gap by appear to be related to the mother’s earnings. examining evidence from several sources, such as the From a public policy perspective, these findings sug- National Longitudinal Survey of Children and Youth gest that, for many mothers, the cost of child care is a (NLSCY), the Canadian National Child Care Survey significant barrier to employment. In some cases, mothers (CNCCS) and the Survey of Household Spending (SHS), as may have to adjust their employment circumstances in well as surveys on the quality of child care services con- order to decrease the financial burden of nonparental ducted in Quebec and Canada. In so doing, they unearth care. Alternatively, families may decide to patch together new evidence on patterns of child care use, child care arrangements of dubious quality at a lower cost. spending and child care quality. Given the widespread use of nonparental care, these After summarizing prevalent policy perspectives on findings also raise the question of quality. In the third sec- child care in Canada and describing the policy context in tion of the paper, the authors review evidence currently which the child care debate is played out, the authors available, and find that the level of quality is too often examine child care use patterns among Canadian fami- inadequate to support child development. Although the lies, and in particular how many families use nonparental vast majority of regulated care is supportive and safe, care for their children, how patterns have evolved since much of the regulated care that is being provided does not 1995, whether families in which the mother is neither optimally develop children’s cognitive, language, behav- employed nor studying use nonparental care, and which ioural, social and motor skills. The small amount of evi- type of child care is used by whom. dence that is available on unregulated child care suggests While the authors consider all types of families, their that the average quality of this type of care is even lower. main focus, because of data availability, is families in The authors argue that, given that mothers are in the which the mother is employed or studying. Moreover, they labour force to stay, and given that Canadian children are define the term child care very broadly, to include all types overwhelmingly in nonparental care for significant portions of parental or nonparental care, including school-based of each week, parental and nonparental care should be kindergarten. The inclusion of kindergarten in discussions viewed as complements rather than as substitutes. More of child care is unusual, but the authors explain that specifically, they call for a strategy to make high-quality kindergarten services are often close substitutes for day- types of early learning and child care more affordable for care and thus should be included in the analysis. all families. They review the strengths and weaknesses of They find that nearly 80 percent of preschool children the following forms of public assistance in this regard: (1) with employed or studying mothers receive nonparental cash or tax credits; (2) vouchers tied to the use of non- care regularly. Half of these children are primarily in reg- parental care; (3) enhancement of existing child care sub- ulated forms of child care, and if kindergarten is included sidy systems for low-income families; and (4) direct as a type of regulated care the proportion is closer to funding of child care centres, nursery schools, preschools, two- thirds. The authors also point out that a consider- regulated family child care providers and kindergartens. able number of children with stay-at-home mothers Their analysis suggests that the last two of these options, attend nursery school, pre-kindergarten or kindergarten especially direct funding to improve affordability and quali- or use some other form of regulated child care. ty, are likely to have the greatest positive impact overall. 42