The National Care System in Uruguay

Demographic and social changes have posed challenges to social protection systems in both developed and developing countries. In the former, population is gradually aging due to the ongoing demographic transition, characterized by fewer children being born and more adults surviving into old age. At the same time, the prevalence of significant gender gaps in the provision of care, with links to documented differences between men and women in the labor market, reinforce the need for innovation and new approaches to the provision of social protection. In this context, many countries are implementing important efforts to extend child- and elder care services. In the case of child care, evidence is abundant on the importance of the early years of life in terms of individual health and cognitive, social, and emotional development (e.g., Heckman et al. 2010a; 2010b). Child care services may also be seen as an instrument to encourage women's labor force participation rates, as well as strengthen their position within the labor market. In the case of elder care, even if elderly morbidity and disability continue to improve, a growing number of the old will need help to lead satisfactory lives. Improving living standards for elderly people provides a rationale for new policies, as does the aim of relieving relatives from the burden of providing permanent care in elders’ advanced stages of life. Uruguay is a country that stands out in Latin America because of its relatively older and more mature social protection system (Ocampo and Gómez-Arteaga 2017). It has recently made advances in the implementation of a National Care System (NCS), which includes both child care and elder care for dependents. The policy also offers care services for persons with disabilities, although the analysis we present in this article does not cover this population. The NCS can be included among the recent interventions that address new social risks-sometimes referred to as social investment interventions-borne by groups such as the young, the low-skilled, and women, as opposed to such traditional social risks such as old age or unemployment, which had their roots mostly in the protection of male breadwinners (Cantillon 2011; Morel, Palier, and Palme 2012). Although social investment interventions do not necessarily crowd out spending on traditional risks, the possibility of some shift raises the issue of the redistributive impact of these new interventions. How to evaluate their effects in redistributive terms (Vaalavuo 2013; Van Kersbergen and Hemerijck 2012)?1 Redistribution-at least vertical redistribution-is not always the rationale for public services (Vaalavuo 2013), but the discussion about redistributive impacts is still important to understanding, among other things, the political economy of these policies. Different countries have adopted a variety of models to implement and finance child care and promote early child development. European countries tend to rely on publicly funded programs, whereas the United States relies more on subsidies and tax credits to reimburse parents for child care expenses (Waldfogel 2001). In the Nordic countries, parents’ contributions are set nationally at a maximum level decided by the state and depend on household earnings and the number of children benefiting from child care. In addition, there may be rules about maximum payment (Eydal and Rostgaard 2013). The final result of differential combinations of child care policies is that in most countries, services are not entirely free, but parents pay subsidized prices (Plantenga and Remery 2005). With regard to long-term care, given the rapid process of aging and reduction in the supply of informal care due to employment of traditional (mostly female) caregivers, governments all over the world have found organizing and funding these services to be a major challenge (Barr 2010; Costa-Font, Courbage, and Zweifel 2017; and Wouterse and Smid 2017, among others). Public long-term care coverage systems across countries in the Organisation for Economic Co-operation and Development (OECD) require users to share a portion of the cost for their care. In Latin America, the prevalence of fiscal constraints, the limited supply of long-term services by public health service providers, and the lack of regulated private markets for long-term care are of particular concern. It is reasonable to expect-both for the region and for Uruguay-that society will have to address the growing risks of dependency at older ages through publicly financed mechanisms, focused initially on the poor and vulnerable (Caruso, Galiani, and Ibarrarán 2017). Doing this may entail implementing tax incentives or subsidies (means-tested or universal), which may in turn affect savings. In this article, we assess the distributional impact that the NCS would have in Uruguay through the calibration of a static tax benefit model which places values on the public services provided and allocates them among households. We estimate the impact of two alternative scenarios that reflect different allocations of social investment as between children and the elderly. While children are overrepresented among the poor, the elderly have above-average income and low poverty rates. Also, the scope of the intervention is different between the two populations-we simulate universal coverage among the preschoolers and assistance rates over 50 percent for 2-year-olds while determining elderly coverage by the person's dependent condition. Thus, the expected redistributive effects should be stronger for children than for the elderly; in the latter case, the effects may even be regressive. We consider both populations together, as their inclusion in the same initiative is a distinctive characteristic of the Uruguayan proposal and institutional design, but we also report separated effects, due to the differences explained above. As a funding strategy, we simulate an increase in marginal income tax rates. The analysis we present is a first approximation of the potential impact, in that it considers only first-round effects, excluding long-term redistributive effects that may arise from increases in labor supply and consequently in household income. If increases in female labor force participation were to take place-mainly among women from lower-income households, for whom the current labor supply is lower-redistributive effects would be higher. Also, given that children's enrollment in formal child care or preschool education has proven beneficial for their future school achievement and earnings (Heckman and Masterov 2007; Heckman et al. 2010a, 2010b; Duncan and Sojourner 2013), these policies entail a further potential redistributive effect in the medium- and long runs, as higher equality could result over time. Moreover, these effects tend to be stronger for children who come from vulnerable backgrounds (Cascio 2015), improving the potential distributive effects in the long term. Another important issue not considered in this article is the fiscal pressure that arises from demographic trends. Uruguay was one of the first countries in Latin America to embark on the demographic transition. These trends will have important implications for the social protection system and labor market (Rofman, Amarante, and Apella 2016). The expected increase in the elderly population will probably increase the total demand for care, even considering the predicted reduction in the demand for child care due to declining fertility. At the same time, there will probably be a reduction in the availability of informal and familial care due to the increase in female labor force participation. Our analysis only considers the effects arising from the current design of the programs up to 2020, based on the 2014 population structure. Evidence on the redistributive impact of these policies is still scarce, especially in developing countries. Nevertheless, a new strand of literature provides useful information about the effects of public care services, and our article aims at contributing to the ongoing discussion. On that note, the extension of child care services has been found to reduce income inequality (Hufkens et al. 2015; Matsaganis and Verbist 2009; Vaalavuo 2013). However, these results depend heavily on who uses the services and on the tax structure (Van Lancker and Ghysels 2011). In every case, the impact is relatively small compared with traditional welfare state spending, though it may be equalizing or pro-poor (Vaalavuo 2013). Less research has been carried out for elderly care services, although scarce existing results seem to be along the same lines as those for child care (Vaalavuo 2011, 2013). The NCS's objective is to expand available care services for the dependent population, as well as to create new services. For the NCS, the dependent population consists of children aged 0-12 years (prioritizing those aged 3 and younger), persons with disabilities, and those elderly who are not autonomous in their daily lives. A synthesis of the policy process that led to the design of the Uruguayan NCS can be found in MIDES (2014), and details about the current policy design can be found in Junta Nacional de Cuidados (2015) and in the preamble to the 2015 Budget Law (MEF 2015). For younger children, the policy consists of the expansion of child care services, whereas for the dependent disabled and elderly, it focuses on the provision of home-based paid care. Care services for persons with disabilities are not considered in our distributional analysis, due to the lack of suitable statistical information. The new policy has a universal claim but foresees a gradual implementation. In the first stage (until 2020), the aim is to install and develop diverse programs directed toward the dependent elderly population but targeting the benefits to those in greater need. The 2015 Budget Law sent to Congress covered up to the year 2017, establishing additional resources for the NCS and setting coverage goals. However, it did not include any new tax revenues to finance the proposed expansion of the NCS. Instead, the law based the funding of the new programs on the proceeds of economic growth. The estimated annual additional resources approved for 2017 were US$67 million (prices as of January 2015); more than one-third of this sum was allocated for early childhood services (36 percent), while 29 percent was set aside for the elderly, 22 percent to people with disabilities, and 12 percent to administrative expenses. Child care and education, while both depending on the national government, are separate policy domains in Uruguay. Pre-primary education starts at 3 years of age but is compulsory from age 4. Primary schooling starts at age 6. For children aged between 1 and 3, child care is available through a wide variety of public and private services. The main public service is Plan CAIF, a program that emerged in the late 1980s with a clear target of serving vulnerable children. Public preschool and child care are free of charge and state-funded, although child care is privately organized. Public child care services at Plan CAIF offer day care for children ages 2 and 3, and an incipient service is proposed for 1-year-olds. CAIF also provides weekly workshops on childrearing and child development guidelines for families with children under 2. Originally, the daily services covered four hours a day; they are now expanding to cover six and eight hours per day. Public preschool started to expand in the 1990s to cover children aged 4 and 5 and is currently committed to universal coverage for 3-year-olds. Figure 1 shows assistance rates by age for day care and preschool at public and private services, by income groups. For 3-year-olds, both types of services (day care and preschool) overlap, so the rates refer to the global coverage. The rates increase with age, as well as the weight of public provision: overall, 42 percent of children aged 1-3 use child care or preschool centers on a daily basis-13 percent of 1-year-olds, 42 percent of 2-year-olds, and 69 percent of 3-year-olds. As expected, assistance rates hide important differences between income groups, being significantly higher among the better-off households, which concentrate their use in the private sector. In the lowest quintiles, the rates are lower, and the public sector predominates. Age-specific assistance rates and female employment rate, by per capita income quintile, Uruguay, 2014 NOTE: Assistance rates do not include CAIF's family workshops for 1-year-olds. SOURCE: Authors' calculations, based on the Continuous Household Survey (INE 2014). At the same time, female employment increases with income, presenting an important socioeconomic stratification. The correlation between female employment and child care has been a debated issue in economic studies (Baker, Gruber, and Milligan 2008; Cascio 2009; Del Boca, Pasqua, and Pronzato 2009). In effect, the obvious correlation suggests that publicly provided or subsidized child care could induce increases in female employment. However, it is also possible that these policies may crowd out other forms of care, resulting in smaller-than-expected increases in employment (Havnes and Mogstad 2011). In Uruguay, an impact evaluation of an expansion of public preschool services for children aged 4 and 5 during the 1990s found that this expansion crowded out attendance at private schools, particularly among children of high-skilled mothers. Among children of low-skilled mothers, attendance increased, but the policy did not have any effect on mothers’ labor market outcomes (Nollenberger and Perazzo 2016). These results are important for the discussion about the potential impacts on female employment of the NCS. Turning to the elderly, public care services are a very recent development. Besides some examples of institutional elderly care, the first program intended to address care of dependent elderly-the Programa de Asistentes Personales-was implemented in 2014, and its coverage was very limited. In contrast, the pension and the health systems have almost universal coverage in Uruguay, and their quality is better than the average of the region. In Uruguay, people aged 64 and older have higher incomes than the national average and exhibit the lowest poverty rates in the population (INE 2016). The NCS has distinguished four levels of dependency based on a person's capacity to perform daily life activities without help: nondependence, mild dependency, moderate dependency, and severe dependency. Following the objectives of the NCS in this first stage, this article focuses on severe dependency. Severe dependency affects 3.1 percent of persons aged 65 and older (Table 1). The incidence is significantly higher among those aged 85 or older, which poses a threat for the future of social policies in a demographic context where the population of the “oldest old” is expanding more rapidly than that of the elderly as conventionally defined. If the dependency rate does not change, the proportion of severely dependent people in the population may be sizable by 2050. The limited scope of this article does not allow an analysis of these trends in greater detail or their implications in terms of public spending and distributive impact. These issues should be covered in future research.2 Owing to imputation procedures detailed in the next section, the income distribution of the severely dependent is assumed to follow that of the elderly. This implies that about one-third belong to households in the richest quintile and only 5 percent to those in the poorest one. It derives from this procedure that the incidence of severe dependency is similar among all the income groups, which may introduce some bias into our results.3 The relationship between income and dependency is not clear, but previous research for Uruguay shows higher rates of disability in poorer households (Bagnato, Luzardo, and Padula 2011). This may not stand so clearly for the elderly, as life expectancy is lower for people who belong to vulnerable households, whereas old-age dependency is highly age-related.4 Our objective is to analyze the impact of the new NCS programs on the budgets of affected households. To do so, we calibrate a static tax benefit model. We first apply the rules that define the social programs involved to identify the beneficiaries. Then we assign public services a monetary value. The final step consists of deducting the payments that each household would have to make to finance the NCS. By comparing the original and counterfactual vector of household income, we identify “winners” and “losers” from this intervention and quantified its redistributive impact. When the counterfactual household income is higher than the original one for a certain household, it is considered a winner, and when it is lower, it is considered a loser. Households whose income does not change (they do not receive benefits nor pay taxes) are considered unaffected. Our analysis is based on the 2014 Uruguayan Continuous Household Survey, a reference survey for income, living conditions, labor market status, and education conducted by the National Statistical Institute (INE 2014). In 2014, this nationally representative survey covered 132,000 people in 49,000 households. It provided information on the use of child care and preschool services and accurate income data. Unfortunately, it did not collect information on dependency. We took that information from the Longitudinal Social Protection Survey (LSPS) (BPS and IDB 2013). (Details on how we combined these datasets are provided below.) Additional data on the costs of service provision were derived from the 2015 Budget Law. Given that the household survey collected information on net income (after taxes and social contributions), gross income was calculated by applying the rates of taxation and social contribution. Our analysis was that we did not take into the previous we the of the benefit to the gross income of the household in the case of new beneficiaries. This is especially for the case of as the coverage of public child care is important and the is not included as income in the We as NCS those people who would use the system in the reference so we took a static and of for the of especially when we considered and a it is highly that will use the NCS at some in as a or as a or of one. Our analysis covered those programs whose potential could be the data child care services of CAIF for children aged expansion for 3-year-olds, and home-based care for the dependent elderly. to the preamble to the 2015 Budget the programs that were included in our analysis make up percent of the 2017 allocated to the NCS. The expansion and coverage of the programs for and 2017, as well as the global costs of each were in the preamble to the 2015 Budget Law. and were for on this information and on further of each structure of CAIF CAIF and of Uruguay the expansion scenarios for 2 shows the number of and the and to the for In the case of children, coverage rates are proposed for in the 2015 Budget Law. To perform the we assumed that all of the expansion will be publicly provided and that all of the places will be This implies that the coverage expansion will to public assistance rates. These may be very as enrollment and assistance rates for these services and as some of the expansion be privately the number of Evidence from European countries shows that the cost and availability of child care are not the only that the use of child care may also a 2015). research for Uruguay has that the of family may the use of child care services for very children and 2013). We do not take these into in our 3 shows the assistance rates and the The that proposed in the preamble of the 2015 Budget Law in the for 1-year-olds, where the law includes the coverage to family which is not considered in this The distributional impact on the income However, information does not for the new beneficiaries. The of the scope of the intervention and its of As in the given that child care as a for it is reasonable to that the demand for these services will be higher among parents (Cantillon 2011; and Van Lancker 2013). On the targeting children in greater need provides further incentives for the labor participation of their and may have greater redistributive impact. To estimate reasonable for the impact, we present two alternative The first that will be based on their demand for child care. We estimated a model for child care assistance and and the children not assistance by their predicted of who have a higher who are not child are as the number of by age is This model the use of the available services, which implies higher for higher The alternative that is some the poorer children of each age. This traditional in child care, which is directed toward the households. We the of the child care programs to the for the new care centers in to target children from lower income households. and the of the new CAIF centers by considering that the potential within a and that at least percent of them belong to households for the program as de or For the expansion of a was also proposed by and who where the expansion should for better considering that it in with children. on their we the total places by the new centers proposed to the number of total in 2. we them from the where the centers would be considering the proportion of by age and the proportion of that the had in each the elderly, the preamble to the 2015 Budget Law proposed percent coverage of the severely dependent by The challenge is dependent people among the elderly, as dependency is not about in the household survey and from every age 64 help in daily life We to the 2014 which included a set of of the four of dependency by the NCS in the nondependence, mild dependency, moderate dependency, and severe MIDES (2015) that the was a accurate data for the of severe dependency. To identify the population in the household we the severe dependency rate in the to the population. The of among the population estimated in the household survey was based on demographic and income We percent of the severely dependent population, that the and distribution of the is the same as for the total elderly population. The income groups were based on the that the program for the The age and income of the are in This is based a number of due to the lack of dependency data. In that our results should be considered as and need further better data are Figure 2 the distribution of the and by In both the potential distributive effects would come from children, as their distribution is more in the lower of the income to that of the elderly. differences are found between the alternative scenarios of of child care beneficiaries. As expected, in the demand the distribution among income groups is almost whereas the alternative poorer households percent of the are among the poorer are in the than distribution of by income SOURCE: Authors' calculations, based on the Continuous Household Survey (INE 2014). Turning to the of public services care and preschool we the in the literature of to the the average costs of the This is a as it does not for the quality and of service Another is that we do not consider the of use of to average The services are available for or eight hours per but we do not have information about the distribution of the new between Given that our at the potential effects of the it an analysis based on costs as in We took the and for each program from the preamble to the 2015 Budget Law and calculated the This process was for as it only referred to children aged For child care and we had to for the that the costs for each age in addition, for children aged the program included two care and the but only the was of for this the distribution of the program between ages and the number of in each age, we an We the costs an for each The program is a for an The of the is there are four household income groups, and the of the with income. we to the the monetary The is set at The lower-income the the next percent of the the and is to the The final step consists of deducting the payments that each household would have to make to finance the NCS. The funding consists of the income Uruguay has a tax of a income tax a de or which a tax for labor income with a low tax rate on income. The labor income of the consists of marginal income tax rates, from percent in the lowest to percent in the with This tax is only to formal contributing to the social who percent of all are by a similar tax a de or income is at differential rates 3 percent to 12 depending on the We a 5 percent increase in the marginal rates of income tax to finance the costs of the expansion of NCS. on the current and proposed rates are in The expected from such a change would be about which would almost cover the total as in The programs that we are are to percent of in 2014 and of the of a that has been to reduce inequality by one of the in Uruguay

Business · Gender, Labor, and Family Dynamics · Intergenerational Family Dynamics and Caregiving · Poverty, Education, and Child Welfare

• Gender Inequity

  Open Access•Maria Amparo Cruz-Saco, Mirian Gil et al.•Social Inclusion•2021

  Incluir item

• International actors and national policies

  Open Access•Meika Sternkopf•Policy Sciences•2024

  Incluir item

• Implementing policy integration

  Open Access•Guillermo M Cejudo, Cynthia L Michel•Policy Sciences•2023

  Incluir item

• Maternity, motherhood and access to contributory pensions

  Open Access•Ianina Rossi•International Social Security…•2025

  Incluir item

• The rate of return to the HighScope Perry Preschool Program

  Open Access•James J Heckman, Seong Hyeok Moon et al.•Journal of Public Economics•2010

  Incluir item

• Motherhood and market work decisions in institutional context

  Daniela Del Boca, Silvia Pasqua et al.•Oxford Economic Papers•2009

  Incluir item

• Why We Need a New Welfare State

  Anton Hemerijk, Duncan Gallie et al.•Why we need a new welfare state•2002

  Incluir item

• Social Spending, Taxes and Income Redistribution in Uruguay

  Marisa Bucheli, Nora Lustig et al.•2013

  Incluir item

• Demographic Change in Uruguay

  Rafael Rofman, Amarante et al.•Demographic Change in Uruguay•2016

  Incluir item

• Money for nothing? Universal child care and maternal employment

  Open Access•Tarjei Havnes, Magne Mogstad•Journal of Public Economics•2011

  Incluir item

• The Productivity Argument for Investing in Young Children

  James J Heckman, James Heckman et al.•Applied Economic Perspectives and…•2007

  Incluir item

• Three Shortcomings of the Social Investment Perspective

  Open Access•Bea Cantillon, Wim Van Lancker•Social Policy and Society•2013

  Incluir item

• International Policies toward Parental Leave and Child Care

  Jane Waldfogel•The Future of Children•2001

  Incluir item

• The paradox of the social investment state

  Open Access•Bea Cantillon•Journal of European Social Policy•2011

  Incluir item

• Long‐term Care

  Open Access•Nicholas Barr•Social Policy and Administration•2010

  Incluir item

• Universal Child Care, Maternal Labor Supply, and Family Well‐Being

  Michael Baker, Jacob Gruber et al.•Journal of Political Economy•2008

  Incluir item

• Two Decades of Change in Europe

  Open Access•Kees Van Kersbergen, Anton Hemerijck•Journal of Social Policy•2012

  Incluir item

• The Redistributive Impact of 'Old' and 'New' Social Spending

  Open Access•Maria Vaalavuo•Journal of Social Policy•2013

  Incluir item