The socioeconomic gradient in child health through childhood and across time
There is strong persistence in social status across generations. Parental socioeconomic status may affect child health through channels such as parent's income, education level, health and relative social status. As a result, transmission of health potentially contributes to the intergenerational persistence in social status. The aim of this project is to increase our understanding of the socioeconomic gradient in child health and its consequences for social mobility. We study how the socioeconomic gradient evolves through childhood and what types of health problems are most frequent at different ages. Understanding these patterns is potentially important for our understanding of the gradient and for the design of policy. We will also study how the socioeconomic gradient in child health has responded to the recent increases in income inequality, which will further contribute to our understanding of its underlying determinants. Finally, the project explores how health problems in childhood affect school performance and labor market outcomes, a finding that will help uncover how the socioeconomic gradient transmits over generations. The project will rely on register data combining information on children and their parents from social insurance-, patient-, prescription drug- and birth registers, with information from the employment agency and statistics Sweden 1987-onwards.
Final report
The purpose of the project is to increase our understanding of the socioeconomic gradient in health and its consequences for social mobility. The project consists of three parts: i) a descriptive study of the social gradient using in- and outpatient data and medical prescriptions; ii) a study of causal effects of income (inequality) on health; and iii) investigation into the consequences of poor health on long-term outcomes, such as labor market outcomes.
The work relies heavily on a database with connected data from the National Board of Health and Welfare, Statistics Sweden and the National Insurance Board. A major obstacle has been the late arrival of data due to severe delays at the National Board of Health and Welfare, which has also entailed more work from the researchers in the project. The database was not complete until the fall of 2018 and, due to the delays, the project is still ongoing and will be finished at a later stage.
The first part of the project, describing social gradients, is proceeding according to plan. The second part, with the aim of analyzing the causal effects of a family's social status on child health, has been slightly adjusted. The idea was to use a quasi-experimental approach to isolate the effect of family income on child health from other factors, such as parental health, educational level and factors that could be related to both family income and child health. Unfortunately, the proposed identification strategy, using a so-called Bartik instrument, did not work as planned. Therefore, the project has investigated other identification strategies, with the aim of estimating causal effects of social status on child health. One study uses matching strategies to investigate the causal effects of exposure to parental job loss on child health and human capital development. In another study, we investigate whether access to preschool during parental leave affects child health and education, and whether the effects differ depending on parents’ income and education level. Another strategy we are working on, with the purpose of identifying causal effects of social status on health care utilization, is to use the reductions in psychiatric hospital care for children in the 80s-90s. The last part of the project, to study the effects of childhood illness on later outcomes, is delayed due to late data delivery.
Among the three most important results so far are, firstly, a negative social gradient in hospital care and specialist care. The socio-economic differences in health outcomes emerge early. Children of low-income parents are at higher risk of being born with a low birth weight and poor health. Although there is a social gradient for most causes for care, the association is most evident for mental health diagnostic groups, such as behavioral problems, depression, abuse, etc. Furthermore, the relationship between family income and care is stronger for hospital care than specialist care. Possible reasons might include the following: poorer families do not seek care for less serious illnesses and injuries; they do not receive the same care, perhaps because they are not as argumentative; wealthier families seek care at an earlier stage and therefore less frequently need hospital care.
A second important finding is that exposure to workplace closure has negligible negative consequences for children. Although job loss reduces family income and has negative effects on parental health, there are no, or only minor, effects on child health and schooling outcomes. This result is somewhat at odds with results in studies from other countries. A possible interpretation of the result is that children do not suffer as in other countries because of universal access to a social safety net in the form of preschool, school, child health care and unemployment benefits for the parents. However, preliminary results from the study of access to preschool for children with parents on parental leave do not indicate that access to preschool is of crucial importance for the social gradients in children's early health and school outcomes in the Swedish context.
The third finding is methodological. Identification strategies that rely on changes in revenue streams at the national level to estimate the causal effects of income changes at the local level rely on assumptions that are often not fulfilled. A more accessible way forward, to identify causal effects of social status, is to look for quasi-experiments due to policy changes, for example. This conclusion has led use to investigate new avenues of research. We have collected information about the number of places in psychiatric inpatient care in different regions during the 80 and 90s, a period when counties greatly reduced the number of such places. Since the counties reduced the number of places at different times, we can study how this change affects the care given to children with different parental backgrounds, while controlling for trends in the care needs of different income groups as well as persistent differences between counties. The preliminary results indicate that a reduced number of places in psychiatric inpatient care mainly affected the care received by children from low-income families. Further research questions raised by the project are how changes in support services affect the well-being of children in recipient families? A doctoral student linked to the project group has begun a study of the effects of the national standard for pre-employment support on child health and human capital development. The project's database provides even more opportunities for future studies.
The results so far have been presented in national and international conferences, workshops and seminars, both directly to stakeholders and the general public. Presentations include: The National Conference, Växjö 2018; Örebro University 2018; SACO, Saltsjöbaden 2018; IFAU / UCLS / HEFFU Workshop on the Social Gradient in Human Capital Formation, Uppsala 2019; AASLE, Singapore 2019; “22nd Annual Enterprise and Competitive Environment Conference”, Brno 2019; the pod "Rather rich and healthy", organized by SNS and Kapitalet, 2020; and accepted to EALE / SOLE, Berlin 2020.
Together with Eva Mörk, Uppsala University, the project group has also organized a research conference on the topic "The Social Gradient in Human Capital Formation" in Uppsala May 27-28, 2019, with participants from 8 countries. The keynote speaker was Gabriella Conti from University College London.
Some results from the first part of the project will be published in Ekonomisk Debatt this spring, and the plan is to complete a scientific article including in-depth analysis later. The results from part two are published in IFAU WP 2019: 3 and we have high hopes that it will be published in Labor Economics after requested revisions. The study on access to preschool for children with parents on parental leave will be completed in 2020/21 and published as an IFAU working paper. We also hope for publication in a highly ranked scientific journal.
The work relies heavily on a database with connected data from the National Board of Health and Welfare, Statistics Sweden and the National Insurance Board. A major obstacle has been the late arrival of data due to severe delays at the National Board of Health and Welfare, which has also entailed more work from the researchers in the project. The database was not complete until the fall of 2018 and, due to the delays, the project is still ongoing and will be finished at a later stage.
The first part of the project, describing social gradients, is proceeding according to plan. The second part, with the aim of analyzing the causal effects of a family's social status on child health, has been slightly adjusted. The idea was to use a quasi-experimental approach to isolate the effect of family income on child health from other factors, such as parental health, educational level and factors that could be related to both family income and child health. Unfortunately, the proposed identification strategy, using a so-called Bartik instrument, did not work as planned. Therefore, the project has investigated other identification strategies, with the aim of estimating causal effects of social status on child health. One study uses matching strategies to investigate the causal effects of exposure to parental job loss on child health and human capital development. In another study, we investigate whether access to preschool during parental leave affects child health and education, and whether the effects differ depending on parents’ income and education level. Another strategy we are working on, with the purpose of identifying causal effects of social status on health care utilization, is to use the reductions in psychiatric hospital care for children in the 80s-90s. The last part of the project, to study the effects of childhood illness on later outcomes, is delayed due to late data delivery.
Among the three most important results so far are, firstly, a negative social gradient in hospital care and specialist care. The socio-economic differences in health outcomes emerge early. Children of low-income parents are at higher risk of being born with a low birth weight and poor health. Although there is a social gradient for most causes for care, the association is most evident for mental health diagnostic groups, such as behavioral problems, depression, abuse, etc. Furthermore, the relationship between family income and care is stronger for hospital care than specialist care. Possible reasons might include the following: poorer families do not seek care for less serious illnesses and injuries; they do not receive the same care, perhaps because they are not as argumentative; wealthier families seek care at an earlier stage and therefore less frequently need hospital care.
A second important finding is that exposure to workplace closure has negligible negative consequences for children. Although job loss reduces family income and has negative effects on parental health, there are no, or only minor, effects on child health and schooling outcomes. This result is somewhat at odds with results in studies from other countries. A possible interpretation of the result is that children do not suffer as in other countries because of universal access to a social safety net in the form of preschool, school, child health care and unemployment benefits for the parents. However, preliminary results from the study of access to preschool for children with parents on parental leave do not indicate that access to preschool is of crucial importance for the social gradients in children's early health and school outcomes in the Swedish context.
The third finding is methodological. Identification strategies that rely on changes in revenue streams at the national level to estimate the causal effects of income changes at the local level rely on assumptions that are often not fulfilled. A more accessible way forward, to identify causal effects of social status, is to look for quasi-experiments due to policy changes, for example. This conclusion has led use to investigate new avenues of research. We have collected information about the number of places in psychiatric inpatient care in different regions during the 80 and 90s, a period when counties greatly reduced the number of such places. Since the counties reduced the number of places at different times, we can study how this change affects the care given to children with different parental backgrounds, while controlling for trends in the care needs of different income groups as well as persistent differences between counties. The preliminary results indicate that a reduced number of places in psychiatric inpatient care mainly affected the care received by children from low-income families. Further research questions raised by the project are how changes in support services affect the well-being of children in recipient families? A doctoral student linked to the project group has begun a study of the effects of the national standard for pre-employment support on child health and human capital development. The project's database provides even more opportunities for future studies.
The results so far have been presented in national and international conferences, workshops and seminars, both directly to stakeholders and the general public. Presentations include: The National Conference, Växjö 2018; Örebro University 2018; SACO, Saltsjöbaden 2018; IFAU / UCLS / HEFFU Workshop on the Social Gradient in Human Capital Formation, Uppsala 2019; AASLE, Singapore 2019; “22nd Annual Enterprise and Competitive Environment Conference”, Brno 2019; the pod "Rather rich and healthy", organized by SNS and Kapitalet, 2020; and accepted to EALE / SOLE, Berlin 2020.
Together with Eva Mörk, Uppsala University, the project group has also organized a research conference on the topic "The Social Gradient in Human Capital Formation" in Uppsala May 27-28, 2019, with participants from 8 countries. The keynote speaker was Gabriella Conti from University College London.
Some results from the first part of the project will be published in Ekonomisk Debatt this spring, and the plan is to complete a scientific article including in-depth analysis later. The results from part two are published in IFAU WP 2019: 3 and we have high hopes that it will be published in Labor Economics after requested revisions. The study on access to preschool for children with parents on parental leave will be completed in 2020/21 and published as an IFAU working paper. We also hope for publication in a highly ranked scientific journal.