The Household Income – Child Health Gradient: Evidence from Ukraine
by
Gavryliuk Olga
A thesis submitted in partial fulfillment of the requirements for the degree of
Master of Arts in Economics
National University “Kyiv-Mohyla Academy” Master’s Program in Economics
2008
Approved by
Mr. Volodymyr Sidenko (Head of the State Examination Committee)
Program Authorized
to Offer Degree Master’s Program in Economics, NaUKMA
Date
National University “Kyiv-Mohyla Academy”
Abstract
The Household Income – Child Health Gradient: Evidence from Ukraine
by Gavryliuk Olga
Head of the State Examination Committee: Mr. Volodymyr Sidenko,
Senior Economist Institute of Economy and Forecasting, National Academy of Sciences of Ukraine
Previous studies have shown that there is a significant impact of family income on the child health. Nonetheless, such relationship is not universal for all the countries. This work estimates such relationship for Ukraine and finds it significant. However, the sign of the relationship depends on the family type. As part of child health – income gradient, this work investigates impact of income on chronic diseases. Results reveal that income does not cause chronic conditions and there is only partial evidence that it can affect the degree of harmful influence of chronic disease on health.
Table of Contents
Chapter 1. Introduction…………………………………………… ..1
Chapter 2. Literature Review………………………………………...4
Chapter 3. Methodology…………………………………………….10
Chapter 4. Data Description………………………………………...14
Chapter 5. Empirical results…………………………………………18
Base health-income gradient model estimates…………………18
Psychological factors estimates………………………………..23
Chronic condition models………………………………….…24
Chapter 6. Conclusions………………………………………..….…28
Bibliography………………………………………………………...31
Appendix .………………………………………………………… 33
List of figures and Tables
Number Page
Figure1. Distribution of child health over the child age……………….15
Figure 2 The household income and child health by age groups………………16
Table 1. Summary statistic……………………………………………………17
Table 2. Influence of income on child health. Probit regressions……………...21
Table3. Impact of physiological perception of economic problems on
child health. …………………………………………………………….……24
Table 4 Influence of income on chronic conditions……………………..……25
Table5. Influence of chronic conditions on health: role of income……………26
Table A1. Construction of variables…………………………………………..33
Table A2. Comparison of basic regression for two samples……………...……35
Table A3. Impact of reported family expenditures on child health……………36
Acknowledgments
The author wishes to express her sincere appreciation to thesis supervisor,
Dr. Olena Nizalova, for her guidance and help, patience and support, optimism and strong believe in the author and her research.
The author wishes to thank to Dr. Tom Coupe for his responsiveness and valuable comments and to all research workshop professors for their patience in reading thesis.
Thank to Ganna Bielenka for friendly help. The author is deeply grateful to parents for their support at every day of the life.
Glossary
Health-income gradient – positive relationship between income and health; empirical fact that people with higher income also have better health.
Liquidity constraints - problems for family to find financial recourses in the very short term
Chronic conditions - a disease that is long-lasting or recurrent
32
Chapter 1
Introdiction
A strong and positive relationship between health and economic status, called income-health gradient (or income gradient in health), has been documented for many countries. This is, of course, not surprising. According to the health production model, income can influence health directly through better investment in health, and also indirectly by increasing the efficiency of other inputs to health production. Indeed, people with higher income can afford not only better health services, but also better food, accommodation, education, all of which has its own influence on health.
Of course, while income influences health, health also influences income. People may have lower education and productiveness due to health problems, both of which have a direct impact on income, resulting in endogeneity problem. Theoretical and empirical discussions of the causality question still remain unsolved as summarized by Smith (1999).
In the context of the health-income relationship researchers have also investigated the effect of income on child health (Case et al. 2002, Currie and Stabile 2003, Currie et al. 2004, and others). For child health, the endogeneity problem is not as straightforward as for adult health, as children do not earn income directly. This makes the effect of income on child health easier to identify compared to the effect of income on adults’ health. The impact of income inequality on child health is an important policy issue because health in childhood has a strong influence on the health in adulthood, on education, and therefore on labor market outcomes (Conely et al. 2001). The impact of economic status on health is especially strong during critical periods of growth and development of a human body (Smith, 1999). For instance, health in early childhood can predict cardio vascular, neurological and respiratory diseases, all of which are common in Ukraine. Hence, children’s health has a long lasting influence on the health at middle and older ages as well as important economic outcomes.
This work investigates the impact of current income on child health based on data on Ukrainian households. As has been explained above, studying the influence of income on child health is important for a country. Studies for Great Britain, Sweden, United States, and Canada have shown that income gradient and its changes for different age groups vary across different countries. That is why it is difficult to apply results from other countries to Ukraine and hence a country specific study is necessary.
In Ukraine, over the last fifteen years income and life expectancy has dropped substantially. Health care and social systems have seriously deteriorated. These tendencies increase the importance of reform in these spheres. This study can show most important determinants of child health and that is why to have important policy implications. In particular, this research aims to answer questions whether income has an influence on child health and degree of health income inequalities, whether such influence is the same for all children and constant over the child age, what parent’s characteristic’s can decrease influence of family income. The answers to these questions will help to increase efficiency of various policy measures.
The rest of the study is organized as follows. Chapter 2 discusses literature on general health-income gradient as well as specific issues concerning child health-income relationship. Chapter 3 gives some theoretical background of the model, well developed in other studies, and presents empirical models estimated in this work. The next chapter focuses on data description and primary analysis of child health-income gradient. Then various estimations of different modifications of the model are presented in Chapter 5, which is followed by conclusion.
Chapter 2
Literature Review
A great number of empirical studies have investigated the relationship between income and health in different countries. This review briefly discusses how income can influence health and how health can influence income in general health-income relationship. Then the discussion will switch to child health-income gradient. In this part, the main differences in findings on gradient and its behavior over the child age are presented. After that, the review discusses some of the specific questions related to income measurements and instruments.
The researches have shown that there is a positive significant association between income and health: people with higher income on average have better health (Marmot, 1999, Deaton and Paxson, 1998; Mirovsky and Ross, 1999; Robert, 1999; Smith, 2004 and others). In order to investigate income-health relationship, researchers use various measures of health and income. Health is measured by self-rated health, mortality, specific health problems, chronic conditions as well as psychological health. Income is measured as current, permanent income (Ettner, 1996, Meara, 2001, Meer et al., 2003), wealth (McDonough et al., 1997; Feinstein, 1993; Smith, 1999). Endogineity problem leads to application of various instruments for income, such as unemployment rates, parental education, partner’s characteristics, work experience, inheritance (Ettner, 1996, Meer et al., 2003, others). However, the results are ambiguous, and researchers still argue what is the cause and what is the consequence.
The traditional views on how income influences health are risk behaviors that are more common for lower income and less educated groups of people (Marmot, 1999) and worse access to health care. However, an increase in well being over the decades, has led to the changes in main health risk factors and causes of death (Smith, 1999).The main risks are not already severe epidemics and infections but chronic conditions and life-style risks These new circumstances have required new explanations of health-income gradient. Among them are negative impacts of low income during the critical periods of development, such as childhood (Smith, 1999). Indeed, poor health and nutrition in early childhood has long-lasting impact on health. Researches also emphasize on psychological factors relate to income level that can influence health such as social position and income inequality (Marmot, 1999).
Health can also influence income, wealth, and economic status. First, it can influences expenditures on health care directly or on insurance. For people with poorer health the probability of great health care expenses is higher, so people may give up part of their wealth in order to offset possible future health expenditures. What is more, health status can influence labor market decisions. Healthy people can work longer hours and be more productive (Wu, 2003).
There can be also a third possibility. Some other factors affect both health and income. For example, people who value future earnings and health less than present can invest less in the human capital and that is why have worse health and income in the future (Barsky et al., 1997).
As mentioned above, health at some critical periods of life, such as childhood, can have long-lasting impact on person’s health and income gradient. That is why some of the researches focus on child health-income gradient. Moreover, investigating child health-income gradient allows partial mitigation of the endogeneity problem, because a child does not participate in labor market.
The main question in focus of such types of the research is the magnitude of income gradient and the changes in the gradient over the child age. All researchers agree that income has indeed significant impact on child health. However, findings on changes of income gradient over the child age differ among the studies. Research on Canada and the USA find a significant income gradient, which increases with child age (Case at al. 2002, Currie, 2003). In order to check the robustness of the result, authors add important variables that could reduce the importance of income. Adding health at birth and chronic conditions do not eliminate the effect of income inequality on health. Even more, the presence of chronic conditions increases the role of income in maintaining health status (Case, 2002). Children from the households with different socio-economic position are able to overcome a health shock, namely, chronic conditions, with similar success. However, children with low socioeconomic status have worse health because they have more shocks to their health due to life-style, nutrition, worse detection of disease, etc. (Currie, 2003).
Finally, parents’ health is added to a model. Parents’ health and income are highly correlated: genetically healthier parents can have healthier children and have higher income. That is why samples of adoptive and own birth children are separated in order to investigate the effect of genetic ties. Nonetheless, the difference in income effect for different parental types is not found, suggesting that the income effect is not driven by parents’ health as an omitted variable (Case, 2002).
The research considered above finds a growing role of income in determining health for older children. In contrast to the studies discussed above, research on Great Britain, Ireland, and Sweden, shows little difference in income gradient for different age groups of children (West, 1997, Currie 2004; Nahum, 2007).
First works comparing the association between the class status of a family on different health outcomes, for young males and females are done without computing income gradient (West, 1997). It is found that health inequalities due to the socioeconomic position decrease in youth. More complex studies confirm the result for these countries showing lower magnitude of income gradient and decrease of income gradient in youth (Currie at al. 2004; Nahum, 2007). Nonetheless, these results can be sensitive to the sample size construction of variables and composition of the sample (Case et al., 2007).
Different explanations for an existence of such income-health gradient are provided. First of all it is secondary school influence that can reduce the effect of home background and neighborhood by promoting cross class relations and activities, which can influence health directly, or indirectly through creating common youth culture and values (West, 1997). Another possible explanation is free access to efficient health care (National Health Service) (Currie et al. 2004)
The size and significance of the income-child-health gradient depend on the measure of income used. All the researchers find a significant impact of current income. However, average or permanent income appears to have stronger influence for all age groups, supporting permanent income hypothesis (Curtis et al. 2000, Case et al. 2002). On the other hand, limitations on very short run income also appeared to have a significant effect. Liquidity constraints (problems with finding financial recourses in the short term) increase the probability of a child having worse health more than just current income. This fact can be explained by the influence of liquidity constraints on living conditions and parents’ worries about financial situation (Nahum, 2007).
While all the works mentioned above documented important findings, still there can be a problem in the research of this kind. Child health can in fact influence family income. In their work Case et al. (2002) mention that for children the effect of child health is at least not as direct as for adults, as children do not earn money. However, children with bad health may need more attention and some specific treatment, often provided by parents, which may negatively influence family income. Following Case et al. (2002), most authors do not pay much attention to the endogeneity problem. That is why instruments are not widely used in child health-income gradient research. One of the attempts has been to use grand parents smoking as an instrument for parent’s income (Doyle et al., 2007) The problem is that this instrument may have direct influence on child health, not only through income, so this instrument also may give inconsistent result.