Health is undoubtedly an important dimension of well-being. As soon as child well-being is concerned, health is even more important because early life health is a key determinant of future health as well as future economic outcomes.
The direct impact of health policy on the health of children is studied since long. Less is known about the indirect impact of cash transfers to parents on the health of their children. The most important cash transfer policy in the US is the earned income tax credit (EITC). This policy was introduced in 1975 and it has been expanded several times. Since the 1996 reform, no less than 25 millions of households benefit from EITC, for a total cost of above $50 billion. The targeted households, though, remain the low- to moderate-income families with children, especially single mother families.
The EITC works as follows. When the fiscal authority computes the tax of a given household, based on the income earned the year before, an income subsidy, or tax credit, is added to the earned income. The amount of that subsidy is a fraction of earned income when it is low (36% for taxpayers with one child and 40% for taxpayers with two or more children), then the subsidy is constant (at $3,094 for one child and $5,112 for two or more children families, statistics of 2011), and for yet larger incomes the subsidy is decreasing.
The EITC has two direct impact. First, it increases the after tax income of the poor households. Second, it incentivizes people to find jobs, as it amounts to increase the wage rate by 36% to 40% for low-skill workers.
Hilary Hoynes, Doug Miller and David Simon (2015) propose a detailed study of the indirect impact of EITC on the health of infants. More precisely, they focus on single mothers, aged 18-45, with a child under 3, and who have a high school degree or less. Note that 42% of them benefit from EITC. For those families, the authors study the impact of benefitting from EITC on low weight birth, that is, weighing less than 2,500 grams. Low weight birth is highly correlated with early mortality and other health or socio-economic outcomes.
In principle, there are many channels through which benefitting from EITC may impact weight at birth. First, richer mothers are likely to spend more on goods they like. Infant care is certainly such a good, but smoking, drinking and eating low quality food can also be stimulated by higher income while being detrimental to the development of the fetus. Second, EITC is also supposed to incentivize people to look for and accept jobs, and many studies have shown that the labor supply of low skill workers has indeed increased. Having a job or working longer may decrease the time a mother has for her children or divert her attention away from her pregnancy. It may also play the role of increasing the social pressure, for instance to smoke less or to eat and drink in a way that is better for the fetus. Third, EITC, because it increases available incomes, may push single mothers above the income threshold that make them benefit from Medicaid, the free public health insurance program. As a result, mothers may be disincentivized to see doctors during their pregnancy.
All this shows that the indirect impact of EITC on infant health in general and low weight birth in particular cannot be predicted with certainty. It has to be empirically estimated. This is what the authors do in their paper. Having one more child increases the amount of EITC one is eligible to, but that amount, as explained above, will only become available the year after at the time of the tax refund. As a result, because low weight birth depends on the health and behavior of the mother during her pregnancy, EITC can only impact the health of the second born, or the third born, etc. This is why the authors concentrate on the effect on low weight birth of an increase in income due to the EITC program during the last three months of pregnancy.
In their main estimation, the authors compute the low weight birth rate in each State, for each year between 1991 and 1998, as a function of the rank of the child among siblings and the demographic group of the mother. The authors then estimate whether the difference in low weight birth rate between first born and later born children was larger after the 1993 reform that increased the amount of EITC.
The main result is that EITC entails a significant decrease in low weight birth. No less than 10.2% of the children of those single mothers had low weight at birth. This fraction decreases by 0.35 percentage points for second or later born children after the 1993 reforms. The effect is larger for the third born child than the second born. The effect is larger for children of black women. This means that positive effects dominate negative ones. The authors propose many variants of this estimation strategy. All results converge to the same effect.
Then the authors proceed with an inquiry into the possible channels. They point out several effects, even if the source of those effects cannot be clearly identified. Income increases due to EITC lead single mothers to smoke less. Whether it comes from a decrease in stress, an increase in the social pressure on the work place to quit smoking or from another source is still unknown. Mothers also drink less alcohol. As expected, many women loose their right to public health insurance. Many of them, though, substitute it with private insurance, which is often linked with the job they get. The overall insurance coverage stays stable. Nonetheless, prenatal care increases, which certainly favors the development of the fetus, and which, through doctors’ recommendations, may itself explain the changes in mothers’ behaviour.
The conclusion of the paper is that a more generous EITC policy, either through the increased income or through the increased labor force participation of single mothers, has a positive indirect impact on the health of their infants. This is good news. It also calls for taking the situation of the mothers seriously into account in the design of policies dedicated to improve the health of the children.
 Hoynes, Miller and Simon, ‘Income, the Earned Income Tax Credit, and Infant Health,’ American Economic Journal: Economic Policy 2015, 7(1): 172–211.
 Technically, they do a difference-in-difference regression, controlling for State and year economic variables, such as unemployment rate and welfare reforms, and adding State, year, rank in siblings and demographic group fixed effects.