For instance, Currie and Neidell (2005) examine the impact of air pollution (CO, O3, and Pm10 4 ) on low birth weight. 5 To address this, they used fixed effects models at the individual level, controlling for zip code-month fixed effects. To associate exposureto air pollution with low birth weight, they impute prenatal pollutionexposure in each trimester using a radius of 10 kilometers (km) (6.2 miles) around the meter device. Results show no significant effect on low birth weight when the mother is exposed to air pollution during pregnancy. Similarly, using fixed effects at the individual level, Currie et al. (2009b) examine the effects of pollution (CO, O3, and Pm10) on birth weight and prematurity. For birth weight, they utilize a panel with a pollution monitor and mother locations fixed effects, in which averages of exposuretopollution are imputed for the three trimesters of pregnancy. Results show that a one-unit increase in CO during the third trimester leads to an average birth weight reduction of 16.65 grams. Currie et al. (2009b) regress levels of pollution during the three trimesters of pregnancy to different birth outcomes (including a model for child mortality). These authors use a rich set of controls as well as fixed effects for the closest air pollution monitor, an interaction between the monitor effect and each quarter of the year (to capture seasonal differences), and mother-specific fixed effects to control for time-invariant characteristics of neighborhoods and mothers. Results show that a one-unit increase in CO during the third trimester reduces birth weight on average by 16.65 grams (results were found at lower levels of CO). Currie and Walker (2011) exploit a policy that reduced traffic congestion in the U.S., in which electronic toll collector technology was implemented to look at the effects of traffic congestion on newborn health. This policy allowed them to implement a difference-in-differences design, in which the treatment group is made up of mothers living within two km of a toll plaza, while the control group is made up of those who live close to a highway, but between two km and 10 km of a toll plaza. Results suggest that implementing the E-ZPass 6 is associated with significant reductions in prematurity, by 8.6%, and in low birth weight, by 9.3%. Finally, Coneus and Spiess (2010) present a study using mother fixed effects and year/zip code effects together with an ample set
We propose that the expansion of the Swedish welfare state and the public sector (mainly employing women) starting in the 1960s (Sundin and Willner, 2007) is a potential mediating mechanism for our long-term findings. Figure 8 shows that female employment rapidly increased from about 800,000 employed women in 1950 to about 1,200,000 in 1970 while male employment stayed rather constant over time. While until the 1950s mainly single women participated in the labour market, also married women were integrated in the Swedish labour force in the mid- 1960s (Stanfors, 2003) when our cohorts were about 30 years old. Alongside the development of other parts of the welfare state, there were specifically large investments in publicly provided child care for pre-school children (Datta Gupta et al., 2006). By the expansion of child care a lot of female labour was released (Bergh, 2009). Women could join the labour force due to relief in child raising, but the expansion also generated new workplaces for women – see e.g. Figure 9 illustrating the increase in pre-school teachers among females 1960-1975. Public sector employment thus offered possibilities to combine career and family responsibilities (Datta Gupta et al., 2006). 28 Overall, it seems likely that the expansion of the Swedish welfare state pulled women into the labour force from 1960 and onward. Although education only explains about 10% of the income effect it seems plausible that the better educated females, being more competitive, were the ones that benefited from the expanded demand for women. In comparison to singles or low educated women they were not obliged or “pushed” to work in order to gain their living, but were probably rather career oriented and attracted by attractive “women friendly”
In the first part of this paper, we estimate causal impacts of an intervention targeting infanthealth on cognitive performance in grades 1 and 4 of primary school. Since we also have records of sickness-related absence from school, we are able to (weakly) differentiate the impacts of contemporaneous from early life health on test scores. Our first contribution is hence to a fairly thin strand of the literature that seeks to link early life health interventions to cognitive attainment. Analysis of health interventions include Chay et al. (2009) who study black-white convergence in test scores as a function of hospital de-segregation in America, Bharadwaj et al. (2013) who document impacts of neonatal care on school test scores in Chile, and Bhalotra and Venkataramani (2013) who investigate impacts of early life exposureto a clean water programme in Mexico on cognitive attainment in middle and late adolescence. Accidental exposureto radiation from the Chernobyl disaster is examined in Almond et al. (2009). Other important studies that analyse impacts of early life health rather than of health interventions or shocks, include Black et al. (2007) and Figlio et al. (2014), who use twin or sibling estimators to identify the impacts of birth weight on later outcomes including cognitive performance in Norway and Florida respectively. Like Figlio et al. (2014), we are able to assess impacts of infanthealth on cognitive scores at different ages and by the socio-economic characteristics of parents. However, while they analyse impacts of birth weight, we analyse impacts of an intervention. This is important because, as they state, “While we have strong evidencefrom twin comparison studies that poor initial health conveys a disadvantage in adulthood, we have little information about the potential roles for policy interventions in ameliorating this disadvantage during childhood”; also see Heckman et al. (2014).
with an 8% increase in the Covid-19 death rate.
While the above studies provide useful preliminary evidence, Conticini et al. (2020) and Ogen (2020) offer only geographical correlations between Covid-19 cases andpollutionexposure, whereas Travaglio et al. (2020) take a similar approach but control only for differ- ences in population density and do so across only 7 relatively large regions. Establishing a convincing link between exposuretopollutionand Covid-19 cases requires individual-level data with the ability to control for individual characteristics, such as age and the presence of underlying health conditions. Since individual-level data on Covid-19 infections is not available, the next best alternative is to examine a large number of small geographic re- gions with detailed data on the characteristics of those regions. This allows the researcher to assess whether any correlation between Covid-19 andpollutionexposure still holds once differences in social deprivation, population density, ethnic composition, and other factors are controlled for. While Wu et al. (2020) come closest to doing this, US counties are still relatively large, raising the question of how well such aggregated data can capture the local
largely on adult mortality. This is all the more surprising because studies on adult mortality suffer from conceptual and methodological problems that studies on infant mortality obviate (Chay and Greenstone 2003b). First, the excess deaths attributable to air pollution may occur primarily among the old and infirm, who may not have lived long anyway. Of course, this “harvesting” is also important for studies on infant mortality that identify effects on the basis of short term variation in mortality andpollution. But if decreases in air pollution are found to reduce infant mortality rates overall, the number of life-years gained is large. Second, there is considerable uncertainty regarding the life-time exposure of adults. As Chay and Greenstone (2003b) argue, this uncertainty is greatly reduced by the low migration rates of pregnant women and infants. These advantages sparked a recent interest in the effects of air pollution, not only among epidemiologists but also among economists (see references below).
G Appendix: Child Day Care Expansion
From the early 1940’s there was political debate in Sweden over state provision of preschool child care. At the time there were a few day care institutions run by charity organisations or the community (SOU, 1972)) or informal care. 49 Since 1943 the state provided partial salary funding of educated personnel (Hammarlund, 1998; Westberg, 2015) in day care centres, but these grants covered only about 10 per cent of total costs. In March 1963, the national government considerably extended this, covering a larger share of total costs and funding the training of more child-carers. As a result, many municipalities built new day care centres and employed day carers in the mid 1960’s. The new state subsidy also led to the municipality taking over governance of most of the day care centres previously run by charity organisations (Westberg, 2015). From 1963 to 1970 the number of day care places increased from 10,300 to 29,200. As there remained an excess demand for day care, from 1968, the government implemented an expansion of family day care (familjedaghem), involving a day carer look after smaller groups of children (1-4) in their home (Karlsson, 2002). From January 1, 1969, every municipality was eligible for funding as long as the regional Child Welfare Board approved each centre and the carers are employed by the municipality (SOU, 1972). From 1969 to 1972, the number of places in municipality governed family day care increased from 10,000 to 42,000 (SOU, 1972).
In a recent review of the literature, Almond et al. (2017) argue that the effects of the early life environment on long run outcomes are often heterogeneous,“reflecting differences in child endowments, budget constraints, and production technologies”. Here, we highlight the role of opportunities. While our findings reinforce the small body of evidence demonstrating causal effects of infanthealth on cognitive performance in the school years, they also highlight that the average earnings payoff to cognitive skills is uncertain, being dependent upon distributional effects that may determine higher education choices and on demand conditions. While previous work has discussed changes in the relative demand for female (vs male) labour stemming from recession, war, or technological change, we provide a new perspective, emphasising how expan- sion of broad-based public services will tend to change the relative demand for female labour. This is of potential relevance to understanding historical trends in women’s employment, and the prospects for women in developing countries that are currently witnessing large-scale expansion in the provision of schooling, public health services and, potentially, pre-school centres.
24 as evident in their HFA z-score. Since height is correlated with cognitive ability (Case and Paxson 2008) and given the brawn-based nature of the colonial Indian economy, it is possible that over time, women and girls would have had relatively higher growth rates of cognitive skills with respect to the previous generation of women and girls, but also in relation to the cognitive growth rates of men and boys in the same generation. We see some evidence for this from the 1901 census since Christian males and females had literacy rates that far exceeded those of the Hindu and Muslim populations (Figures 2A and 2B) and growth rates in female literacy that exceeded growth rates in male literacy at least as far back as 1881 (Figures 3A and 3B), also when the population is restricted to those who are 15 years or older (Figure 3C). Could mothers with higher cognitive ability (compared to previous generations) bequeath relatively the same size in HFA advantage to their sons and daughters, but the effect is measured more precisely for daughters in comparison to sons? There is evidencefrom Silventoinen et al. (2003) that this might be the case. If you consider determinants of height, Case and Paxson (2008) argue that about eighty percent is genetic (heritability) whereas about twenty percent is the environment (uterine environment, nutritional status and the disease environment). Genetics change slowly but the environment may be shaped more rapidly and Silventoinen et al. (2003) indicates that the latter is more important for determining height in female populations. In colonial India, the “signal” would have been the strongest in terms of changing and reinforcing the environment that lower-caste mothers faced upon conversion to Christianity; this may thus lead to the reduced “noise” that we document in the child height measures for Christian girls today.
Of course, our study is not without limitations. As with most datasets, vital statistics data do not contain information on authorization status, so we must instead follow the literature by employing a proxy based on citizenship, birthplace, ethnicity, and education. Next, while the reduced form relationship between immigration enforcement and birth weight is itself important, we are unable to precisely identify the mechanisms driving this pattern. Instead, we present suggestive evidence that interior immigration enforcement measures increased stress for likely undocumented pregnant women, a pathway in line with the prior literature on the fear and stress induced by immigration enforcement (Becerra et al., 2020; Cardoso et al., 2020; Potochnick and Perreira, 2010). We also present suggestive evidence on access to care, such as inadequate prenatal care visits and having a midwife instead of a doctor present at delivery. Finally, while we detect a meaningful increase in the likelihood of low birth weight, we are unable to fully measure how in-utero exposureto immigration enforcement may affect these infants throughout their lives (Barker, 1990; Aizer and Currie, 2014; Almond et al., 2018).
year fixed effects, which account for differential outcome trends across states, any state time- varying policies, and the fact that we observe states in different sets of years in the HCUP data. δ c,m × f (y) are county-by-calendar-month-specific trends (e.g., Queens-County-by-January-specific
trends), which we model with a quadratic polynomial. We also control for the average number of mothers (or infants) per 100, residing in zip codes in different quartiles of the median income distribution. We weight all regressions by cell size, and cluster standard errors on the county level. 8 Our model identifies the effects of extreme heat exposure using year-to-year deviations in tem- perature from the county-month trend within each cell. As a concrete example, consider a black woman giving birth in Queens county, New York, in August 2010 and a black woman giving birth in the same county in August 2011. Our empirical strategy leverages the arguably exogenous differ- ence between them in the temperature deviation during their pregnancies from the Queens-specific quadratic trend among all August births, while controlling for the average difference in temperature exposure between all New York state births in 2010 and 2011.
data set contains health information about approximately 45,000 children. Around 28% of the main insured persons received welfare during the complete 12 months after birth. Children in this group born after 1.1.2011 are treated by the welfare reform. The comparison of the sociodemographic characteristics between the two groups reveals that persons on welfare are younger when a child enters the household and more often without German nationality. The main insured person is more often female in the welfare group which might result from more mother lead lone parent households in this group. The welfare households also have more children which might be one reason for their welfare dependency. In the welfare group 18% and in the non welfare group 15% of the children live together with other children who are also born between 1.1.2009 and 31.12.2011. There are slightly more boys in the welfare group than in the non welfare group which indicates a mild boy preference in the welfare group. Twins and triplets occur more often in the non-welfare group which might relate to the more frequent use of fertility treatment in this group.
The dependent variable in all regressions is the state-level infant mortality rate. Public Health Expenditure = expenditure on medical & public health, water & sanitation, and nutrition (% of state GDP); Female Literacy = proportion of women aged 15 years and more who can read, write and carry out simple arithmetic calculations; PCNSDP = per capita net state domestic product (at 2004 − 05 prices); Sex Ratio = females per 1000 males; Urbanization = proportion of population living in urban areas. P-values (clustered by state) appear in parenthesis below estimates; *p<0.10, **p<0.05, ***p<0.01. Specification 1 is estimated by OLS. For all other specifications public health expenditure has been instrumented with “own” tax revenue, “own” non-tax revenue, and effective number of parties in the government. Models 1 through 6 have been estimated by 2SLS; model 7 has been estimated by LIML. The KP (F-stat) refers to the Kleibergen-Paap rk Wald statistic for weak identification; the J-stat refers to Hansen’s overidentification test.
pop 15-64 102 57.30 6.71 46.33 76.81
intensive industrial units. Urbanization also provides a unique opportunity for people to access politicians and policy makers, which may be not the case in a country with a higher share of the rural population [Torras and Boyce (1998), Rivera-Batiz (2002) and Farzin and Bond (2006)]. It is, however, unlikely that the benefits of urbanization outweigh its negative consequences. The second demographic variable is population density. It is often emphasized that a high population density leads to an unsustainable exploitation of the environment [Hilton and Levinson (1998)]. The age structure of population, in particular the working age population (15-64 years old), can have an effect on the environmental degradation. Some scholars refer to the positive role of this part of the population in reducing pollution, while some others oppose this view. Farzin and Bond (2006) have explained these different views on the effects of age composition on environmental quality. They point out, for example, that young people can bear more risks of environmental pollution compared to the older part of population. The youngers have a higher option value of waiting for future improvements in environmental quality. Older people feel the health problems of pollution more directly and are willing to put higher pressures on the government for stricter environmental regulation. Older people also may have more spare time to participate in local NGOs, supporting environmentally-friendly policies by the government. Thus, on the basis of Farzin and Bond argument, we expect that a higher share of the working age population on total population (pop 15-64 ) increases the environmental degradation.
A possible source of bias in our estimates arises in our inability to observe the characteristics of individuals who died before they could be observed in our data. As a robustness check, we use the 1980 census—as mentioned in the data section, it is more “representative” of the cohorts born in the colonial period—to re-estimate the malaria exposure effect on education. 24 The OLS results are reported in Panel A of Table 5, while the 2SLS results are reported in Panel B. The dependent variable is the dummy variable indicating whether an individual received any schooling. The OLS regression results from the Census data are consistent both in signs and sizes with the results obtained from our original panel data. The 2SLS results are also consistent in sign and are slightly larger than those obtained from our panel data.
realized in the short to medium term if individuals feel less healthy due to ozone exposure. In such cases they may be expected to use health services, decrease their labor supply and/or suffer losses in productivity, even if the impact on objective health is limited. Conversely, a moderate deterioration in objective health may not lead to economic cost in the short to medium term, if the individual generally feels well. Evaluating the effect of ozone on subjective well-being is also of interest, as it allows us to evaluate indirect effects of pollution: While pollution may not significantly affect a generally healthy adult directly, it may affect the individual indirectly through others sharing the same household, e.g. children affected by ozone. Such indirect effects may affect the subjective well-being of the individual. Such changes in well-being may also have economic consequences, such as absences from work to care for affected relatives or losses in productivity due to well-being effects, even if the person’s own health is not directly affected by ozone concentrations.
Cancer (IARC) concluded that there is sufficient evidenceto classify outdoor air pollution, in particular, PM as carcinogenic (Loomis et al. 2013).
With reference to urban air quality in Asian cities, a rise in types and number of emission sources of air pollutants due to the intensive growth in industrial sector, urban population and automobile vehicles in the region has resulted in poor air quality with levels many times higher than WHO guidelines and standards in developed countries (Gurjar et al, 2008; Hopke et al, 2008). A comprehensive review on ambient air pollutionandhealth in developing countries of Asiaby HEI International Scientific Oversight Committee (2010) has concluded that although there has been some improvements in ambient air quality the level of air pollutants in different cities in Asia are well above WHO guidelines. They have also argued that the public health impact of ambient air pollution will grow over periods of time due to demographic and epidemiological transitions and increased prevalence of risk factors.
Regarding our findings on the health shocks of air pollution, we first report that our first stage results indicate that wind speed is strongly negatively correlated to PM10, suggesting that the local air quality improves when heavy winds blow. Results for the second stage con- firm that PM10 is harmful for health, consistent with previous studies, and does create a health shock in the SPMA. In fact, hospitalization rates for respiratory diseases rise when exposureto air pollution increases for children aged one to five years. We also find suggestive evidence that the length of hospital stay might be shorter, but the estimates are not statistically sig- nificant. Hospital admissions for asthma and pneumonia for that age group follow a similar pattern as for the overall hospitalization rate for respiratory diseases. On the other hand, ad- missions for influenza may be driven by infants. We caution, though, that our instrument may not be ideal for this last group because young children under one year of age may get less ex- posed given that they stay indoors more often. OLS estimation disregarding the endogeneity of exposureto air pollution seem to underestimate the parameters of interest in our context, and robustness checks support our main findings on the health impacts of air pollution. 8
This paper investigates effectiveness of an infant intervention that, in addition to clinical services, provided universal access to “soft” inputs including information, monitoring and support with a large home-visiting component. The intervention was implemented as a trial in the early 1930s in Sweden, and was a precursor to legislation that rolled it out nationwide from 1937 as part of the expansion of the welfare state. Using birth certificate data from parish records matched to death registers, we find that the programme led to substantial improvements in infant survival andto improvements in longevity over and above this. This is consistent with the hypothesis that infant morbidities create structural changes that predict the onset or progression of chronic disease, but that are latent earlier in the life course. We estimate that an individual with the average duration of program exposure in infancy experienced a 23% reduction in infant mortality risk and a 6.5% reduction in the risk of dying by the age of 75 (an age by which almost 40% of their cohort had died). The results contribute new evidence on the potential benefits of postnatal care programs that are being introduced or modified in many contemporary settings. They are of particular relevance in today's poorer countries, which are carrying the double burden of infectious disease and chronic diseases.
ozone) as 150 equivalent gasoline cars. 3 Hereafter, we refer to cars with “clean diesel” technology as cheating diesel cars.
We exploit the dispersion of these cheating diesel cars across the United States as a natural ex- periment to measure the effect of car pollution on infantand child health. This natural experiment provides several unique features. First, it is typically difficult to infer causal effects from observed correlations of healthand car pollution, as wealthier individuals tend to sort into less-polluted ar- eas and drive newer, less-polluting cars. The fast roll-out of cheating diesel cars provides us with plausibly exogenous variation in car pollutionexposure across the entire socio-economic spectrum of the United States. Second, it is well established that people avoid known pollution, which can mute estimated impacts of air pollution on health (Neidell, 2009). Moderate pollution increases stemming from cheating diesel cars, a source unknown to the population, are less likely to induce avoidance behaviors, allowing us to cleanly estimate the full impact of pollution. Third, air pol- lution comes from a multitude of sources, making it difficult to identify contributions from cars, and it is measured coarsely with pollution monitors stationed only in a minority of U.S. counties. This implies low statistical power and potential attenuation bias for correlational studies of pollu- tion (Lleras-Muney, 2010). We use the universe of car registrations to track how cheating diesel
impact of air pollution on infant mortality in Mexico City. Finally, Greenstone and Hanna (2011) examine the impact of air and water pollution regulations on infant mortality in India.
This paper extends the growing literature on air pollutionandinfant mortality in developing countries by using a novel source of variation from Turkey. In particular, we exploit the widespread replacement of coal by cleaner-burning natural gas for residential and commercial space heating and cooking in Turkish provinces, which has been made possible by the expansion of a network of natural gas pipelines originating from Russia in the 1980s. 2 Burning of natural gas emits virtually no sulfur oxide, which is a key component of acid rain. Emissions of total particulate matter, carbon monoxide, and nitrogen oxide, which cause acid rain, smog, and contribute to global warming, are also at much lower quantities from burning of natural gas compared to burning coal. First introduced to households and businesses in Ankara in 1988 and followed by Istanbul and Bursa in 1992, the number of provinces with natural gas has grown rapidly over the last two decades. As shown in Figure 1, 61 of the 81 Turkish provinces now have access to a natural infrastructure. On the one hand, the expansion of natural gas networks in Turkey has caused considerable improvements in air quality by reducing emissions of pollutants such as particulate matter, sulfur dioxide, carbon dioxide, and nitrogen oxides. On the other hand, Turkey has also experienced significant reductions in infant mortality rate during the same period. For example, between 1990 and 2010, infant mortality rate (per 1,000 live births) decreased from 33 to 10. In fact, Turkey has already surpassed one of the Millennium Development Goals, which called for a reduction of the under-five year old mortality rate by two thirds between 1990 and 2015. Nonetheless, Turkey still has the second highest infant mortality rate among the OECD countries after Mexico.