Alan Barreca, Tulane University; Melanie E. Guldi, Mount Holyoke College; Jason M. Lindo, University of Oregon and NBER; and Glen Waddell, University of Oregon
Running and Jumping Variables in Regression Discontinuity Designs
Barreca, Guldi, Lindo, and Waddell demonstrate that regression discontinuity designs will arrive at biased estimates when attributes related to outcomes predict heaping in the running variable. They discuss several approaches to diagnosing and correcting for this type of problem. Their primary example focuses on the use of birth weights as a running variable, explaining why prior estimates of the effect of very low birth weight classification have been severely overstated. They also consider the use of day-of-birth and student GPA, gaining further insight into the issue, and highlighting that it is important for applied researchers to consider generally when employing RD designs.
Prashant Bharadwaj, University of California, San Diego, and Christopher Andre Neilson, Yale University
Early Life Health Interventions and Academic Achievement
Do early life health interventions affect schooling outcomes later in life? Bharadwaj and Neilson exploit medical recommendations and rules for treating very low birth weight (VLBW) infants, and find that infants who receive extra medical care do better in school. Using detailed administrative data on schooling and vital statistics from Chile, they find that children who receive extra medical care at birth obtain scores that are around 0.18 SD higher in math. In addition, they examine the impact of a specific policy: Chile's national surfactant policy which was introduced in 1998. Because surfactant therapy was especially recommended for VLBW infants, the authors discover that it helped in raising test scores even more. They also find that interventions targeted towards VLBW infants significantly decrease infant mortality. The treated infants are 5 percent less likely to die within a year of birth than the untreated infants. These results are robust to a wide variety of regression discontinuity checks, including checks aimed at detecting (and accounting for) non random heaping of data, which might occur in the case of birth weight.
Sonia R. Bhalotra, University of Bristol, and Atheendar S. Venkataramani, Yale University
The Long-Run Effects of Early Life Pneumonia: Evidence from the Arrival of Sulfa Drugs in America
Bhalotra and Venkataramani exploit the introduction of sulfa drugs in 1937 to identify the impact of exposure to pneumonia in infancy on later life well-being and productivity in the United States. Using census data from 1980-2000, they find that cohorts born after the introduction of sulfa experienced increases in schooling, income, and the probability of employment, and reductions in disability rates. Importantly, these improvements were larger for those born in states with higher pre-intervention pneumonia mortality rates, the areas that benefited most from the availability of sulfa drugs. While men and women show similar improvements on most indicators, only the estimates for men are robust to the inclusion of birth-state-specific time trends. With the exception of cognitive disabilities for men and, in some specifications, family income for men and women, estimates for African-Americans tend to be smaller in magnitude and less precisely estimated than those for whites. The authors speculate that this may be because of barriers in translating improved endowments into gains in education and employment in the pre-Civil Rights Era.
Douglas Almond, Columbia University and NBER; Hilary W. Hoynes, University of California at Davis and NBER; and Diane Whitmore Schanzenbach, Northwestern University and NBER
Childhood Exposure to the Food Stamp Program: Long-run Health and Economic Outcomes
Almond, Hoynes, and Schanzenbach evaluate whether increasing resources available in utero and during childhood improve later-life health and economic outcomes. In particular, they focus on the introduction of the Food Stamps Program, which was rolled out across counties in the U.S. between 1961 and 1975. They use the Panel Study of Income Dynamics to assemble unique data linking adult health and economic outcomes to family background and county of residence in early childhood. The identification comes from variation across counties and over birth cohorts in exposure to the food stamp program. Their findings indicate that the FSP has effects decades after initial exposure. Multiplier effects are largest for health in adulthood. They find weaker positive evidence on adult economic outcomes. Overall, these results are consistent with fetal origins literature on long-run health effects of early life interventions.
David Frisvold, Emory University
Nutrition and Cognitive Achievement: An Evaluation of the School Breakfast Program
Frisvold investigates the impact of eating breakfast through the School Breakfast Program (SBP) on cognitive achievement. Using a regression discontinuity design, he compares the academic achievement of students in schools where the number of free and reduced-price eligible students is just below the mandated threshold to students in schools where the number of free and reduced-price eligible students is just above the threshold. His results show that state mandates increase the availability of the SBP in schools, which increases math, reading, and science achievement in elementary school, particularly for boys.
Niholas J. Sanders, Stanford University, and Charles F. Stoecker, University of California at Davis
Where Have all the Young Men Gone? Using Gender Ratios to Measure the Effect of Pollution on Fetal Death Rates
Sanders and Stoecker estimate the causal impact of ambient prenatal pollution exposure on fetal deaths. Since a complete census of true fetal deaths is impossible to obtain, they exploit the differential in fetal susceptibility to environmental stressers across genders to estimate this effect. Males are more vulnerable to maternal stress in utero, and thus are more likely to suffer fetal death due to pollution exposure. The authors use the Clean Air Act Amendments of 1970 (CAAA) as a source of exogenous variation in county-level ambient total suspended particulate matter (TSPs). They find that a single standard deviation increase in TSPs decreases the percentage of live births that are male by 0.9 percentage points. They then use the observed differences in neonatal and one-year mortality rates across genders in response to pollution exposure to estimate total fetal losses in utero. Their preferred calculations suggest that the pollution reductions from the CAAA prevented approximately 36,000 fetal deaths in 1972.