Research 

Abstract: This paper examines whether the expansion of mobile broadband access affects birth outcomes among young mothers. The introduction of the iPhone in 2007, and its exclusivity to AT&T through early 2011, generated large geographic differences in access to mobile broadband. We combine the universe of U.S. births to mothers aged 10-24 over 2000-2015 with county-level AT&T coverage data in a difference-in-differences design. Greater coverage after 2007 reduced average birth weight by roughly 3.9 grams per standard deviation, reduced fetal growth and increased the incidence of low birth weight and small-for-gestational-age births, while having little effect on gestational length or preterm birth. The effects emerge after 2007 and grow as smartphone adoption diffuses, with limited evidence of pre-trends in the event studies. We also find that mobile broadband exposure altered prenatal-care utilization and smoking behavior along with dropping their sleep duration by 16 minutes among young women, and changed the observable composition of women giving birth. The results are robust to a set of falsification exercises, including placebo treatment dates and alternative carrier coverage measures, which yield little evidence of comparable effects outside the AT&T coverage setting. The findings show that the consequences of the smartphone revolution extended beyond fertility and into the health of the next generation.

Abstract: This paper examines the effects of the 2005 Texas Medicaid funding cutbacks on prenatal care utilization, birth outcomes, and infant mortality. Using the universe of U.S. natality and mortality records from 2000 to 2010, we implement event study and difference in differences designs comparing Texas to other states. The cutbacks led to sizable reductions in Medicaid receipt, healthcare utilization, and prenatal care, including fewer visits and delayed initiation of care. These disruptions were accompanied by declines in birth weight and increases in adverse measures of health at birth, including low birth weight, very low birth weight, extreme low birth weight, and very preterm birth. Infant mortality rose by roughly 7 percent in respect to average infant mortality. Heterogeneity analyses show that these adverse effects occurred across all racial, ethnic, and educational groups, with especially large percentage deteriorations in severe outcomes, such as very preterm birth and extreme low birth weight, among Black and Hispanic mothers. Overall, the results demonstrate that Medicaid retrenchment meaningfully undermined early life health and underscore the importance of stable public insurance coverage during pregnancy.

Abstract: This paper evaluates the welfare effects of Medicare Part D in the United States. A central question is whether the benefits of public prescription drug insurance justify its fiscal cost. Using Medical Expenditure Panel Survey (MEPS) data from 2000 to 2012 and an Instrumental Variable (IV) estimation, we estimate that Part D improves health by 0.017 quality-adjusted life years, reduced out-of-pocket prescription drug spending by about $490, and increased total prescription drug spending by about $432. These estimates imply that Part D improved financial protection while also increasing access to medications and generating measurable health gains. To translate these reduced-form effects into welfare terms, we incorporate the estimates into two complementary frameworks; a complete-information approach and an optimization-based approach. These frameworks allow for a direct quantification of recipients’ willingness to pay for Medicare Part D relative to the public cost of provision. The gross government cost per recipient is estimated at $3,221, while the net resource cost is $922. The difference, about $2299, reflects transfers to private companies who provide subsidized insurance. We estimate the willingness to pay for the Medicare part D is between $1826.51 and $1281 per year, and marginal value of public funds is between 1.389 and 1.9806 based on the different approaches.

Abstract: Can schooling protect against mental health problems decades later in life? We address this important policy-relevant question for older adults in Brazil, China, India, and Mexico using aging studies modelled after the US Health & Retirement Study and a nonparametric partial-identification approach that bounds the causal effect. We document important effects at the lower part of the schooling distribution. An extra year of schooling reduces the probability of likely being depressed (depressive symptoms) by 0.0-3.3% (0.0-0.03 SDs) in Brazil; 0.3-5.5% (0.01-0.06 SDs) in China, 1.1-3.8% (0.01-0.04 SDs) in India, and 0.6-6.2% (0.01-0.07 SDs) in Mexico. We further investigate heterogeneity by gender and social background and estimate effects of schooling on potential mechanisms (general and cognitive health, SES). We find suggestive evidence that the effects of primary schooling may (1) be working through better cognitive health and higher SES, (2) differ by gender in India, and (3) differ by social background in China and India. At other schooling margins (e.g., increasing schooling from lower- to upper-secondary, or from upper- to post-secondary), we cannot statistically distinguish between null, small, moderate and large effect.