Research
Working Papers:
The Hidden Cost of Firearm Violence on Infants in Utero (NBER Working Paper No. 31774)
(with Janet Currie, Michael Hatch, and Erdal Tekin)
Abstract
Firearm violence is a critical public health crisis in the U.S., marked by a significant number of homicides involving firearms, including indiscriminate shootings in public spaces. This study investigates the largely unexplored impact of such violence on newborn health. We adopt two approaches. First, we analyze the 2002 'beltway sniper' attacks in the Washington DC metropolitan area, using administrative birth records with maternal residential addresses in Virginia. The beltway sniper attacks, a series of random shootings in the Washington DC metropolitan area, caused widespread terror and disruption over three weeks. Leveraging both spatial and temporal variation, we compare outcomes of children exposed to the attacks in utero due to timing or having a residential address near a shooting location to those who were not exposed. Second, we investigate the impact of in-utero exposure to mass shootings on infant health using restricted-access U.S. Vital Statistics Natality records between 2006 and 2019 and leveraging variation in the timing of mass shootings in counties with at least one shooting. We document substantial, previously overlooked costs on pregnant women and infants. Exposure to the beltway sniper attacks during pregnancy increased the likelihood of very low birthweight and very premature birth by 25%. The analysis of national mass shootings confirms these findings with slightly smaller effect sizes. These results emphasize the need to consider the broader impact of violence on vulnerable populations when assessing the cost of firearm violence. The estimates suggest additional costs of $155 million (2023 dollars) for the beltway sniper attacks and $75 million annually for mass shootings.
Juvenile Courts and Recidivism
(with Nils Braakmann and Diego Zambiasi)
Abstract
Many countries operate special courts for offenders below a certain age. These courts have often been criticised, as juveniles sentenced by youth courts display high recidivism rates. Using administrative data, we exploit a discontinuity in the assignment to juvenile courts at age 17 in the UK in a fuzzy regression discontinuity design to study the causal effect of juvenile, rather than adult, court convictions on recidivism. While recidivism rates are high in all courts and juvenile and adult court sentences differ in behaviour, we find no evidence that juvenile court convictions increase recidivism for either first or serial offenders. These findings suggests that juvenile courts do not impact the likelihood of recidivism.
Energy Price Shocks and the Demand for Energy-Efficient Housing: Evidence from Russia’s Invasion of Ukraine (IZA Discussion Paper No- 15959) (R&R, Economica)
(with Nils Braakmann and Harry Pickard)
Abstract
This paper examines how the energy price shock post-Russia’s invasion of Ukraine in 2022 affected property transactions. Using administrative data on transaction price and energy efficiency in the UK, we explore if buyers adjust behaviour in response to greater energy costs. Using hedonic difference-in-differences models, we find that prices increases by £1,000 for high-energy efficiency properties and decreases by £2,400 for low-rated ones. Properties with
higher potential energy efficiency also see value upticks. We also find that high gas-consuming neighborhoods experience fewer house sales post-invasion. The size of the effect is explained by buyers perceiving the energy price hike as temporary. Our results shed light on the interplay between energy prices and the property market, offering insights into buyer behaviour and the valuation of energy efficiency.
Spatial Inequality in Unsolved Crimes: Evidence From Small Neighbourhoods (R&R, Journal of Regional Science)
(with Nils Braakmann and Diego Zambiasi)
Abstract
Using administrative data on the universe of police recorded crime linked to judicial outcomes for England and Wales from January 2013 to December 2018, we document – for the first time – large and persistent spatial inequalities in the proportion of solved and unsolved crimes across small neighbourhoods covering a whole country. We find substantial differences across neighbourhoods in the same municipality or police force. Fixed effects decompositions suggest that neighbourhoods have different clearance rates across different crimes and that high-crime neighbourhoods also have high clearance rates, but with substantial heterogeneity across offences. Clearance rates correlate systematically with neighbourhood composition.
Publications:
Mandatory Schooling of Girls Improved Their Children's Health: Evidence from Turkey's 1997 Education Reform (NBER Working Paper No- 23492, Accepted, Journal of Policy Analysis and Management)
(with Resul Cesur, and Inas Rashad Kelly)
Abstract
The impact of maternal schooling on child health is examined using an education reform that increased compulsory schooling. The analysis employs two large data sets from Turkey, a developing country where female education is stigmatized. Results show that an increase in mother’s schooling improves child health at birth (e.g., reduces low-birthweight and premature birth) and lowers child mortality. This research arguably provides the strongest evidence to date in favor of the argument that mandating girls to attend school may have substantial non-pecuniary benefits in terms of the health of the offspring in developing countries with low female schooling.
Curriculum Reforms and Infant Health (Forthcoming, The Review of Economics and Statistics)
(with Ozkan Eren, and My Nguyen)
Abstract
This paper examines the effects of high school curriculum reforms on infant health by exploiting sharp and staggered changes across states in core course requirements for graduation. Our results suggest that curriculum reforms significantly reduced the incidence of low birth weight and prematurity for black mothers. For white mothers, the estimated effects are small and generally insignificant. Improvements in maternal health behaviors and labor market outcomes appear to explain a non-negligible fraction of the observed effects. Finally, we calculate a large social gain induced by favorable infant health outcomes. Several robustness checks and different placebo tests support our findings.
The Role of Education on Unemployment Duration (Forthcoming, Economic Inquiry)
(with Duha T. Altindag, and Elif S. Filiz)
This paper investigates the role of education in the job search process of the young unemployed workers. Exploiting the variation in education induced by a reform that caused a dramatic increase in the exposed cohorts’ educational attainment, and using data obtained from administrative unemployment insurance (UI) records, we identify the reduced form impact of education on unemployment duration. We show that high-educated individuals, compared to their low-educated counterparts, stay unemployed longer, and they are less likely to transition into employment before their UI benefit periods expire, suggesting education may be increasing one’s selectiveness over jobs. This difference in unemployment duration between the high- vs. the low-educated diminishes when the jobs are scarce, i.e., during the recessionary periods and in the regions of the country where the unemployment rate is high. In a supplementary analysis, we show that an extension of UI benefit period causes a differential impact on workers’ unemployment duration based on their educational attainment. Particularly, a longer benefit period increases the unemployment durations of the high-educated workers more than the low-educated. In addition, a prolonged benefit period increases low-educated unemployed workers’ probability of finding a job, but it has a null effect on the high-educated. Our findings highlight the importance of taking worker characteristics into consideration when designing the UI system.
The Impact of Education on Health and Health Behavior in a Middle-Income, Low-Education Country (NBER Working Paper No- 20764) (Economics and Human Biology. September 2018, 31: 94-114)
(with Resul Cesur and Naci Mocan)
Abstract
Although the impact of education on health is important for public policy everywhere, the overwhelming majority of research identifying the health returns to education has focused on developed countries. We use data from multiple waves of nationally-representative Health and Tobacco Surveys in Turkey, and exploit an education reform that increased the mandatory years of schooling from 5 to 8 years in 1997. Using exposure to the reform as an instrument for completing at least eight years of schooling, we examine the impact of education on health indicators and smoking among young adults. We find that extending schooling on this margin impacts men and women differently. Our results indicate that while a one-year of extra schooling increases the likelihood of being obese among males by 9.9 percentage points, the same increase in schooling improves the probability of women being in the healthy weight range by 15.5 percentage points. Consistent with this result, an extra year of education increases women’s propensity to self evaluate their health as excellent by 4.3 percentage points. Additional analyses reveal that education makes men (but not women) more likely to spend time on computers, using the internet, and to spend time on social media, suggesting that differential time allocation between men and women, triggered by enhanced education, may be a mechanism behind the differential results between the sexes. Education has no impact on smoking for men or women regardless of the measure of smoking.
Transforming Lives: The Impact of Compulsory Schooling on Hope and Happiness (Journal of Population Economics. July 2016, 29: 3 )
(with Resul Cesur)
Abstract
This is the first article examining the causal impact of extended primary schooling on happiness of young adults. We rely on a law change that raised compulsory schooling from 5 to 8 years in Turkey to address the endogeneity of education to happiness. Our study shows that, for females, holding at least a middle school diploma increases the likelihood of being happy, and the probability of being satisfied with various life domains. Descriptive tests suggest that being hopeful about one’s own future wellbeing partly explain the relationship between women’s schooling and happiness. For males, although relatively imprecisely estimated, we find some evidence that holding at least a middle school degree causes a decline in subjective wellbeing. Supplemental analysis develops evidence consistent with the view that an imbalance between aspirations and attainments, induced by extended primary schooling, may be the reason behind this counterintuitive finding among men.
Work in Progress:
Curriculum Reforms, Adult Health, and Mortality (with Ozkan Eren, Kathryn Johnson, and Erdal Tekin)