Research

Research Papers

Sleep and Adolescents’ Mental Health and Educational Attainment (Job Market Paper)

Studies consistently show that short sleep duration is associated with mental health problems and poorer cognitive performance in adolescents. Yet relatively few studies have attempted to establish the extent to which these correlations represent causal relationships. This is especially important considering that adolescents are sleeping less and experiencing poorer mental health – early school start times and the increasing temptation of technology, such as social media and streaming services, make it difficult for adolescents to obtain sufficient sleep. This paper uses data on US middle and high school students and exogenous variation in sleep duration induced by sunset time to investigate the impact of sleep on adolescents’ mental health and long-run educational outcomes. A later sunset will induce a later bedtime, but due to sticky school start times, adolescents will not wake up later. Suggestive evidence indicates that sleep deprivation increases depressive symptoms and decreases the probability of graduating from high school and attending college among US teenagers. These results indicate that sleep deficits can have a lasting impact on human capital formation.

Public and Private Information on Nursing Home Care Risk (with Owen O’Donnell & Teresa Bago d’Uva)

Inaccurate expectations of nursing home care needs can distort willingness to pay for long-term care insurance and cause financial hardship in old age. Using the Health and Retirement Study, we measure the accuracy of older Americans’ subjective probabilities of moving to a nursing home within five years and identify the contributions of public and private information to that accuracy. We find that while the subjective probabilities have predictive power and are reasonably close to the objective risk on average, individuals, particularly the least wealthy, make noisy predictions that are highly inaccurate. In part, the inaccuracy stems from inappropriate weighting of risk factors that insurers can observe and use in pricing contracts. Subjective probabilities capture only 30% of the potential discriminatory power of this public information. Nevertheless, they are predictive of long-term care insurance coverage, even when controlling for public information. This indicates that long-term care insurance may be afflicted not only, or even mainly, by adverse selection arising from private information but also by behavioural selection – suboptimal insurance decisions resulting from misuse of public information.

Gender and Mortality in Russia: A Decade of Lifestyle Changes (with Teresa Bago d’Uva)

Virtually all around the world, women are less likely to die than their male counterparts. This gender gap in mortality is especially striking in Russia, where the gender difference in life expectancy exceeds 10 years. This paper examines the changing contribution of lifestyles to the gender disparity in 5-year mortality in Russia by decomposing the gender mortality gap for two relevant periods a decade apart: 2000-2003 and 2010-2013. Our results indicate that gender differences in lifestyles explain a substantial part of the gender mortality gap in the early 2000s (47%) and the early 2010s (29%). Most of this lifestyle contribution is due to gender differences in smoking behaviour, with alcohol consumption contributing relatively little. Over time, the gender gap in mortality has shrunken considerably. Along with it, there has been a noticeable decline in gender differences in lifestyles. The results point to men’s declining tobacco consumption as the main lifestyle factor responsible for narrowing the gender gap, but men’s decrease in excessive alcohol use has also contributed, albeit to a much lesser extent.

Research in Progress

Teachers’ Beliefs and Students’ Human Capital: The 2015 CITO Reform (with Vahid Moghani)

In the Netherlands, students in their final year of primary school are tracked by ability for secondary education. Before 2015, this tracking was mainly determined by the result of a, mostly central, test taken in the final year of primary school, the CITO test. However, since 2015, this test result is no longer leading, but the recommendation of the teacher, determined before the student takes the test, mainly determines the track. This research investigates the consequences of this reform by looking at teachers’ behaviour and students’ mental health and educational attainment following the reform.