| Abstract: |
Children and adolescents spend more than one-third of their time sleeping.
Yet, we know little about the causal impact of sleeping on their development.
This paper is the first to exploit variation in local daily daylight duration
measured on pre-determined diary dates across the same individuals through
time as an instrument in an individual fixed effects regression model to draw
causal estimates of sleep duration on a comprehensive set of child development
indicators. Applying this model to about 50 thousand time use diaries from two
cohorts of Australian children spanning over 16 years, we first document that
children sleep substantially less on days with longer daylight duration. Our
results show that sleeping longer improves selected general developmental,
behavioural and health outcomes in children and adolescents. By contrast,
sleeping more statistically significantly increases the BMI scores, mainly by
increasing the risk of being overweight. Moreover, while the impact of sleep
duration on general and behavioural outcomes is more pronounced for females or
older individuals, the effect on BMI is largely driven by males. The results
indicate a null or relatively small positive impact of sleeping longer on
cognitive skills. |