Search

• We utilized unique historical datasets on the per capita consumption of tobacco and the heights of twenty-year old conscripts in the second half of the 19th century Italy. • Linear models for longitudinal data for Italian provinces are used. • We find a strong negative effect of smoking in the period immediately before and after birth on the height at age 20. This paper studies the effect of passive smoking on child development. We use data from a time when the adverse effects of smoking on health were not known and when tobacco was not an inferior good. This allows us to disentangle the effect on foetuses and infants of smoking from that of other indicators of social and economic conditions. We exploit a set of unique longitudinal historical datasets defined at a detailed level of geographical disaggregation, namely the 69 Italian provinces. The datasets record precise information on the per capita consumption of tobacco products, the heights of twenty-year old conscripts in the second half of the 19th century Italy, and other relevant controls. We find a strong negative effect of smoking in the period immediately before and after birth on the height at age 20. Results are robust to changes in specification and consistent across the height distribution. [ABSTRACT FROM AUTHOR]