While dowry is regularly adduced as a motivation for son-preference in India, there is no available evidence on how son preferring behaviors respond to changes in the cost of providing dowry. Dowry is not widely measured and, where it is, the size of dowry payments is likely to be endogenous, varying with family preferences in micro-data and with population sex ratios in time series data. This paper exploits exogenous variation in the burden of dowry created by a sharp rise in the price of gold in 1980 of unprecedented magnitude. We use a regression discontinuity design on birth month, investigating whether girls born in the months before the price shock are more likely to survive the neonatal (first month of life) and infant (first year) periods compared with girls born in the months following the price shock, with boys acting as a control group. On the premise that Muslim households are much less likely to give dowry, we estimate a difference in discontinuities to investigate birth (month) cohort differences in girl relative to boy outcomes in Hindu relative to Muslims families. We find an increase in excess girl mortality in Hindu and not Muslim households. Hindu Girls born just after the gold price hike were between 6 and 9 percentage points more likely than boys to die before the age of one. The increase in excess girl mortality is larger in Hindu households in which the first born child is a girl consistent with such families having a greater incentive to exercise son preference in the allocation of early life nutrition and care. Our findings indicate son preference and they undermine the alternative hypothesis that the birth of a girl, on account of dowry, acts as a negative wealth shock that hurts boys and girls equally.
Presented by:
Selim Gulesci (Bocconi University)
Date & time:
December 1, 2014 4:00 pm - December 1, 2014 5:30 pm
External seminars home