Sex-Selective Abortion in India: The Impact on Child Mortality B ...

Several studies in India have found that relative mortality of girls is lower among mothers with more education (Bourne & Walker, 1991; Rosenzweig & Schultz, 1982; Simmons, Smucker, Bernstein, & Jensen, 1982), and Murthi, Guio, & Dreze (1995) found at the district level that female literacy was associated with lower excess female child mortality, which did not vary by geographic (North/South) region. Male literacy was associated with greater excess female child mortality at the district level due to greater reductions in mortality for boys than for girls. Das Gupta (1987), on the other hand, found that in Punjab excess female mortality was higher among daughters of women with some education than among those of women with no education. These discrepant findings may result from the context-dependent social and behavioral effects of female literacy.

Physical and Social Security

Oldenburg (1992) studied the correlation between violence and population sex ratios in Uttar Pradesh, finding a correlation between high M:F sex ratios and murder rates at the district level. He hypothesized that variability in sex ratio by region can be accounted for by the level of violence and by the need for males in the family for day-to-day protection and security as well as by physical support in family and village affairs.

In a nationwide study, Dreze and Khera (2000) found that the population sex ratio was the strongest predictor of murder rates at the district level. Most murders were committed by young men, and the majority were disputes over property or over women. To sort out the direction of causality, a two-stage analysis was performed using female labor force participation as an instrumental variable for sex ratio.