Excess Female Infant Mortality And The Gender Gap In Infant Care In Bihar, India

ABSTRACT

This paper analyzes excess female infant mortality and the gender gap in infant care in the East-Central state of Bihar, India. Using pooled data from three waves of the National Family Health Surveys (1992–3; 1998–9; 2005–6) in a panel analysis, it compares female infant mortality rates in Bihar to those in thirteen major Indian states. These comparisons suggest that females in Bihar experience a statistically significant excess infant mortality when compared to female infants in less gender-biased states in eastern, western, and southern regions, but not when compared to more gender-biased states in northern and central regions. An estimated 23 percent of female infant deaths in Bihar are excess. Examination of infant care demonstrates that a gender gap in Bacillus Calmette–Guérin (BCG) vaccination is the single most important driver of excess female infant mortality, followed by gaps in vaccination for polio; diphtheria, pertussis, tuberculosis (DPT); and measles.

KEYWORDS:

• Excess female infant mortality
• India
• Bihar
• gender gap in infant care
• vaccination

JEL Codes:

• J16
• O53
• J13

NOTES ON CONTRIBUTOR

Sanjukta Chaudhuri is Associate Professor of Economics and a women's studies program affiliate at the University of Wisconsin–Eau Claire. Chaudhuri specializes in feminist economics, women's empowerment, human rights, and gender issues of South Asia. She has published articles in World Development, Women and Health, International Perspectives on Sexual and Reproductive Health, Healthcare for Women International, Review of Radical Political Economics, International Journal of Economic Policy in Emerging Economies, and International Journal of Public Policy. In 2012, she received a research publication grant from the American Association of University Women (AAUW). She is a 2014 honoree for the University of Wisconsin system outstanding women of color in education.

Notes

¹ In this paper, infant mortality rate refers to the number of infant deaths between 0–12 months of age for every 1,000 live births.

² Throughout the paper, Bihar is compared with thirteen major states of India including West Bengal and Orissa (East), Uttar Pradesh and Madhya Pradesh (Central), Punjab, Haryana, and Rajasthan (North), Maharashtra and Gujarat (West), Andhra Pradesh, Karnataka, Kerala, and Tamil Nadu (South). These thirteen states together with Bihar are referred to as “national.”

³ In 2001, Bihar was bifurcated into two states: Bihar and a new state named Jharkhand, which was carved out of southern Bihar. All analyses in this paper refer to an undivided Bihar. Since Jharkhand was included as a separate state in the NFHS-3, all observations on Jharkhand from this wave were recoded as Bihar.

⁴ Supplemental Table 1 is available in the supplemental tab on the publisher's website.

⁵ Supplemental Table 2 is available in the supplemental tab on the publisher's website.

⁶ The SRS is administered by the Office of Registrar General of India.

⁷ Ancient Hindu religion stratified society into five distinct groups or castes. The caste system made caste hereditary, something one was born into and which could never be changed. Caste affected every aspect of life, determining a person's occupation, education, marriage, privilege and status in relation to other castes. The lowest caste was shudra, consisting of people who were socially shunned from birth to death and who performed the most menial jobs. The shudras were considered so loathsome that they were referred to as achhuts or untouchables. In 1949, the Indian constitution adopted a permanent affirmative action program that privileged the shudras in matters of education, employment, government, and legislative power. The list of groups that were to be beneficiaries of this affirmative action came to be called the scheduled caste (SC) and scheduled tribes (ST). Given that a quarter of India's population belongs to SC and ST, this variable is an important control variable included in the analysis (Ashwini Deshpande 2001).

⁸ Controls used in the analysis are similar to Clark (2000), Deepankar Basu and Robert De Jong (2010), Jayachandran and Kuziemko (2011), and Oster (2009a).

⁹ Predicted probability of female IMR in Bihar is estimated as follows: Expected Pr (mortality) = . Note that Tables 3–5 report odds ratios only. The coefficient βs are not reported but are available from the author upon request.

¹⁰ See Table 9 notes for sources of data.