Women empowerment and infant mortality in Bangladesh

Abstract

This study attempts to investigate the effects of women empowerment on infant mortality in Bangladesh. Four indicators have been considered to measure four dimensions of women empowerment: the level of education, participation in household (HH) decisions, autonomy in movements and employment status. The rotated factor analysis technique is used to construct the last three dimensions. The Weibull parametric survival model has been specified and estimated using the Bangladesh Demographic and Health Survey (BDHS) data of 2011. The results show that three measures of empowerment (the level of education, participation in HH decisions and autonomy in movements) contribute significantly to the reduction of infant mortality. But women employment is associated with increased mortality for infants. The findings of this study have a number of policy implications on this issue for a developing country like Bangladesh.

Keywords:

• women empowerment
• infant mortality
• survival model
• rotated factor analysis
• Bangladesh

JEL Classification:

• I12
• O15
• I10

Notes

¹ The whole country was divided into clusters. The sampling technique consists of two-stage stratified sampling. In the first stage, 600 clusters (207 from urban and 393 from rural) were selected with probability proportional to the size of clusters. In the second stage, 30 HHs, on average, were selected from each selected cluster using the systematic sampling technique.

² Women empowerment has been specified in the literature in many ways. For example, Duflo (2011) highlighted three dimensions of women empowerment: education, participation in the decision making process and involvement in economic activities. On the other hand, Bhagowalia et al. (2012) used mobility, decision making power and attitudes towards verbal and physical abuse as indicators of women empowerment. Following the literature, the present study identifies and measures the dimensions of women empowerment based on principal component analysis.

³ Respondent’s level of income is more appropriate to be included in the set of covariates. But, unfortunately respondent’s level of income is not available in the dataset. Therefore, wealth indices are used to proxy the economic status of respondents.

⁴ The value of Wald statistics increased by more than 100 in the full model compared to that of the basic model.

⁵ In the survival model, the hazard ratio of the reference category is considered to be 1. Furthermore, it assumes that the hazard of a particular group with specific characteristics is proportional to that of the reference category.

⁶ In Bangladesh, there are six administrative divisions (Barisal, Rajshahi, Khulna, Sylhet, Chittagong and Dhaka). Five divisional dummies are used to represent six divisions with Barisal being the default division.