Boys' versus Girls' Schooling in Nepal: Does It Vary by the Extent of Mothers' Autonomy?

Abstract

This paper hypothesises that resource allocation affecting the decisions relating to sons' versus daughters' schooling in Nepalese households is dependent on the extent of the mother's autonomy. Here, we posit that women's autonomy is a relative concept as a woman has degrees of decision-making power within her household. The results indicate that daughters' education is more likely to benefit when mothers solely make the decisions, but when decisions are made jointly with her spouse then the decisions are more likely to be in favour of sons' education. Our results indicate a marked gender difference in parental decisions over children's education, in the direction posited above, and less than 10% of mothers in the sample have complete autonomy over such decision-making. These results are important for policy-makers wishing to decrease gender bias in children's educational outcomes.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ Discussed in detail in a later section.

² Centers for Disease Control and Prevention, Reproductive birth outcomes. Sex ratio and the environment. Retrieved from http://ephtracking.cdc.gov/showRbSrEnv.action (2013).

³ All data are self-reported and therefore suffer the biases associated with self-reported data. The argument against self-reported measures is that these measures are based on very specific questions, which by their nature constrain the likelihood that respondents rationalise their own behaviour through their answers. The problems relating to using self-reported data range from the individual's tendency to justify his/her own behaviour via the response to the unknown level of comparability of these evaluations across individuals (Currie & Madrian, 1999). However, any analyses based on household surveys suffer from such limitations since there is no other way of gathering information about household dynamics without learning about it from the respondents. The reader thus needs to interpret the results from this analysis with caution given it is based on self-reported data.

⁴ For a comprehensive study and discussion of Nepal's education system, refer to Graner (1998, 2006), Bhatta, Adhikari, Thada, and Rai (2008) and Bhatta (2004, 2005). For ethnographic research addressing education, particularly gender dynamics, refer to Rothchild (2006) and Ahearn (2001).

⁵ The mother's autonomy being referred to applies only in terms of decision-making as far as the children's education is concerned.

⁶ There are papers which also make counter arguments (Gupta, 1987).

⁷ The reason for including families with at least one son and one daughter is to be able to see whether or not there is any preferential treatment shown towards sons versus daughters.

⁸ Estimations were also carried out using the proportion of school-age boys and girls who have been withdrawn from school, and the proportion of illiterate boys and girls in the household as dependent variables. The results relating to these other dependent variables were almost identical in terms of signs and statistical (in)significance, and thus these results are not being presented here. For example, full autonomy of mothers reduces the proportion of girls withdrawn from school but this impact is statistically insignificant when mothers have partial autonomy. When mothers have partial autonomy, it is found to have a statistically significant negative impact only on the proportion of sons withdrawn from school. In terms of proportion of illiterate sons and daughters, mothers having full autonomy has a statistically significant negative impact on the proportion of both illiterate sons and daughters but this impact on daughters becomes insignificant when the decision is made jointly with the husband.

⁹ About 4% of sons in this sample were withdrawn after being sent to school and about 7% of daughters were withdrawn.

¹⁰ The literature on female autonomy commonly finds factors such as outside work opportunities (Anderson & Eswaran, 2009; Rahman & Rao, 2004), infrastructure at the village level (Rahman & Rao, 2004), access to resources (Agarwal, 1994), age (Mason & Smith, 2003), religion (Morgan, Stash, Smith, & Mason, 2002), and caste (Jejeebhoy & Sathar, 2001) to have important influence over the level of female autonomy in the household.

¹¹ Since the analysis carried out in this paper is at the household and not the individual level, it does not take into account the age and/or sibling position of individual boys or girls which have been shown to influence household decision-making (Punch, 2001).

¹² Another potential variable to measure the opportunity cost of schooling children would have been to see whether children are sent to private or public school and how this varies by the extent of mother's autonomy. However, we do not use this criterion since this would probably provide more meaningful results if the analysis was being carried out at the individual level.

¹³ For instance, Jacoby and Skoufias (1997) find that child labour helps smooth the incomes of rural Indian families, given the poorly developed credit and risk markets common in these instances.

¹⁴ What this implies is that simply increasing female autonomy may not have the intended consequences for girls' versus boys' education unless the women who would be making the decisions due to increased autonomy have a more feminist attitude that would actively promote girls' education.