The anomaly of women’s work and education in India

There has been a decline in the female labour force participation rate and size of the female labour force in India in recent years. This article looks at this trend in conjunction with female educational attainment in the country. It finds that states with high levels of patriarchy are also the ones where a high proportion of women with graduate and postgraduate degrees are out of the labour force.

There has been a fall in female labour force participation rate (FLPR) in India despite robust economic growth, rising incomes, fall in fertility rates, and improvement in female literacy levels. This precarious trend has received much attention in the last few years. Estimates from the National Sample Survey (NSS) data reveal that between 1993-94 and 2011-12, the FLPR (‘usual status’ approach¹has declined from 42.6% to 31.2% for ages 15 and above (Andres et al. 2017). In a recent study (Ghai 2018), I look at FLPR in conjunction with female education using the most recent data from the Labour Bureau’s Annual Employment and Unemployment Survey (EUS) (2015-16).

Drop in FLPR and concomitant decline in size of female labour force

Firstly, I find there has been a drop in FLPR from 31.1% in 2013-14 to 27.4% in 2015-16 (usual status approach). Moreover, not only has there been a fall in FLPR in recent years, the size of the total female labour force has also shrunk between 2013-14 and 2015-16, which is a subject of much concern. In 2013-14, the size of the female labour force was 136.25 million, which declined to about 124.38 million in 2015-16, a drop of 11.86 million² (the biggest decline observed since the decline between 2004-05 and 2009-10).

Table 1. Size of the female labour force (in millions); usual status approach

	FLPR (usual status)	Population projection	Size of the female labour force	Change in the size of the female labour force
Ages 15 and above
4^th EUS (2013-14)	31.1	438.11	136.25
5^th EUS (2015-16)	27.4	453.96	124.38	-11.86

Source: Author’s estimates from labour force participation rates and projected population in the 4^th and 5^th Report of the Labour Bureau Annual Employment and Unemployment Survey.

Examining the relationship between education and FLPR

Secondly, the U-shaped hypothesis (Klasen and Pieters 2013, Andres et al. 2017) that traces the relation between education and FLPR is reaffirmed: FLPR is high at very low levels of education, plummets into a downward trajectory with improvements in educational attainment, and rise again at higher levels of education. When education levels are low, women work out of economic necessity – mainly in agriculture. At medium levels of education, women are reluctant to work manual jobs at low wages. As the financial necessity of women to engage in outside work drops as well, families are keen for women to stay at home as it is reflective of a rise in social status. Women at higher education levels are able to find white-collared jobs, and are less restricted by social norms and family circumstances.

Conversely, by looking at the percentage of women out of the labour force at each level of educational qualification (2015-16), an inverse U-shaped curve is obtained: there is an increase in the percentage of women out of the labour force as one moves from illiterate (70.3%), to middle school- (79.3%) to higher secondary education (85.5%) levels and subsequently, a decline as one moves from diploma/certificate (70.9%) to graduate (68.4%) and finally to post-graduate levels and above (51.8%). This inverse U-shaped relationship holds true for women in both rural and urban areas as well as across all age cohorts (ages 15 and above, 18 to 29, ages 30 and above).

Thirdly, I re-examine the argument that increasing education enrolment has led to a drop in FLPR (Thomas 2012, Rangaranjan et al. 2011, Abraham 2013). Andres et al. (2017) note that between 1993-94 and 2011-12, for ages 15 and above the combined participation rate (sum of the FLPR and education participation rates) declined, suggesting that the decline in FLPR may not necessarily have stemmed from rise in educational attendance. Rustagi (2013) notes a secular decline in the FLPR between 1999-00 and 2011-12 across all age cohorts and not just the 15-29 year olds (the cohort most likely engaged in acquiring education). Ghose (2013) too argues that the non-student FLPR (the female labour force as a percentage of non-student population) displays the same pattern (decline) as that of the FLPR between 1999-00 and 2011-12.

All of the literature cited above has made use of the NSS quinquennial employment and unemployment survey, the latest data for which is available upto 2011-12. On examining the data from the Labour Bureau, it is also revealed that the drop in FLPR cannot just be attributed to higher educational participation among the young cohort but must depend on other heterogeneous factors. Between 2013-14 and 2015-16, there has been an increase in the percentage of women out of the labour force at all levels of education and across all age cohorts. For instance, with respect to the age cohort 30 and above, the percentage of women graduates out of the labour force has increased from 62.7% to 65.2% while the percentage of illiterate women out of the labour force has increased from 67.6% to 70.1%.

Fourthly, in understanding the phenomenon of drop in FLPR among women with high educational attainment, I identify four pathways such as the link between education and marriage markets, education and social norms, the poor demand conditions in the labour market for educated women, and quality of education. For instance, there is evidence that suggests that part of the expansion of education in India has been to improve the marriage prospects of women, rather than their employment prospects (Klasen and Pieters 2013). In 1987, primary education alone accounted for high returns in the marriage market but in 2009, the marriage prospects of women with higher education were better (Klasen and Pieters 2013). At the same time, adverse social norms for women place a higher premium and social prestige on households that keep the women secluded which offsets the impact of education on FLPR.

There is also the case of oversupply of educated women relative to growth in jobs that are considered appropriate by them (formal sector jobs), resulting in ‘crowding-out’ effects. Further, significant gender wage differential exists in the labour market even at high levels of education. The wage gap for women graduates and above vis-à-vis male counterparts was as high as 31.3% in rural areas and 24.3% in urban areas, in 2011-12. Further, education in most developing countries seeks to ‘domesticate’ rather than ‘empower’ women. The Hindu (2017) reported that there was absence of gender parity in textbooks for children in India for grades 2-5 on subjects such as Hindi, mathematics, English, and environmental studies. In all of these textbooks, women are confined indoors and shown to be adept only at domestic chores while men are shown to be participating in outdoor activities. While men are shown to be the head of the family, women are portrayed to be primary caregivers. Moreover, there was a significant ‘masculisation’ of jobs in these textbooks. For instance, in the class five mathematics textbook, all the jobs (in the illustrations) including that of a farmer, milk-seller, and shopkeeper are performed by men. Thus, education does not challenge the traditional elements that shape marriage dynamics in terms of decisions regarding labour force participation and domestic work. In September 2018, Madhya Pradesh’s Barkatullah University announced a three-month course to instil sanskar (traditional values) among the young generation. Vice-chancellor Prof. D.C. Gupta explained the objective of the course was to “make girls aware so they can adjust to new environment after marriage…and prepare brides who will keep families intact”. All of this speaks to the poor quality of education and institutionalised gender biases in the country.

Role of social norms

Lastly, I test the hypothesis that prevailing social norms hinder the participation of women in the economy despite higher levels of education. For the purpose of quantifying social and cultural norms, I constructed an index for patriarchy at the state level³. The variables used in the index to quantify patriarchy are child sex ratio, participation of women in household decisions, and spousal violence. Though solely correlational, notably it is found that states with high levels of patriarchy, as estimated by the index, are also states where high proportion of women with graduate and postgraduate degrees are out of the labour force. For instance, Haryana that scores very high on the patriarchy index has almost 70% of women graduates out of the labour force. On the other hand, Nagaland that scores low on the patriarchy index has only about 29% women with graduate degrees out of the labour force.

Concluding thoughts

Along with focus on female education, government schemes must also target the cultural and social forces that impede FLPR, and facilitate behavioural changes that are conducive to the acceptability of female employment. For instance, implementing reflective programmes on gender equality in secondary school that improve the cognitive ability of young girls to question gender roles, and launching high-octane campaigns at the local and national level that redefine norms around masculinity and men’s and women’s roles in the family. These, coupled with policies that address some of the other demand- and supply-side constraints on FLPR, are likely to bolster the effort and result in discernible positive outcomes for women.

Notes:

The usual status (principal status plus subsidiary status) approach takes into consideration both the major time criterion (183 days or more in an economic activity) and shorter time period (30 days or more in an economic activity). Thus, a person who has worked even for 30 days or more in any economic activity during the reference period of last 12 months is considered as employed under this approach.
The Labour Bureau’s 4^th and 5^th EUS report the population projections for ages 15 and above using the Census 2001 and 2011 data. The reference period for the 4^th EUS is January 2014 to July 2014 and the population data is as on 1 March 2014. It is estimated using the formula A=A1*{[1+(R/100)]^(52/120)}, where ‘A1’ is the census population as on 1 March 2011, ‘R’ is the percentage decadal change in population between 2001 and 2011, and ‘A’ is the projected population as on 1 March Similarly, the reference period for the 5^th EUS is April 2015 to December 2015 and the population data is as on 1 July 2015 using the same formula.
The index is constructed using the technique of principal component analysis, which simplifies the complexity in high-dimensional data while retaining trends and patterns. It does this by transforming the data into fewer dimensions or variables, which act as summaries of its features.