Could better jobs for men have improved gender equality?

Evidence indicates that economic growth can improve gender equality. In this post, Sujata Balasubramanian suggests that India’s high-growth period from 1982-83 to 2011-12 failed to do so substantially. She examines structural changes over those three decades, concluding that the failure was due to insufficient employment – not just for women, but also for poorer men. The analysis therefore emphasises the importance of employment for both pro-poor growth and gender equality.

Economic theory suggests two ways in which growth may increase gender equality: directly, by raising female employment, and indirectly, by reducing poverty (Kabeer 2016). When economic growth lowers poverty, this reduces the need for poorer families to discriminate against females when allocating essentials like food and healthcare (Duflo 2012). In a recent study (Balasubramanian 2023), I focus on this latter scenario, using a case-study method to assess whether India’s high growth, between 1982-83 and 2011-12, was likely to have reduced gender inequality simply by lowering poverty.

Theoretically, therefore, my study builds on two, previously unconnected strands of the literature on economic growth: an older strand about pro-poor growth (Ravallion and Chen 2003), and a more recent strand relating growth to gender equality (Kabeer and Natali 2013). My analysis links these two strands to emphasise that, if poverty reduction converts growth into gender equality, then the factors that create pro-poor growth should also enhance gender equality.

As a first step, I rule out any direct impact of growth via higher female employment. The National Sample Survey (NSS) data show that higher growth actually correlated with a fall in women’s employment from 30% in 1983 to 22% in 2011-12. In the prime working-age group of 25 to 59 years, female employment fell from 34% in 1983-84 to 28% in 2011-12 (Lahoti and Swaminathan 2016). Consequently, female employment could not have improved gender equality.

This leaves poverty reduction as the only factor potentially converting growth into gender equality, and my study illustrates that:

(i) Patterns of structural change between 1982-83 and 2011-12 left India overdependent on agricultural employment. Growth in manufacturing and services was capital-intensive and skill-intensive, and unable to create sufficient employment even for poorer, less-educated men.

(ii) Additionally, growth was not pro-poor, because of low investment in basic education. Accordingly, high growth failed to produce commensurate poverty reduction.

Since there was limited poverty reduction, I hypothesise that growth was also unlikely to have substantially improved gender equality: a brief review of progress on equality supports this conclusion.

Structural changes and employment: 1982-83 to 2011-12

Table 1 illustrates that agricultural employment is disproportionately high relative to the sector’s share of GDP (gross domestic product): agriculture employed 69% of workers in 1982-83, and 49% even in 2011-12 – despite contributing only 18% of GDP. Substantial income gains could have resulted had more workers shifted from lower-productivity agriculture to manufacturing and services.

Table 1. Structural composition of employment: 1982-83 to 2011-12

	1982-83		2011-12
Sector	Share of GDP	Share of employment	Share of GDP	Share of employment
Agricultural sector	35.2%	68.5%	17.9%	48.9%
Manufacturing	17.6%	10.6%	14.7%	12.6%
Services	37.5%	17.5%	54.9%	26.9%

Note: GDP figures are based on factor cost and current prices.

Source: Author’s calculations based on data from various sources.

Unfortunately, manufacturing failed to create jobs on the scale seen in other industrialising, Asian countries. Manufacturing employment rose extremely slowly, from 10.6% in 1982-83 to a peak of just 12.6% in 2011-12. By contrast, around Taiwan’s industrial peak, manufacturing provided 34% of employment. In Malaysia, manufacturing generated 26% of employment in 1995 and 23% in 2005. Even in Sri Lanka, manufacturing provided 16% of employment in 2005 (Athukorala and Sen 2015).

Indian services also generated remarkably low employment. In both developing and developed economies, the contribution of services to employment is estimated to be roughly equal to their share of output. However, Indian employment was just over half this predicted level: services provided only 28.9% of employment in 2010, although they should have contributed 53.8% based on their contribution to GDP (Ghose 2014).

Relatedly, the skill-intensity of both services and manufacturing held back poorer workers. Notably, only about 14% of agricultural workers have a secondary or higher level of education. But the percentage of employees with a secondary or higher education rises to 43.2% in registered manufacturing, and sharply, to 84% in financial services and insurance, and 89% in educational services (Amirapu and Subramanian 2015). Thus, while both manufacturing and services generated low employment, the jobs typically required more educated workers. In the prime working age group of 25 to 54 years, however, only 57% of males have completed secondary school (World Economic Forum, 2017). So, 43% of men may be unqualified for many of the higher-productivity service and manufacturing jobs.

In short, growth did not create sufficient remunerative employment even for poor, less-educated male workers – quite apart from not raising female employment. Moreover, while investments in primary and secondary education can prepare poorer workers for manufacturing and service jobs, government investment during that period tended to prioritise tertiary education.

Higher education was subsidised substantially more than primary schooling, even when compared to countries that are better-off in terms of per-capita GDP: the Indian ratio of public expenditure, per student in tertiary education to that in primary education, in the year 2006, was 6.2, in comparison to a ratio of 2.1 in Indonesia in 2007, and 1.6 in Thailand in 2004 (Ghose 2014). Meanwhile, the Annual Survey of Education Report (2024) reveals that the quality of primary schooling in rural areas remains abysmally poor. Inevitably, this failure to ensure access to high-quality primary and secondary education has compounded India’s inability to create jobs for the poor.

Growth was not pro-poor and increased inequality

Not surprisingly, the growth elasticity of poverty – that is, the percentage reduction in poverty caused by a 1% increase in per capita incomes – was extremely low during these high-growth decades. In a comparison across all developing countries, from 1981 to 2005, India’s growth elasticity was among the weakest. For example, from 1981-1990, for both the US$1 and US$1.25 poverty lines, the average growth elasticity of poverty for developing countries was more than seven times that of India. Indian elasticities for the US$2 and US$2.50 poverty lines were nearly insignificant from 1981 to 2005 (Lenagala and Ram 2010).

Relatedly, income inequality increased from the 1980s, paralleling India’s high-growth period. Chancel and Piketty (2017) show that between 1980 and 2015, the share of total income growth accruing to the bottom 50% of Indians was only 10-11%, while the top 1% captured 28-29% of the gains. By contrast, despite slower growth during 1951-1980, the bottom 50% captured 28% of the gains, with their incomes growing by 87%, compared to 5% income growth for the top 1%.

Progress on gender equality

Since the high-growth period had a very limited impact on poverty, I infer that it could not have substantially improved gender equality either, by reducing poverty-related gender discrimination. Lastly, therefore, my study assesses progress on gender equality from 1991, using the female to male sex ratio as an indicator. This ratio reflects excess female mortality, and is likely to be affected by growth, and, more importantly, by poverty reduction.

The Indian sex ratio in 1991 was substantially lower than ratios in Europe, North America and Sub-Saharan Africa. Sen (1992) estimated that 37 million Indian women were “missing”, due to both sex-selective abortion and other, scarcity-driven discrimination (like females receiving less food and healthcare in poor families). While sex-selective abortion may actually rise with growth (due to increased access to and affordability of pre-natal sex determination techniques), scarcity-driven gender discrimination can be expected to decline – if growth reduces poverty. And notably, sex-selective abortion accounts for only 30% of missing Indian females. Approximately 70% of excess female mortality is due to other causes – primarily poverty-related discrimination (World Bank, 2011).

The Indian sex ratio did improve between 1991 and 2011, indicating some reduction in both sex-selective abortion and other gender discrimination. However, “The stock of missing women as of 2014 was nearly 63 million and more than 2 million women go missing across age groups every year (either due to sex selective abortion, disease, neglect or inadequate nutrition)” (Government of India, 2017).

Thus, high growth was unable to eliminate excess female mortality, with the total number of missing women rising from 37 to 63 million between 1991 and 2014. Conversely, if growth had generated substantial poverty reduction, this could have increased the access of poor girls and women to essentials like food and healthcare, thereby eliminating, or at least sharply reducing, the roughly 70% of excess mortality caused by poverty-driven discrimination.

Furthermore, for girls under five, the trend of excess mortality worsened between 1990 and 2012. Consequently, by 2012, India had by far the highest excess mortality in the world among girls under five, at 13.5 per 1,000 livebirths, followed by Afghanistan at 5.2 per 1,000 and Pakistan at 4.7 per 1,000 (Alkema et al. 2014). These statistics are in line with my inference that growth did not substantially improve gender equality through poverty reduction.

It must be emphasised that excess female mortality is the most extreme indicator of poverty-driven gender discrimination, and consistently underestimates its full extent. Firstly, excess mortality does not capture some forms of discrimination – like lower spending on girls’ education. Secondly, it does not even capture the fullest extent of discrimination in nutrition and healthcare – since, for example, giving girls less food than boys results in higher stunting/wasting but not necessarily death. Therefore, excess female mortality systematically underestimates poverty-driven discrimination.

Policy implications

From a policy perspective, my study, firstly, illustrates how structural changes can alter the impact of growth on gender equality via employment patterns. Second and relatedly, it underlines the importance of pro-poor growth for gender equality. Finally, it highlights India’s failings on two essential dimensions of pro-poor growth – job creation and basic education – which limited the power of growth to reduce both poverty and gender inequality. While female employment actually fell, growth did not generate sufficient remunerative employment even for poor males. Conversely, and counterintuitively, better-paid jobs for poor men could have reduced poverty-driven discrimination, thereby improving gender equality.