Reflections on the Less Visible and Less Measured: Gender and COVID-19 in India

Abstract

The gender effects of COVID-19 are complex, and extend much beyond the issues of care work and domestic violence that have captured global attention. Some effects have been immediate, such as job losses, food shortages, and enhanced domestic work burdens; others will emerge in time, such as the depletion of savings and assets and pandemic-related widowhood, which would make recovery difficult. I use examples from India to outline the complexity of such outcomes, the limitations of the many telephone surveys conducted during the pandemic, and the importance of anticipating both the immediate and the sequential effects.

We can anticipate these effects by drawing on our knowledge of preexisting gender inequalities and people’s coping strategies under crises, as well as real-time media alerts. Prior conceptualization can help us design better surveys for capturing both the visible and less visible impact of the pandemic, as well as formulate more effective policies for mitigating the adverse effects. I also highlight the advantages of group-based approaches for protecting women’s livelihoods during such crises, and emphasize the need to create a synergy between feminist theory, evidence gathering, and policy formulation.

The global focus on the gendered impact of COVID-19 has been largely on two issues: domestic violence and care work. Both issues are clearly important, but they do not cover the full range of gendered fallouts and their implications. Beyond the visible and short-term impact of the pandemic lie longer-term implications for the many who are unable to recover economically, socially, or emotionally—implications that are likely to play out differently for women and men and across diverse socioeconomic groups. In fact, even the much discussed rise in domestic violence is complex, catalyzed by multiple factors—some hidden, such as social norms and intra-household power relations, and others visible, such as women’s property ownership and job losses under the pandemic. Using examples from India, I share reflections on the complexity of outcomes we might expect, the limitations of existing data and evidence, and the imperatives of tracking and monitoring shifts over time to capture both the immediate and the sequential effects.

The Complexity of Expected Gender Effects

Consider first the effects on jobs and livelihoods. It is notable that although employment has been at the center of global media attention, we hear little about differential job loss by gender and even less about the intersections of gender, caste, and race. In India, as in many developing countries, most workers are employed in the informal economy, women more than men: 90 percent of female workers and 86 percent of male workers in India, for instance, are in the informal sector (Government of India 2019). Moreover, women are concentrated largely at the lower earning end of both rural and urban labor markets. In rural India, 75 percent of women workers are in agriculture, relative to 55 percent of male workers (Government of India 2019).

Women are at particular risk of losing their livelihoods during the pandemic because they have fewer employment options, are in lower-paid less-secure jobs, or in jobs that are “touch heavy” such as domestic help, or are doing unpaid work in family enterprises. Also, once displaced, they are less able to recover. Women who have gone back from their city jobs to their villages (as have many domestic workers) are less likely to return or be hired back. Those seeking to survive with no earnings and few savings will increasingly face poverty and food scarcity and may even lose their tools of trade, such as the carts on which street vendors depend.

In fact, there is usually a sequence of responses over time, as indicated by research on the coping strategies of rural communities in South Asia during climatic disasters such as severe droughts. Rural households facing food shortages, for example, first forage for food in the local commons, then draw on stocks of grain and other items, and then reduce the number of meals they eat, turn to neighbors for support, and borrow where they can, before mortgaging or selling assets (Agarwal 1990; Jodha 1978). If forced to sell assets, families tend to first dispose of smaller items, such as small animals or pieces of jewelry, while holding onto cattle and land to the last (Agarwal 1990; Jodha 1978). The irony is that the smaller assets are also often the only assets women own, whereas the larger assets are usually owned by men. Only about 14 percent of rural women, for example, own farmland in India (Agarwal, Anthwal, and Mahesh 2020). Selling small items appears logical in economic terms, since farming communities need to hold on to land and cattle for production, but this sequence of sales can have hidden gender costs. If women lose their assets while men retain theirs, women’s bargaining power within the home is undermined, leaving them more vulnerable (Agarwal 1997).

Extrapolating from this understanding, we would expect that under the pandemic the jobless would draw first on savings and cut down on essential consumption, then borrow from friends, relatives, and moneylenders, and finally begin to sell assets (Agarwal 1990). We would also expect these consequences to vary by class. Well-endowed households with ample savings and assets could remain secure over the crisis. Households with fewer savings and assets could lose them over time, depending on how long it takes to restore livelihoods. But poor households, with no savings or assets to begin with, could fall into extreme poverty, even destitution, and not recover at all. These sequential costs are likely to have unequal impact by gender because of preexisting economic and social inequalities within families, including gendered social norms that give men embedded advantages. And the inequalities can worsen with the pandemic.

As food scarcity grows, for example, in regions where existing social norms require women and girls to eat last and least, as in northern India, we can anticipate gender inequalities in food sharing and hunger within homes. It is worth considering, however, whether this would be the case only in developing countries or be more widespread globally? In the United States, for example, many families are facing food scarcity under the pandemic. We might ask: Is the impact of this scarcity being borne equally by gender? Similarly, when households draw on savings, whose savings are being drawn upon, men’s or women’s?

Even for domestic violence or care work, we can encounter layers of complexity. For example, although it is well recognized that clustering family members into small spaces under lockdowns could increase the risk of physical and verbal abuse, it is less recognized that male unemployment can also escalate domestic violence, as found in India (Agarwal and Panda 2007). Thus, women can be affected not only by their own loss of earnings but also if their husbands lose their jobs under COVID-19. Moreover, domestic violence can vary by class and asset ownership. In India, women owning immovable property, such as land or a house, are found to be at dramatically lower risk of domestic violence than propertyless women (Agarwal and Panda 2007). A woman who owns a house or land has a credible exit option, or, if it is her house the couple is living in, she can ask her husband to leave. Both factors can deter violence, quite apart from the lure that property itself may hold for the husband. Moreover, women who are economically better off are also better placed to connect with helplines through their cell phones if they face violence during the pandemic, whereas poor women lacking individual cell phones would remain at high risk. A 2019 study of 2,000 rural and urban adults in India found that 63 percent women compared with 79 percent men owned mobile phones, and only 21 percent of women relative to 42 percent men had mobile Internet (Global Systems for Mobile Communications 2020). Hence, to assess the impact of the pandemic on domestic violence, we need to move beyond generalizations to a more intersectional approach that takes into account women’s class and property status.

Similarly, both globally and in India, there is a substantial body of work on the excessive burden of care work borne by women, even before COVID-19 hit us (see International Labour Organization 2018 for an overview). But in discussing the pandemic’s impact on women’s work burden we must differentiate between middle-class urban families and poor rural ones. In India, by the recent 2019 All-India Time Use Survey, females and males were found to spend on average 5 hours and 1.6 hours per day, respectively, on unpaid domestic work (rural–urban differences were narrow) (Government of India 2020a). Under COVID lockdown, as out-of-school children and jobless husbands stayed home, women’s domestic workloads would be expected to increase, but not uniformly across households. Among urban professional couples who have to do without domestic help during the pandemic, the extended domestic work may well be shared. In rural areas, however, what counts as domestic work can involve additional tasks, such as fetching water, gathering firewood (this remains the single most important source of fuel in South Asia and sub-Saharan Africa; see Agarwal 2010), collecting fodder (animal care is largely women’s work), and foraging for supplementary food items from local forests and commons (Agarwal 2010). This could stretch women’s time during the pandemic in ways that are seldom anticipated. Caring for those in home quarantine under COVID is another less-recognized aspect of women’s increased burden of work.

Other types of effects that tend to remain unseen include the impact of a man’s death under COVID on the widow left behind. We know that COVID death rates have been higher among men than women (Global Health 50/50 2020), but there appears to have been no discussion on the economic and social disabilities women might face as widows.

There are also likely to be second-generation gender and class effects, both immediate and long-term. Because of the pandemic, schools have closed globally, often requiring students to learn online. But large numbers of households in developing countries lack computers or mobile phones with Internet access. In families that have one phone, given gender bias we would expect priority to be given to the education of boys over girls. Girls may even drop out of school altogether. Again, although this is more likely in developing countries, it could also be happening in poor communities of the Global North.

Surveys: What They Reveal and What Remains Concealed

To what extent are these complex effects being captured in surveys on the pandemic? The answer is, rather few. In India, for instance, from late March 2020 many organizations began conducting telephone surveys, a few with large samples and covering many states but most with small samples and usually focused on one region or occupation. Of 23 such surveys that I tracked, only 11 had covered any aspect of gender, and most had covered only one dimension (Agarwal 2021). Some surveys focused only on women, with no comparative data on men. And most were limited to one point in time. In fact, it was media reports—the “gray literature”—that provided more nuanced leads on women’s experience of the pandemic. But few, if any, of these leads were followed up in the large-scale surveys.

Telephone surveys also carry a built-in gender bias, since women respondents often lack personal mobile phones or privacy (Alvi et al. 2020). With family members listening in, we are likely to get skewed reporting of domestic violence and even of unequal sharing of workloads or food. With these caveats, what can we learn about the impact of COVID-19 on Indian women from existing evidence? The issues that have been best covered by the systematic surveys are the immediate effects on male and female employment and women-directed government relief schemes; the issues most neglected are the intra-household effects.

On employment, analysis of data from the Center for Monitoring the Indian Economy—the only source of all-India longitudinal data on employment under the pandemic covering several thousand households—showed that in April 2020, soon after India’s stringent lockdown on March 25, 2020, although more men than women lost jobs in absolute terms, women were 20 percentage points less likely to be employed among those employed before the pandemic. And by August 2020, although both had recovered partially, women were behind men (Deshpande 2020). For neither period, however, do the data cover the days of work or level of earnings. Hence, we cannot assess whether those employed are earning enough for subsistence or remain under economic distress with inadequate incomes.

Smaller surveys found, as anticipated, that with job loss and livelihood erosion, families drew first on savings and then borrowed (Civil Society Organizations [CSO] 2020; SEWA Bharat 2020). Unfortunately, this information is not gender-disaggregated in the larger CSO survey and is available only for women in the smaller SEWA Bharat survey. The latter covered members of the Self-Employed Women’s Association (SEWA) ¹ and found that 91 percent of the 300 women interviewed had run out of savings and borrowed from relatives, friends, and moneylenders. However, we do not know whether the women had used their own collateral or their family’s collateral for borrowing, or whether the borrowing was for individual needs or household needs, and who would pay off the debt.

We also do not know to what extent families faced the next sequential consequence, namely, being forced to sell assets, beyond the questions I requested to be inserted in some ongoing surveys. These revealed that in the immediate months following the pandemic, none recorded asset sales beyond, say, a mobile phone, if the family owned more than one. But female street vendors did say that if they needed to sell assets as time progressed, they would first sell their unsold stock and animals before considering carts or rickshaws (SEWA Bharat 2020). These surveys were not repeated, but clearly a follow-up on asset disposal is important since that could indicate whether the women are economically stable, have recovered, or are sliding into poverty.

What is clear, however, is that the most immediate and serious impact of the pandemic was on access to food, not just in India but also in developed countries, as can be seen from the long lines leading to food banks in American cities. In India, although several one-point surveys and media reports (e.g., Scroll 2020) focused on the growing food insecurity and hunger under the lockdown, none provided gendered insights. Even the detailed CSO (2020) survey for rural India mentioned above—which notes a reduction in the number of daily meals and number of items eaten per meal during the pandemic—does not tell us the relative sacrifice made by male and female family members. At best we can infer gender differences from earlier studies of intra-household inequalities in food sharing (Agarwal 1994). Similarly, some qualitative surveys and media reports highlighted the inability of poor pregnant and lactating women to get enough nutrition (National Law School of India University 2020), or to reach hospitals for delivery because of transport restrictions during the lockdown (Bisht, Sarma, and Saharia 2020), but there was no systematic tracking of such cases.

Government relief packages of cash and kind could mitigate the impact on hunger to an extent, and here several surveys did ask if the relief had reached eligible households, including the Rs. 500 ($6–7) per month promised specifically to poor women under the Pradhan Mantri Jan-Dhan Yojna (PMJDY). The surveys consistently found that the majority of women had been left out: For example, one survey of 2,670 respondents undertaken during mid-April–May 2020 across 10 states and 2 cities found that 64 percent of the eligible women did not get their cash relief because they did not have the special PMJDY bank accounts needed to receive this transfer (CSE 2020; see also Pande et al. 2020; Dvara Research 2020).

Information on other types of effects is sparse, however, including the impact on women’s domestic workloads in India. The above-mentioned CSO (2020) survey of a few thousand rural households across 11 states in June 2020 was unusual in this respect. In 53 percent of the 820 households with male migrant returnees on which it had information, women were found spending more time fetching water and 71 percent reported an increase in women’s firewood collection time. The comparable figures for nonmigrant households were 39 and 52 percent, respectively. Among urban middle-class couples, men who had shared some domestic tasks with their wives under lockdown usually reverted to pre-COVID levels after they returned to work (Deshpande 2020).

Other glimpses of complex effects can be found in qualitative surveys and occasional media reports, such as of widows whose husbands had died of COVID being left dependent on sons for social visits (author’s survey with a colleague), or of girl children being married off or even sold in poor families (CNN News18 2020). Media reports of rising domestic violence have also been common, but the figures provided by the National Ministry of Women and Child (Government of India 2020c) are likely to be underestimates, because victims had to submit complaints by email or WhatsApp, which most women do not have access to.

Overall, therefore, gender-disaggregated information on the impact of the pandemic on women in India is, at best, patchy. Few surveys have focused on intra-household effects or identified intersectional differences between women by class or caste. This complexity can remain invisible unless we know what to look out for. In particular, the deeper indirect gendered effects that arise within families, due both to preexisting inequalities and shifts in intergender relations, tend to remain concealed.

Also, the focus of almost all the surveys is on the negative effects of the pandemic, with rather little on the positive side. Yet positive stories are important to probe. In Kerala, for example, most women who were doing group farming before the pandemic, by pooling their leased-in land and labor (Agarwal 2018, 2020), were able to survive economically during the lockdowns (Kudumbashree State Mission 2020), whereas most individual farmers lost out. Similarly, women in Self-Help Groups (there are six million such groups in India) worked together to produce protective gear and hand sanitizers to earn a living during the pandemic, and many also reached out to support the hungry (Government of India 2020b; World Bank 2020). This suggests that group approaches are likely to be more effective than individual approaches in restoring women’s livelihoods and even providing social support.

Concluding Reflections

As argued here, COVID-19 is likely to have a range of complex and often indirect gender effects that could be missed. To identify these effects, we need to draw on our understanding of preexisting gender inequalities and social norms as well as examine literature that illuminates how people cope with economic and social crises, not just immediately but also in the long term. This can help us anticipate not only the direct and immediate impact but also the indirect and long-term impact. Media reports can further alert us about emerging adversities that could be incorporated into surveys. Such a conceptualization and learning from past crises and real-time reporting appears to be largely missing from many of the telephone surveys conducted during the pandemic in India (and likely elsewhere), with the exception of hunger among the jobless.

An anticipation of the diverse potential effects of COVID-19 can help us design surveys that can better capture impact that would otherwise remain invisible, and to guide policy for mitigating adverse effects, including those that are barely apparent yet. This would include (but not be confined to) the effects discussed here, such as an unequal sharing of food and hunger, increase in care work burdens and domestic violence, asset loss, the abandonment of women and girls due to poverty, the insecurities and dependencies experienced by widows whose husbands have died from COVID, and the educational reversals for girls. I believe tracking indebtedness and asset sales would be especially revealing of women’s economic situation and vulnerability within families. Moreover, tracking COVID-driven poverty by gender, beyond the immediate short term, is imperative, since long-term precarities can set in for women with limited resilience.

In conclusion, we might pose ourselves some questions for reflection. Why have many of the intra-household effects mentioned here been missed in most surveys, despite a long history of feminist scholarship in India and elsewhere? Is it a lack of theorizing, or a gap between feminist theory and empirical research? Or are the surveys undertaken largely by male scholars who are unfamiliar with the research? Equally we can ask: Why is there a disconnect between feminist research and action? In most writings on domestic violence under the pandemic, for example, the solutions offered are still helplines, shelter homes, and stricter laws, although research in the Global South shows that women who own a house or land, or have formal jobs, are at substantially lower risk of intimate partner violence than women without property or secure jobs. Yet we see rather little coordination between social movements against domestic violence and social movements for women’s rights in land and property.

Similarly, for several decades now, a diversity of empirical work across disciplines and feminist practice has demonstrated that working in groups empowers especially poor women, both within families and vis-à-vis communities and markets. This needs to inform policy discussions on women and the pandemic. “Building back better” will require creativity on many fronts, not least by creating a synergy between feminist theory, evidence gathering, and practice. Together these could add up to much more than the sum of their individual trajectories.