Conceptual framework

Resilience entails the capacity of individuals, households, or community members to adapt, to change, to anticipate, or to respond to shocks and uncertainty [12], recognizing their agency and ability to make informed decisions that impact their life [11]. Scholars [9] argue that resilience results from interactions (synergies and trade-offs) between three capacities— absorptive, adaptive and transformative—each of them leading to persistence, incremental adjustments or transformational responses, respectively. However, this is not a simplistic linear response—absorbing low-intensity shocks, adapting during periods of greater disturbance, and transforming when conditions become unbearable. Instead, the resilience approach takes a systemic view of socio-ecological interactions by situating vulnerability (the susceptibility to harm) within multiple interacting stressors and power hierarchies, operating simultaneously at various levels and impacting individuals and communities unequally [11]—thereby shaping individual and collective abilities to absorb/adapt to shocks.

Resilience is thus best understood as a characteristic that can be either supportive or erosive, depending on the strategies adopted and their consequences [11]. For example, individuals can experience severe poverty but appear resilient by absorbing/adapting to poverty and deprivation by making choices that compromise their well-being— a concept referred to as adaptive preference. This occurs when individuals living in precarious situations adjust their expectations and aspirations to cope with adverse realities through detrimental practices such as skipping meals, withdrawing children from school, or delaying healthcare. From a social-ecological perspective [9], a crucial aspect of resilience is ‘precariousness’ or closeness of a system to a no-recovery threshold, meaning that adaptive preference might be the only choice in times of crisis when households or communities routinely operate near threshold limits in everyday life.

In this context, to understand the coping responses of households to COVID-19, we use the Sustainable Livelihoods Framework [8](SLF) to situate the households in the broader context of socio-historical origins of structural inequalities and the recent global/local shifts that have compounded existing/created new vulnerabilities.

In India, the onset of neoliberal globalization has spurred greater integration of local economies with national and global circuits of capital accumulation. While this has led to increased aggregate incomes and more economic opportunities for some, it has also created newer forms of rural dispossession and exacerbated livelihood insecurity for marginalized households [8]. The deepening of agrarian capitalism caused by neoliberal policies has increased the casualization of labor, pushing the rural poor to sustain themselves through spatially fragmented and precarious wage work in both agriculture and non-farm sectors [13]. Declining rural and farm incomes, coupled with climate-induced shocks, have forced families to sustain themselves through irregular, precarious work and cyclical loans for agricultural inputs—often making food security and everyday life contingent on loan repayment cycles. More integrated economies have triggered migration outside of the villages, not only reshaping daily interactions in the village but also spatially distributing risks by creating multi-sited, trans-local households that are sustained through remittance networks and affective ties [8].

Rooted within this dynamic context of adverse incorporation [8], wherein poverty is systemically created and maintained through the incorporation of the working poor into markets on adverse terms, are existing social hierarchies or ‘relational power’. In India, these include structures of caste, class, and gender, which shape unequal access to resources, opportunities, and social capital. From a political economy perspective, this relational power is itself embedded in historical access to labor markets, social protection systems, and crucial environmental resources, including agricultural land—the uneven distribution of which structures differential vulnerability and resilience to shocks. Neoliberalism thus amplifies existing social hierarchies, as those with limited access to land, capital, and social networks are systematically excluded from access to socio-economic opportunities and support, and thus disproportionately face the consequences of shocks. For example, the caste-class nexus intersects with the lack of local social support networks to limit the capacity of circular migrant workers who survive on informal daily-wage work to respond to shocks.

In this integrated context of understanding resilience as situated within the SLF framework, we use the typology of Davies’ coping strategies [14] to examine how households managed their basic need for food in the context of COVID-19 (Fig 1). To categorize strategies as either short-term/reversible coping strategies or adaptive strategies, we use Davies’ critical conception of coping strategies as informed by the intensity of their use (how dependent the households are on these strategies compared to ‘normal’ times), the sustainability of these strategies when intensity increases, the motivation behind their adoption, and their effectiveness in meeting food needs. If adaptive strategies adversely affect future coping capacity, they are categorized as erosive.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Integrated Conceptual framework.

https://doi.org/10.1371/journal.pone.0350020.g001

This integrated framework illustrates how the coping strategies households employ are situated in existing ‘precariousness’ —that is, not only shaped by socio-historical processes and local/global market forces but also reinforced through structures of inequality like caste, class, and gender—that mediate unequal access to resources, social support, and entitlements [10]. Through this lens, a major global disruption like COVID produced novel, unequal outcomes, but also exacerbated existing vulnerabilities and inequalities. Disasters such as COVID-19 are not ‘natural’ shocks to an otherwise well-functioning system; they are, instead, a high-visibility indicator of interconnected and longstanding inequalities that created the conditions for systemic marginalization [15].

Building from this conceptual framework, our analysis is motivated by the following questions about households’ experiences and coping strategies during COVID-19:

• What strategies did households adopt to manage food?
• How did these strategies utilize changes to government entitlements?
• How were strategies influenced by geographical and socio-cultural location?

Field sites and sample selection

UP is a north Indian state with a population of more than 241 million. It is primarily agrarian, with 65.5% of the population residing in rural areas, and has a higher-than-average poverty rate (39.8% compared to the national average of 29.5%). Only 10.8% of workers are engaged in regular or salaried work. The female workforce participation was particularly low (17%). While most of the population is Hindu (78%), a sizeable minority is Muslim (20%). We selected four districts (Kanpur Nagar, Kanpur Dehat, Unnao, and Kannauj) to maximize sample diversity and to cover both rural and urban areas. Selecting neighbouring districts reduced travel time and costs, allowing us to spend more time in the field.

In rural areas, the families in our sample were mostly engaged in agricultural labor, relying on seasonal crops, subsistence horticulture, and sometimes livestock farming. Children as young as 8–9 years were engaged in agricultural labor. Housing ranged from sturdy constructions with staircases and courtyards to makeshift shelters with rudimentary roofing. In urban areas, participants included both first-generation migrants seeking livelihood alternatives amidst agricultural challenges and long-term residents. Most of these households relied on contractual/daily wage labor and faced financial instability exacerbated by illnesses, marriage expenses, seasonal changes, and so on. Housing infrastructure and living conditions posed significant risks, with families of 4–5 individuals residing in single-room dwellings lacking basic amenities, including sanitation and electricity. We conducted 238 interviews in 59 households.

In UP, our sample includes two types of migrants: out-migrating individual male workers from rural to urban areas (whose families we have interviewed) and incoming migrant families (in both rural and urban areas). Individual workers were often circular/seasonal migrants, with few connections to the destination. These were also the migrant workers who returned to rural areas due to job losses and a lack of local support networks during COVID-19. Migrant families, on the other hand, are those who have moved to UP from neighboring states like Bihar.

In contrast to UP, Goa is a small coastal state in southwestern India, with the highest GDP per capita in the country and high literacy rates (88%). Manufacturing constitutes more than 40% of the state’s GDP [16] and tourism plays a significant role in the economy. We focused on migrant populations, who experienced heightened levels of precarity. According to the 2011 Census [17], around 43% of inter-state migrants to Goa originated from Karnataka, 26% from Maharashtra, and 12% from the states of Uttar Pradesh, Bihar, and Jharkhand, while only 6.7% came from southern states such as Kerala, Andhra Pradesh, and Tamil Nadu. We interviewed households in urban and peri-urban South Goa, including short- and long-term migrants from states such as Karnataka, Maharashtra, Bihar, Odisha, and Assam. In total, 105 participants from 27 households were interviewed.

In Goa, however, only incoming families have been interviewed (with only one family having a seasonally migrating male worker). These families comprise both first-generation migrants who moved to Goa themselves in search of work and established families here, as well as second-generation migrants (whose parents moved to Goa). They have varying degrees of enduring connections to their places of origin, including both affective ties and financial support from relatives in their home states.

From the full sample of 86 households, the study purposively selected 15 households that had experienced severe COVID-related shocks for in-depth narrative analysis, across both states. Detailed information about each of the 15 households in our sub-sample has been included in a supporting information file (S1 File_Description of sub-sample). The serial numbers in the supporting information file have also been referenced throughout the manuscript, along with narratives from each household. Additionally, any instance where a serial number is not provided indicates that the excerpt is not from the sub-sample.

We used purposive sampling to select households, maximizing diversity across caste, religion, location (rural/urban), household structure (single vs. multifamily), and family type (nuclear/single parent/multigenerational). (Details in Table 1.) For the sub-sample, too, we purposively sought to maximize diversity, focusing on including diverse experiences of COVID-related shocks across migration status (migrant/non-migrant), location (rural/urban), and household structure.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Socio-demographic characteristics of the sample.

https://doi.org/10.1371/journal.pone.0350020.t001

Selection continued until narrative patterns reached thematic saturation.

Data collection

The Faculty of Social Science Ethics Review Committee at Lancaster University, United Kingdom approved our study (FASSLUMS-2022–0953-RECR-3) on 10/11/2022. The study was also approved by Institutional Ethics Committee at Indian Institute of Technology, Kanpur (IITK/IEC/2022–23/I/19), the project’s host institution in India on 09/09/2022.

The fieldwork for this research was conducted between 22 December 2022 and 23 March 2023 by SM, CV, and AR. We collected basic sociodemographic details using a household roster adapted by JF from the Demographic and Health Surveys (DHS). This included questions about the age, caste, religion, education, and occupational status of each household member, as well as general questions about the dwelling infrastructure, sources of drinking water, sanitation, sources of fuel, etc.

After completing the roster with one adult member, we conducted semi-structured interviews in Hindi with multiple members of each household, including as many members aged 7+ as possible. We included children based on evidence that children’s life trajectories are significantly altered by food insecurity, yet older children-adolescents in particular are frequently overlooked in food insecurity studies. [18,19]

Recognizing varied levels of literacy among our participants and the possible stigma associated with limited literacy, in addition to sharing written participant information sheets (PIS), our team (CV, SM, AR) provided a detailed oral summary of their contents to all participants. We explained the objectives of the research, why participants were being approached, and the possible benefits and harms of participation. Additionally, the purpose of the audio recordings was explained to participants (both children and adults), and they were interviewed only if they were comfortable with the recording. Once the participants consented to the interview and audio recording, the recorder was started, noting the date, time, location, and household number of the interviewee on record, and noting the consent in simultaneous field notes. Because we never undertook lone work in the field, this was witnessed by at least one other team member. For children, parental consent was obtained, and then children’s assent was obtained in front of their parents. This procedure was approved by the ethics review boards of both the UK and Indian universities involved in the study. Refusal was acknowledged in various forms throughout the interview, whether verbally or nonverbally, at the start, through skipping a question, or by refusing to answer, even after consenting to the interview itself. The team’s contact numbers, and the heads of their departments were shared, and participants were informed that they could ask for their data to be removed even after participation.

Although we originally did not record conversations while completing rosters (assuming only ‘factual’ details would be shared, which would be documented in the roster), we began recording roster discussions when we realized that the roster sparked rich conversations as well.

We deployed age-specific interview guides (for ages 7–12, 13–17, and 18 + years). These were conceptualized in English by the project lead, JF, with inputs from the Co-I (SV) and project partners, then discussed and revised with inputs from the field team (CV, SM, AR), and collaboratively translated into Hindi by the field team. Questions on meal preparation and consumption, food preferences, and household inequalities were common to all instruments (wording varied somewhat by age group). The children’s guides had unique questions about schooling, meals at school, care work/responsibilities at home, and allocation of their preferred foods at home. Drawing was used as an (optional) strategy to help children open up and discuss food-related issues. We obtained children’s oral assent before photographing their drawings. The adult questionnaire had more questions about livelihoods, food prices and availability, retrospective reflections on COVID and its impact, social protection, climate change, and agricultural practices. We also included a version of the UN Food and Agricultural Organization’s Food Insecurity Experience Scale (FIES) [20] for adults and a modified version for children [21]. We modified the FIES by adding open-ended probes (e.g., ‘Can you tell me more about when that happened?’) to deepen understanding of food insecurity experiences. A summary of core questions in the household roster and the adult and children interview guides has been added as a supporting information file (S2 File_Core questions).

We piloted the interview guides in seven households in UP in early December 2022, including all eligible age groups, and then revised them to thematically cluster questions and avoid repetition. We prepared a small grocery basket (₹800) for each household to express our gratitude to the participants for their time.

We approached both interviewing and data analysis with continuous reflexivity [22], acknowledging how interviewer positionalities can affect participant sharing, interviewer responses, and other interpersonal dynamics. We (CV, SM, AR) are three female researchers born into Hindu families. We are all experienced qualitative researchers with prior experience in collecting data on sensitive issues in various locations in India. All of us have completed our doctorates in social science disciplines. When interacting with participants, we answered queries about the project and ourselves, scheduled interviews at the participants’ convenience, and tried to ensure as much privacy as possible while interviewing. In addition, we made continuous attempts to enable comfortable sharing among participants through regular check-ins and by repeatedly emphasizing that we were trying to learn from their lived experiences; thus, all their experiences were of value.

Audio recordings were translated into English, including pauses, false starts, laughter, and the associated background information/noise. Transcription and translation were carried out by the field team in discussion with JF, and by one additional translator from UP with a social science background, supervised by the field team. The field team also created a translation guideline document and a glossary of terms to be retained in Hindi when an exact translation seemed difficult.

Data analysis

Data analysis was conducted in two stages. First, we conducted a broad thematic analysis by carefully familiarizing ourselves with the data, coding, and reviewing codes for interviews from the entire sample. These were then differentiated by rural-urban location and migrant/non-migrant status to understand their diverse experiences of COVID. Drawing from the above integrated framework, we categorized results into three broad themes: 1) Strategies for food management, 2) Government support easing resource constraints, and 3) Rural-urban linkages. While the first theme keeps its focus centered on the households to understand coping strategies deployed during and after COVID, the second and third themes take a step back and explore the local, national, and global processes, social policies, and structural inequalities that limit a household’s choice of coping strategies and subsequent resilience for such future events.

To grasp the nuances of the strategies households adopt, we delved further into our intrahousehold data to identify diverse experiences within households, some of which were spatially fragmented. In this stage, we paired the thematic analysis with a narrative analysis of 60 interviews with members of 15 households (details in Table 1 and S1 File_Description of sub-sample) that we identified through our process of data familiarization and coding as having been especially impacted by COVID. These included households where members had either died due to COVID/related illnesses or those which faced lingering impacts of untreated illnesses that occurred during COVID; households where additional family members had to be supported due to reverse migration; and families who had not recovered from economic shocks posed by loss of a job/business during COVID. This added layer of analysis was intended to further illustrate the nuances in a household’s response to an external shock by studying the interviews of all household members in tandem. Narratives of each generation reflect a social and historical location from which an individual constructs and articulates their lived reality [23]. Thus, to understand the meanings different members of households attached to food practices as a response to the additive constraints posed by COVID, interviews with all members of one household were examined simultaneously, paying attention to the three-dimensional aspects of sociality, temporality, and physical space [24]. In the case of our transcripts, this meant paying attention to how the COVID stories were shaped not only by individuals’ socio-cultural and geographical locations but also by their positions within the household and the continuity of events/relations before and after the lockdown period.

Transcripts were coded in NVivo 12 and supplemented with additional field notes on the data collection sites and nonverbal cues observed by the interviewers. The coding process was primarily inductive. Initial codes were identified through data familiarization. These were then grouped into relevant themes to improve analytical coherence in interpretation. For example, all food management strategies like reducing consumption, buying less quantity, buying cheaper quality, avoiding certain food items, adults managing by reducing self-consumption, managing food budgets, pooling resources, etc., were clubbed under three sub-heads: 1) Consumption cuts, 2) Changes to household composition, and 3) Tough Trade-offs. A quantitative summary of a few themes/associated codes is available as supporting information (see S3 File_Quantitative summary of themes). For the narrative analysis, we examined interviews from each household together to understand the meanings attached to each decision and its impacts, as shared by all household members.

Inclusivity in global research

Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the Supporting Information (S4 File)

Food management strategies

Adjusting household composition to make ends meet

Another way households cope with crises is by changing their household composition. This could be achieved through household extension via intergenerational co-residence [27] or by sending children away to live with relatives, or by some members moving away to reduce household size [28]. This could be considered an adaptive measure in principle, as it involves a structural change in household composition. However, it had limited sustainability in the context of COVID-19, as compounded stressors like increased medical expenditures and loss of income further exacerbated existing vulnerabilities.

In some families, household extension meant that those previously operating as nuclear units for daily expenses began sharing resources, such as cooking fuel and electricity costs, by jointly preparing meals with relatives living nearby. This temporary shift towards resource sharing reverted to the nuclear model once employment stabilized after the lockdown.

In one such case (Fig 3), three distinct households, comprising members of the same extended family, occupied separate rooms within the same premises both before and after the lockdown. However, the pandemic significantly altered their intra-household dynamics. The daughter (household 2) noted that the families began cooking together to conserve cooking fuel and to ensure that the children did not go hungry. Despite these coping mechanisms, the household was unable to mobilize sufficient resources when her father fell ill, leading to his death during the pandemic. Drawing from Davies’ [14] while pooling resources might have been an effective coping strategy with respect to food, its sustainability became limited as the intensity of shock increased, compounded in this case by health expenditures. During the interview, she became emotional as she recounted how, despite her best efforts, she felt she had failed to provide adequate care. She said,

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Changes to household composition during and after COVID-19 lockdown.

https://doi.org/10.1371/journal.pone.0350020.g003

He became paralyzed later and during COVID...his condition was so bad that he was ill, and he was not taken to the hospital… I used to sell masala (betel nut mixture) all day… and throughout the day, I used to think I should collect enough money before I go home and bring painkillers to my father… feed him fruits etc. so that he becomes well as before…My father was very good. He never let me feel any dearth of anything… like shade from the sun.

(S.no 2.,Woman, 38, urban-non-migrant, Scheduled Caste)

The figure denotes how the composition of three households residing in one premises changed due to COVID-related events and scarcity. The red dotted lines denote each household’s composition before COVID-19 and grey dashed lines denote the household composition of each household when the fieldwork was done (March,2023).

A disabled single mother, she recounted that she remarried during COVID, against her mother’s wishes, hoping for financial stability. Her ten-year-old son remembered the period as particularly difficult, noting that his stepfather was unable to secure employment, which left the family struggling to meet basic food needs. He explained that he sometimes lied to his grandmother that food was cooked in their nuclear household, so she would not worry.

At that time, papa did not have work... so he used to do small work here and there and like… if he got 500 rupees, then that money would run the house…Like sometimes the food was not made… if there is only a little flour in the house, then it used to be made and eaten.

(S.no 2, Male child, 10, urban-non-migrant family, Scheduled Caste)

In Household 1, the daughter-in-law passed away shortly after the father’s death. Her husband (Son, Household 1) was absent from childcare when she was alive; her passing left the children without material or emotional support. Their grandmother (Mother, Fig 3) accepted care responsibilities and used her contacts to obtain a government loan for a small shop to generate income. By our fieldwork in February 2023, these families returned to a ‘new normal’ with three separate households again, although quite different in structure from pre-COVID times. However, the earning members asserted that livelihood opportunities had not returned to pre-COVID levels.

In a migrant household in Goa (S.no. 11, urban-migrant family from Karnataka, Scheduled Caste), the eldest of the four children grew up with the husband’s financially better-off parents to ease the financial burden on their nuclear family. During COVID, another daughter was sent to live with the grandparents when managing finances became difficult. Thus, while livelihoods had limited adaptability due to structural constraints, it can be argued that such adaptations made by families helped to manage resilience by temporarily shifting households away from threshold limits [9] and ensuring that long-term sustainability was not affected.

However, for individual migrant workers, changes to household composition upon their return put additional strain on the already limited resources. Distress-motivated reverse migration during COVID lockdowns led to unexpected changes in household composition, especially in rural areas. Rural-to-urban migrants in UP tried to move back to their villages, assuming that managing food would not be an issue, given access to agricultural stocks and/or social networks. In a joint family in UP, two unmarried brothers returned from working in the city to the rural area where their eldest brother lived with his wife, four children, and an unmarried sister. The married brother (MB) and his wife noted that it became tougher to provide for their children with additional people in the household.

MB: Like during COVID, everything was tight. Now the brothers who lived away…there was no work for them, so they also came here.

Interviewer: So you had to manage for everyone?

MB: Of course, whatever food is there, the entire family will eat... See, this is the problem, if there are two earners then there is more money, if there is a single earner, then the money will be tight.

Interviewer: So how was it managed during that time brother?

MB: With god’s grace, everything went on. And then it is the government. They gave ration etc.…they helped a lot during that time.

(S.no 5, Man, 55, rural, non-migrant, general caste)

The unmarried brothers in the household were also worried about having limited opportunities to return to the city or to find local work. Speaking closely to the idea of their household operating near the boundary of a resilience threshold, the unmarried sister jokingly used a local idiom to compare their household’s situation during lockdowns to a faulty vehicle that must be manually pushed to start (‘Dhakka plate gaadi’), indicating that they had no choice but to push through tough times. All members remembered this as an extremely stressful arrangement, as work opportunities remained scarce.

Tough trade-offs

Davies [14] warns that short-term coping strategies might have lasting consequences on long-term development. This occurs when a choice is forced between basic subsistence and long-term sustainability [29], leading to the adoption of strategies that are erosive in the long term and hamper future resilience [10]. An important reported example of this in the context of COVID was the choice between continuing children’s education and managing everyday food expenditures. We found that some children had not returned to school, as families struggled to regain pre-COVID income levels or were missing school for prolonged periods to augment family income. Intensification of wage work by increasing the number of earning members is an important routine coping strategy for poor households [30]. Numerous children talked about having had to leave/skip school, shift from private to public schools, or start agricultural labor to contribute to household earnings because of financial troubles after the lockdown. Som‌‌e families prioritized resources to educate their sons, while daughters were withdrawn.

In one household in Goa, the father’s small business became unsustainable during lockdown, and he went into hiding to avoid debt repayments, leaving his wife and three young children behind. The wife recalled struggling to find a place to stay, barely feeding the children by working as a house helper and taking informal loans. Although her daughter returned to school post-lockdown, she was grateful that schools were closed during the lockdown, as her older daughter could take on care responsibilities for her siblings.

Good thing the school was closed. Everyone says it would be better if school was open, I pray it was better that the school was closed. Because the children were at home, I got to go for work. Otherwise, how would I have done this, …there was a small child, how could have I done work and all?... daughter used to take care of him, the elder daughter. Both the children.

(S.no 15, Woman, 36, rural migrant family from Karnataka, Muslim)

In other families, people spoke of impossible choices during the pandemic between medical expenditures and managing food—the ‘treat or eat’ dilemma [31]. For example, a woman in Goa managed without medicine for her two-year-old’s high fever, since buying medicine would have meant that family members would go to bed hungry (S.no 13, urban migrant family from Madhya Pradesh). In another household from UP, a participant discussed how they had to eat plain roti when medical expenses cut into food budgets.

Participant: Like if I have 200rs, I spent that on medicines,then I can’t buy vegetables, what will we do then, you tell? So that happens.

Interviewer: So, what is cooked that day?

Participant: Then nothing, just eat it dry (sookhi’- reference to wheat roti). If the entire money is spent on medicine so where will I get it (vegetables) from?

Interviewer: So, does everyone in the family eat dry then?

Participant: What else.

(S.no 5, Woman, 43, rural, non-migrant, general caste)

Non-crucial medical expenses were deprioritized to make room for everyday expenses. In another migrant family in Goa, a father of four delayed operating on his back injury so his wife’s income could be prioritized for children’s expenses during the lockdown (S.no11, Man, 38, urban migrant from Karnataka, Scheduled Caste).

These trade-offs are often undesirable decisions of last resort (adaptive preferences) in precarious households when other social or state support is unavailable to absorb shocks, and they can have lasting consequences for future resilience by eroding earning potential and compromising families’ mobility goals.

Rural-urban linkages: Diverse yet shared vulnerabilities.

In rural areas, several households reported that agricultural/industrial labor still provided jobs (with delayed/erratic payments), especially given the lockdown exemptions for agriculture. In urban areas, it became impossible to find jobs, especially for the daily wage migrant workers, triggering a wave of reverse migration to rural areas.

However, within urban areas too, there were disparities among the laboring families based on the type of work and associated local social networks, which are strongly shaped by caste and class. While most informal laboring migrant families/workers in urban areas faced threat to their means of employment, a few non-migrant families continued to find work due to the association of their work with higher officials/those working in essential services. For instance, despite the widespread precarity experienced by many drivers in urban contexts during COVID, one participant interpreted his relative stability as being shaped by his caste and class position. He repeatedly stressed how his caste location enabled him to access more advantageous social contacts and maintain them. In this sense, both caste and class networks appeared to function as a safety net. When asked about his family’s experiences he shared:

Child, for us, during Corona… there were no problems for us...I was working with a doctor; I did not even have a minute to spare. There was scarcity at labor level work,otherwise the adhikari (higher officials) had a lot of work. And my contact and work is with the adhikari (higher officials), so there was no question of having problems. I did not face any difficulty.

(Male, 46, urban, non-migrant, general caste)

While the need to work amid risk to health and life is also reflective of how informal workers coped with crisis, we found that laboring families belonging to better-off castes and class were able to rely on their caste-based networks and long-term work affiliations to find work to cope with crisis. In a rural higher caste family (S.no.5), when two more brothers returned to the village, unable to find work at the destination, the local married brother who worked as a contractual employee in a factory was able to negotiate some limited days of work for both of his brothers with the owner, based on his long-term association. This was even though the factory was operating at reduced hours. This limited additional income was crucial in helping meet the expenses of the extended family. However, these networks of support were not equally available to all, as some lower-caste-class households, including rural non-migrant daily wage workers, struggled to find work.

In rural areas, ownership of assets, like land and cattle—often embedded in caste-class structures benefiting the most privileged in these hierarchies—also served as important cushions to navigate the crisis and manage food insecurity in the context of the shock produced by COVID. Some agrarian families in rural areas had enough cereal to eat (produce from their own land/return from crop-sharing arrangements) but were struggling to sell produce and consequently had limited/no disposable income. Similarly, while lentils were inaccessible to most families, some cultivating families had stored stocks but limited cash flow to buy other items. Some marginal farmers also began growing vegetables on a portion of their land for family consumption as markets became inaccessible or cashflow became limited. Likewise, milk producers managed excess milk production with no means to sell it or alternative sources of disposable income. In a relatively well-off, higher-caste, joint family in UP (S.no. 6, urban, non-migrant, general caste), excess milk production had to be used creatively in the absence of cold storage infrastructure. This included making and consuming a variety of dairy items such as sweets, ghee, butter, and cheese. Even in these varied forms, milk was too much for family and relatives to consume, so a portion was regularly wasted. Meanwhile, since they were unable to sell their milk products, earnings were limited, and they also had difficulty managing day-to-day expenses, such as grains, vegetables, and children’s school fees.

Thus, while these asset-owning higher-caste-class families were affected by disruptions in crop and milk sales, access to land and cattle allowed them to absorb shocks without the need for erosive coping. In these cases, the severity of impact was also cushioned by multiple food sources and multiple income streams. For example, while the elder son in the cattle-owning family lost his job, the household’s food was not dependent on his earnings alone and thus remained food-secure.

However, marginalized households lacking access to assets faced compounded risks arising not only from their caste–class position and limited access to resources and networks, but also from adverse incorporation under neoliberal restructuring of rural economies. These shifts necessitated migration and created multi-sited household vulnerabilities and translocal risks, as livelihoods became dispersed, precarious, and increasingly dependent on unstable market relations. The most telling example of this was changes to crop-sharing arrangements in the context of reverse migration. In some cases, landowners withdrew from batai (crop-sharing) arrangements to save costs when their sons returned from cities and could work their own fields, leaving landless laborers without access to work and related grain payments at a time of acute need.

Crop-sharing arrangements in UP include batai (one-half), tihai (one-third), chouthai (one-fourth), and on theka (a pre-decided amount of crop must be given to the owner), indicating cost sharing between agricultural laborers and landowners. In some chouthai arrangements, laborers contribute their labor and machinery costs but do not share input costs in return for one-fourth of the produce. One of our agricultural laboring participants (S.no. 7, Male, 44, rural, non-migrant, Other Backward Caste) explained that this arrangement was easier for the landowner to break than when laborers also shared a portion of the input costs. His family was severely affected when the landlord broke the arrangement to save costs. He explained that families like his were heavily reliant on this cereal stock for food throughout the year and had limited cash flow to buy it from the market. This case clearly shows how integrated rural economies are now impacted by livelihood insecurity rooted not just in local contexts but also in broader systems and processes. The shock of COVID-19, which triggered reverse migration to rural areas, further strained already struggling rural economies—thereby creating new forms of vulnerability while exacerbating existing ones.

Table 2 summarizes the diverse yet inter-related vulnerabilities of four key groups in our study: Rural non-migrants, circular migrants, urban migrant and urban non-migrants.For both circular and urban migrants, the political economy of rural livelihoods—characterised by limited diversification of farm incomes and limited availability of both farm and non-farm livelihood opportunities—triggered distress migration to urban areas. For urban migrant families, this meant a structural change in their household location and a susbequent break from rural income sources; whereas for the circular migrants, the aim of migration was to secure diversified access to livelihood opportunities for the individual worker, while sometimes retaining seasonal access to rural livelihod opportunities. This seasonal/circular migration of predominantly male workers and longer-term shift to urban areas of families with dependents created complex, multi-sited households that are maintained through affective and financial ties across geographies.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Experiences of COVID-inter-connections and differences.

https://doi.org/10.1371/journal.pone.0350020.t002

We find that, while all economically precarious families were at risk in the context of COVID-19, circular migrants were particularly vulnerable. While they migrated to urban areas to mitigate/diversify their existing risks, they generally did not have access to local support networks or standard government entitlements at their destination. This was largely due to their mobile lived realities, documentation issues or family members still accessing their entilements at origin due to which entitlement location was consciously not updated. On the other hand, some rural non-migrant families, mainly small/marginal farmers and agricultural laborers, were heavily reliant on local welfare systems like the PDS, and the additional allocations during lockdowns were especially critical for maintaining food security.

While circular migrants and urban-migrants faced similar disruptions in informal job opportunities, some longer-term urban migrants benefitted from established community based/other social networks in the city, a support that individual circular migrants often lacked. Given the interconnectedness of rural-urban economies, reverse migration increased pressure on limited rural jobs, strained household resources, and intensified dependence on family networks and remittances across all groups. Lastly, while urban non-migrants also faced loss of livelihood opportunities due the disruptions, they too, often had access to local social support networks developed over time that helped them to absrob shocks without the need for erosive coping.

Our findings thus suggest that diverse coping strategies and ability to absorb shocks without resorting to erosive coping, then, depended not only upon access to livelihood opportunities but also local support networks (as embedded in structures of caste-class and length of stay in an area) and access to government entitlements, which acted as important buffers for those with access.

Limitations and Strengths

Our study is based on a purposive sample of lower-income households in selected districts of only two states of India— Uttar Pradesh and Goa. The study thus only provides locally contextualized evidence and does not make any generalizations for any group, community, district, state or the country as a whole. While we diversified our sample by caste, religion, household types, and family structures to be inclusive of diverse and marginalized voices, we do not claim to generalize our findings. In addition, it is possible that, despite purposively seeking marginalized households, we missed households facing the greatest precarity. For example, although housing conditions were unsafe, insecure, and unsanitary for some households, these conditions are still not as precarious as having no housing at all.

On the other hand, we use in-depth interviews with multiple members of the same household to provide detailed accounts of the lived realities, meanings associated and mechanisms and support available to diverse groups during a particularly challenging point in time. We foreground participant voices, creating space for marginalized and precarious groups to articulate their realities in their own terms. Qualitative studies like ours thus play a crucial role in deepening understanding by providing evidence of how individuals and households interact with larger socio-economic structures, policies, and hierarchies, thereby contributing to more grounded policymaking. This is particularly relevant to questions of food insecurity, as quantitative studies of nutrition often miss the trade-offs and coping strategies families adopt to address food shortages, which are rooted in socio-cultural and economic structures of inequality that create intersecting vulnerabilities. Additionally, our study makes a unique contribution by interviewing and including the perspectives of older children (7–18-year-olds), who have been labelled a ‘forgotten population’ in the food insecurity literature [18].

Due to practical limitations on fieldwork during COVID, we were unable to conduct interviews during lockdowns and instead relied on retrospective recall, conducting interviews 28 months after the last lockdown (August 2020). While this carries a risk of recall bias, COVID was an extremely memorable, unusual, and disruptive period whose impacts persisted well after the lockdown period. Additionally, even after complete lockdowns were abandoned as a strategy, restrictions on movement and certain types of work remained, affecting livelihoods and incomes. While asking participants to recall experiences after the event carries a risk of recall bias, linking to memorable or emotive events may reduce bias [32–35], particularly when recall focuses on general experiences rather than fine-grained specifics (such as event dates), as was the case in our data. While there is still a risk of recall bias, we note that COVID-19 was a one-of-a-kind, extremely memorable, and emotive period. Lastly, we discuss the ongoing challenges participants link to COVID-19. These include reduced incomes and frequency of workdays and lost businesses, but also long-term effects of shocks such as the death of a breadwinner/support provider or lasting impacts of delaying surgery because of COVID-19. The recollections our participants shared are thus often tied to major upheavals in their lives that were directly caused by the pandemic. Temporal slippage, thus, though a possibility in any study of this nature, might not undermine the discussions as participants relate causality of challenges to the pandemic, even if timelines were blurry.