ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Spatialising Urban Vulnerability Perspectives from COVID-19 Food Relief in Delhi

To address hunger, the Government of Delhi had issued temporary ration e-coupons in the first COVID-19 lockdowns of 2020. This article uses a data set of nearly 17 lakh households that applied for e-coupons to measure and spatialise food insecurity in the city. It does so to measure unmet demand for social protection as well as to draw learnings for the design of urban social protection systems.

In response to the first wave of the COVID-19 pandemic, the Government of India announced a nationwide lockdown on 24 March 2020. The lockdown was seen as a public health intervention, one necessary to curb the spread of the virus. However, as many have argued, it had its own effects and, in particular, affected food security in India’s cities (Lahoti et al 2020; Mishra and Rampal 2020). As lockdowns progressed, a combination of lost income and restricted mobility created large-scale hunger among urban residents. Relief efforts logged innumerable distress calls, hunger helplines were inundated, and the sight of long lines for public and private distribution came from all major cities in the country.1

The state specifically attempted to respond to this food insecurity (Singh 2020). The Government of India, under the Pradhan Mantri Garib Kalyan Anna Yojana (PMGKAY), announced a 50% increase in take-home ration for each individual beneficiary covered under the National Food Security Act (NFSA), 2013. In Delhi, 71 lakh of the city’s 1.68 crore residents were already covered under the targeted public distribution system (T-PDS), which caps enrolment at 40% of the residents for urban areas. With their “ration cards,” as they are colloquially known, they were entitled to dry rations of rice and wheat. The Government of Delhi further extended this entitlement to those without ration cards by announcing a scheme for temporary ration cards, referred to as e-coupons. The Delhi government administered the e-coupon through a mobile and web application, inviting anyone present in the city without a ration card to apply.2 The e-coupon, given per household, would entitle individuals to dry rations as in the PDS system to be collected from temporary ration centres set up at the nearby schools. For both ration card and e-coupon holders, an additional food basket was also distributed, containing pulses, oil, salt, and spices. The first announcement was for 10 lakh e-coupons, arguably the state’s estimation of the additional need they had to cover beyond existing PDS enrolments. Within days, however, in one of the richest urban regions of India, close to 60 lakh people applied.3

The COVID-19 pandemic has made visible foundational vulnerabilities within India’s urbanisation. A growing body of work has sought to document and assess these vulnerabilities, seeking not only to bear witness but also to offer knowledge that could shape a different urban praxis. One strand of this work has amplified specific gaps and deficiencies in current urban systems as they relate, for example, to migrant workers (Aajeevika Bureau 2020; Adhikari et al 2020) to specific caste and gendered impacts on livelihood loss and hunger (Chakraborty 2020; Deshpande and Ramachandran 2020), and to the gaps in social protection entitlements for informal workers that found themselves outside safety nets precisely at a time of deep crisis (Bhan et al 2020; Kesar et al 2021; Rai Chowdhury et al 2020). Others have argued that relief practices can act as templates, experiments, or examples of innovations of what post-crisis urban systems should look like. Writing alongside this body of work, we ask: What can we learn from a temporary expansion in food entitlements about urban vulnerability and social protection systems?

We argue that the Government of Delhi’s e-coupon programme represents an invaluable archive through which to assess both the extent and geography of vulnerability in contemporary Delhi as represented by food insecurity. It does so in two important ways. First, it allows us to get a sense of unmet demand, that is, the number of households outside the PDS system that, arguably, are vulnerable enough to have been included within it. In this, it allows an assessment of social protection based on vulnerability rather than a static, narrow measure of income poverty. We ask: What is the extent of this unmet demand, and “where” in the city is it located?

Second, the data for e-coupons has been made available to us by “locality.” This is a category of internal administration used by the Department of Food, Supplies, and Consumer Affairs to describe the geographical catchment area of each PDS shop. Locality is not a commonly known scale of urban governance, and indeed there are no publicly available maps of localities that we know of. Yet, conceptually, localities represent a sub-ward scale to assess food insecurity using an already existing administrative category. We mine the possibilities of what can be said about food insecurity in particular and vulnerability more broadly at this scale. We ask: In large metropolitan regions, what could a different scalar imagination for urban governance look like when it steps away from the ward, assembly constituency, and district?

We begin by laying out data on the ration cardholders and e-coupon applications, mapped and modelled onto existing PDS shops (n=1,923). We then present estimates of a geography of unmet demand by constructing an unmet ration ratio (URR) that measures the degree to which e-coupons cluster around existing PDS shops. We aggregate and present this unmet demand by existing scales of urban governance, showing how successively higher levels of aggregation distort effective targeting and coverage within a food security system (Rajpal et al 2021). We then, in the absence of any maps of “localities,” posit a hypothetical provocation by using a Voronoi tessellation to suggest what this sub-ward distribution pattern could look like. We argue that even such a hypothetical exercise shows that it is worth considering the locality, and not the ward, as the appropriate scale to assess food insecurity and to design systems for food distribution. In our conclusion, we suggest that such a scalar shift has lessons to offer for other components of urban social protection beyond food, reiterate the critical necessity for sub-ward-level urban data governance, and consider what, using the idea of “localities” more widely, may make it possible for both provisioning and targeting of social security measures.

Data and Methodology

The article uses two different data sets. The first is data on e-coupon applications received by the Government of Delhi. Each application is from a household and, like ration cards, has names of all eligible individuals of a household as long as none is named on an existing ration card. Government declarations suggest 60 lakh individuals applied through the e-coupon scheme.4 The data shared with us and used for this analysis consists of a total of 17.8 lakh households applying for e-coupons. This yields a multiplier of 3.4 to convert from households to individuals. This data is organised at the “locality” level with applications covering 2,268 localities in Delhi.

The second data set is of the PDS system. The locations of PDS shops and the counts of ration cards linked to each shop were scraped from the Delhi government’s NFSA website.5 For ration cards, we used a multiplicative factor of 4.2 to convert from households to number of individuals.6 Our final data, then, consists of the total number of ration cardholders for 1,923 shops in addition to the e-coupon applications from the 2,268 localities.

We used the nearest neighbour algorithm on the data set of e-coupon localities and PDS shops.7 The computation yielded the closest ration shop to each e-coupon locality based on proximity. Thereafter, e-coupon counts associated with every e-coupon locality were assigned to the closest ration shop. This generates two key descriptive variables:

(i) Total ration cardholders

RC= ∑RationCardHoldersi where i denotes a ration shop.

(ii) Total e-coupon holders

EC= ∑ECouponsAssignedtoShopi where i denotes a ration shop.

We chose to use this method because it mirrors the existing governance structure of the PDS system. Ration cards are attached to a proximate ration shop, and the Department of Food, Supplies, and Consumer Affairs, as we argued above, uses the term “locality” in its internal admi­nistration to describe the geographical catchment area of each shop. The logic of this system is that PDS shops must be local, walkable, and accessible from the place of residence. Applications are thus matched to a proximate PDS shop much like in our method. By matching e-coupons to a proximate PDS shop, we are replicating the logic of the present governance structure.8

We then generate a metric to understand the scale and geography of demand for food during the lockdowns. This
is the URR, or the ratio of e-coupon holders to ration cardholders. It is as follows:

where scale denotes a level of administrative/operational scale (PDS shop, municipal ward, assembly constituency, district),

j denotes a unit at particular scale, and i denotes a PDS shop within unit j of any scale.

For PDS shop i,

becomes

since ration cards and e-coupons are already aggregated at the level of ration shops.

For higher levels of aggregation, URR, as defined above, becomes the ratio of the sum of e-coupons assigned to all the shops to the sum of ration cards linked to all those shops within a unit in that level of aggregation. For different levels, URR are defined as below:

(i) URRDistrict,j denotes URR for jth district where j is one of the 11 districts of Delhi.

(ii) URRConstituency,j denotes URR for jth assembly constituency where j is one of the 70 assembly constituencies of Delhi.

(iii) URRWard,j denotes URR for jth ward where j is one of the 272 municipal wards of Delhi.

For example, at the scale of districts, URRDistrict, North Easis the ratio of total e-coupon counts across all the ration shops to the sum of ration cardholders across these shops within the North East District. For all the wards and constituencies, their respective URRs are similarly tabulated.

How does URR relate to unmet demand? At URRRation Shop of zero, there were no additional applicants at a particular PDS shop during the lockdown. This implies that current enrolment in the PDS system may have been sufficient to ensure food security in that locality. As URR becomes greater than 0, it implies that residents in the locality not covered by the PDS had unmet need. At a URR of 0.5, the number of such residents is 50% of the total number enrolled in the PDS system. At URR=1, there are as many e-coupon holders outside the PDS system as the ration cardholders within it. For a URR>1, then, there are more e-coupon holders outside the PDS system than those enrolled within it, indicating a significant unmet need. We report URR findings and discuss this in greater detail.

Unmet Demand: What do URRs Tell Us?

From the 1,923 ration shops within our data, no additional e-coupons were reported from 855 (approximately 44.5%) of shops (Table 1). Of the remaining, the largest (approximately 55%) share is of URRs up to 1, implying that e-coupon applications were up to the same number of individuals as those enrolled in the PDS system and attached to that PDS shop. What is concerning is that, for 415 shops (approximately 21.6%), the URR is between 1 and 5. For every one of these shops, not only were the number of people applying for e-coupons greater than those enrolled in the PDS system, in many cases, this number was closer to two, three, four, and even five times the numbers currently enrolled; for 70 shops, the ratio is higher than 5. For these shops, however, it is possible that such high URRs are the result of the method of allocation to the closest shop rather than a genuine reflection of the concentrated demand, especially if several shops are close to each other as well as the e-coupon points; these shops make less than 4% of the total PDS shops in our sample.

How do we interpret the URRs? One line of interpretation is that they represent a specific form of vulnerability: temporarily increased demand due to an external shock like the first wave of the pandemic and the lockdowns. This is certainly correct. The e-coupons measure a crisis-specific food demand that may not translate or sustain into the post-crisis need. Taking this to be true, it implies that the URR is a map to understand the need for disaster preparedness, marking localities where impacts and vulnerabilities are higher. Measured at the shop level, this yields a geography of vulnerability far more granular and specific than what would be evident, say, at the scale of the ward. We return to the question of granularity later, and in the conclusion, we will discuss the implications of specific spatial preparedness that can anticipate the temporary increases in demand in the specific localities and to the specific PDS shops.

A second interpretation, however, is that, at least for some portion of households, the URRs are a revealed vulnerability that is not just a temporary rise in demand in response to crisis. Put simply, e-coupons measure households that are vulnerable to falling into food insecurity quite quickly, perhaps, at a much smaller trigger than a lockdown and therefore, arguably, should be enrolled into the PDS system for precisely this reason. Such enrolment would imagine the PDS system as a protective and transformative system of social protection, one capable to preventing a fallback into poverty and insecurity, rather than just a preventive one that acts as a bare safety net (Devereux and Sabates-Wheeler 2004).

Within e-coupon applicants, how many hold such vulnerability? It is worth remembering that the e-coupon scheme was started two weeks after the first lockdown was announced on 24 March 2020, and the delivery of e-coupon ration started in mid-April. The shock here is not, then, the full extent of the pandemic or the lockdowns but food insecurity that results from restrictions on income for about three to six weeks. Other studies have shown similarly that urban workers in India’s metropolitan centres began to run out of savings within three weeks (Adhikari et al 2020). Further, accessing e-coupon ration was not without its own risks. One would have to venture out without any available public transport to a distribution centre in violation of lockdown guidelines, be exposed to others and risk infection, and, no doubt, wait and handle uncertainty of delivery and supply, all in order to get basic dry-ration food supplies. Accessing e-coupon ration entitlements, in other words, had significant associated costs and risks indicating that food insecurity would have to be strongly felt by those applying.

It is also possible, however, that some may be accessing e-coupons more as a precautionary measure, unsure of how long the lockdowns would last, and not wanting to be caught unprepared. These households may be seeking to protect savings for other uses, or responding more to the barriers of physically accessing food in the lockdown rather than being unable to afford it. Another category of applicants could be of people present in Delhi during the lockdowns but not normally residents there or living in arrangements without formal proof of address, in which case, they would be ineligible for the PDS system altogether. In this case, they would be outside the system purely because of the eligibility criteria. For them, e-coupons would be a way to access entitlements usually not available to them.

It is not possible in our analysis to quantify these segments precisely to compute a finite number of applicants that, say, should be added to the PDS system. Yet the scale of the response to the e-coupons (approximately 56 lakh individuals), and the fact that more than half of the PDS shops have people near them who are, or can quickly become, vulnerable and in need of food aid, indicate strongly that the current enrolment in the PDS system is severely inadequate, especially if it is to address not just food poverty but food insecurity. Current enrolment is limited, as we argued in the beginning, at 40% of the residents under the targeted PDS scheme that Delhi follows, a constraint that seems to emerge from supply considerations than actual food insecurity. The data reminds us of the artifice of this limit, and suggests that that an expansion is essential.

Spatialising and Scaling the Unmet Demand

What should inform such an expansion of enrolment within the public distribution system? Such an expansion must take the spatiality of unmet demand seriously by working to build capacity at specific PDS shops (and their associated localities), which have greater need. This is especially important if—as is almost certainly going to be the case—the system’s expansion is incremental with limited resources that can be put to use at a time. The “where” of food insecurity must direct new investments in the system’s capacities be it for general expansion or for more limited plans of disaster preparedness. The e-coupon data offers an invaluable archive to make such a geography visible.

Yet this geography cannot be accurately understood at current scales of urban data governance. If we assess the URR at different administrative levels, this becomes apparent. Figure 1 shows the URRWardURRAssembly, and URRDistrict. The loss of granularity as we move from ward to district is stark. The district map flattens variations in URR, unable to hold the specificity of variation. Wards with a URR of up to 10 are visible clearly in the ward map, and this specificity is lost at each aggregation at the assembly constituency, and then at the district. Further, approximately 21.9% of the wards that had a URR>1 and <5 are spread across districts. If a phased expansion of the PDS system was to proceed at the district level, it would risk an inefficient and inequitable allocation of resources. In a city like Delhi, where empowered governance actors are at the district and not the ward level, this is a serious reminder of the consequences of the inadequate localisation of social protection systems “within” megacity regions, and not just between federal–state–local governance structures.

Beneath the Ward

Yet our intention is not to argue for either data or governance mechanisms at ward level alone. Scholars have shown, convincingly, that, in the megacity, the ward is itself a complex and heterogeneous scale with significant variation within it (Baud et al 2009; Bhan and Jana 2015; Bharathi et al 2021). As we dig deeper into our own data, the inadequacies of the ward as a scale to assess food security are indeed apparent. To illustrate, we take a set of four wards with the same number of total ration shops (six, to be precise) and shops with a URR>1. Figure 2 shows, within wards of roughly similar sizes, shops that have unmet demand and how they are distributed within the ward can vary greatly. For example, in Budh Vihar, both shops with the majority of unmet demand were close to each other in a specific area on the northern edge of the ward. In Kishanganj, not only do three PDS shops cluster in one part of the ward towards south-west, one of them holds unmet demand while the others do not, indicating perhaps a locality of concentrated need that is not being met even by three adjacent PDS shops. In Milap Nagar, the largest unmet need is at the western edge of the ward, indicating perhaps a spillover into a neighbouring ward that has implications for governance at this scale.

Within the ward, then, our attention must seek a different scale such as the category we introduced earlier in the article: a locality. If the unmet demand is spatially concentrated in specific localities within the ward, it may also indicate an overlap with other forms of vulnerabilities (are these localities marked by other concentrations of social or occ­upa­tional groups?), or tenure considerations (do the localities intersect with planning categories of jhuggi jhopdi clusters, urban villages, resettlement colonies, or unauthorised colonies?). Each of these could explain why unmet demand exists in that locality as well as represent factors that should shape a governance response for more effective delivery.

We do not, as we said at the start of this article, have any data on the spatial boundaries for what the Department of Food, Supplies, and Consumer Affairs calls a locality. We cannot, therefore, assess whether it overlaps with more colloquial categories like colonies, neighbourhoods, camps, or nagars. What we do know is that the category is meant to capture an immediate spatial area around a PDS shop in order to meet the principle of proximity and access. This is why, in Figure 2, the different spatial distributions of PDS shops within a ward matters; this is equally a distribution of localities.

What could the boundaries of a locality look like? What would a distribution of URR tell us at the scale of locality rather than the ward? Here, we undertake a conjectural exercise, working with the data set we have as well as its limits. Using the PDS shop data, we create a Voronoi tessellation; voronoi polygons reflect a division of a plane into areas or regions or all points closest to a given set of objects. In this case, each PDS shop with unmet demand (n=1,068) acts as an object, giving us a set of polygons that simulate what localities could look like and allow us to assess URR distribution across them. This is presented in Figure 3. It is worth remembering that this is a subset of the actual number of localities that exist in the Department of Food, Supplies, and Consumer Affairs database, where PDS shops are mapped onto close to about 3,000 localities. What we see here then does not represent the granularity of a full localities map, but certainly takes us some distance from the 272 wards in Figure 1.

Even at this first-level simulation of localities, the distinction from the URR distribution at the ward level is striking. At this scale, specific focus on the localities with extreme concentrations of unmet demand becomes possible. This could allow an incremental expansion of enrolment into the PDS system to begin from the most underserved localities first, targeting with far more efficiency below the ward, let alone the district. Further, the ability to granularly assess the reason for this concentration and its relationship with other dimensions of vulnerability such as income, identity (caste, gender, religion), built environment, or planning typology would allow not just an expansion of food security but a focused and more integrated development strategy for the locality itself. It could allow spatialising the approaches to multidimensional vulnerability.

Concluding Notes

Urban research and practice in India are hampered by the absence of adequately scaled, public, and relevant data. Household-level data are not publicly shared by the Census of India and aggregated data at the ward or district level are deeply inadequate for urban governance, especially in metropolitan regions like Delhi. Large-scale data sets like the e-coupon database then represent an opportunity for insight—even if it is necessarily exploratory and conjectural—that we have sought to leverage. In closing, we offer a set of concluding notes.

First, this data set adds to calls for a reconsideration of the NFSA’s caps on enrolment in urban areas. Indeed, the extent of the URR in one of the country’s richest urban regions lends strong support to the long-standing arguments for the universalisation of the PDS (Khera and Somanchi 2020), or at least the use of what is called “an expanded PDS” as currently used by states such as Tamil Nadu, Chhattisgarh, and Odisha to go beyond the mandates of the NFSA 2013. Dismissing the surge of applications for e-coupons to a rare event, like a pandemic, would be a misrecognition of the everyday vulnerability of urban residents (Krishna 2010), particularly post COVID-19 (Kesar et al 2021), and would run the probability of undermining the risks that even temporary food insecurity poses to human development outcomes. Discussions on reform within the PDS system have focused more on portability of benefits—“one nation, one card”—without an equal emphasis on expanding enrolment. Our analysis strongly cautions against this.

Second, however, it is imperative that our emphasis not only be on expanding enrolment but improving design and delivery of the system. We make a limited contribution to this question, drawing from our arguments about spatialising the unmet need. If the expansion of the system is to be incremental, then its phasing must be informed by the specific spatial distribution of food insecurity and by understanding the degrees of vulnerability within those currently outside the system. There will always be trade-offs in practice as the public systems incrementally expand and these must be addressed transparently. For the segmenting populations, relating e-coupon data to other governmental measures of household vulnerability (for example, occupational categories, below-poverty-line lists, social identities such as caste and religion) may precisely allow the segmentation that we spoke about earlier. Such a segmentation allows us to try and assess how many of those that applied for e-coupons during the pandemic would be likely to do so even after lesser welfare shocks such as injury, unemployment, illness, or debt for life events like marriage, death or birth. Similarly, because the e-coupons were open to “migrants,” that is, those without Delhi-based Aadhaar cards, it also offers the possibility to assess how many residents, at any given time, would be left outside the universal PDS system unless eligibility criteria were not amended.

For better spatial planning, our analysis offers a way to more precisely target phasing, that is, focusing on the localities with higher URRs and working backwards. In the case of food security, given that “locality” is already a category of internal administration within the Department of Food, Supplies, and Consumer Affairs, this is a viable framework to use in order to assess where distribution must be improved first. It would also allow a way to understand the relationship bet­ween density, built environment, and demand—asking how and where supply within a more universal system can be expanded through upgrading particular PDS shops, creating clusters of multiple shops in a particular areas (such as the concentration we found in Kishanganj earlier), or even experimenting with new institutional models for food security, including community kitchens.

Rescaling Social Protection

Third, it is critical to keep pushing beneath the ward as a scale at which to assess and administer many components of urban social protection systems. The logic of assigning a ration card to a locality is proximate access. This principle is relevant not just for PDS shops but many kinds of social infrastructure such as public sanitation facilities, primary healthcare comprising dispensaries or mohalla clinics, open spaces, or community centres. We accessed e-coupon data for 2,268 localities, and simulated locality boundaries for 1,028 PDS shops with unmet demand. There are, to our knowledge, closer to 3,000 localities to which PDS shops are assigned in Delhi. What would it mean, then, to think of the locality as the scale of diagnosing vulnerability not just for food but, for example, also for access to primary health?

Using this scale would enable us to understand not only a finer and more granular distribution of need but could enable us to layer data on other attributes that could explain why concentrations of vulnerability seem to exist in particular localities. For example, does a locality’s URR correlate to particular planning typologies like jhuggi jhopdi clusters, urban villages, resettlement colonies? Does it follow a geography of migrant housing and intra-city mobility? Does URR cluster around industrial areas? Does it overlap with social geographies of caste, for example, shown in recent work to be more finely and fractally distributed in our cities than previously imagined (Bharathi et al 2021)? How do these material, tenurial, and social structures shape food insecurity? Such a layering is important so that an explanation of food insecurity is not simply reduced to income poverty but rather read alongside a multidimensional vulnerability that is not be limited to the “slum,” on the one hand, or an inadequate aggregation at ward level, on the other. The spread of e-coupon data across the city and the sheer number of applications, both underline the fact that existing vulnerability is greater and more complex than our current assessments, and suggests the need for continuing research on patterns of intra-city distribution of different forms of vulnerability.

For praxis, such a layering is key to the design of delivery mechanisms that will, for example, be different in an unrecognised jhuggi-jhopdi cluster than in a resettlement colony; in mixed income localities than those more homogeneous; in areas of concentrations of “migrant” workers, or particular caste or religious geographies; in denser versus more sprawled layouts; in localities with dominant occupational or employment types; and in those located in central city areas versus those more peripheral. Such a layering could—echoing an older conceptual debate on the idea of “locality” within urban geography (Massey 1993)—allow the design of social protection systems to both appreciate the particularity of a locality as well as the need to relate it to other localities along particular characteristics—a system for industrial clusters, for example, that looks different from the one for resettlement colonies or urban villages. Universal access must be aided by localised and particular delivery mechanisms. Wards cannot hold these characteristics, but localities perhaps could.

Fourth, we reiterate the necessity of more publicly available data at the household level that would allow an assessment of distributions not just at the scale of the locality but other possibilities too. The vulnerability of households scales incrementally—to their street or gali, their mohalla or neighbourhood, their locality, and then to the city and beyond. Scale, as geographer Neil Smith reminded us, is produced relationally and dynamically (Jones III et al 2017). The complexity of megacity governance insists on data infrastructures that are equally relational and dynamic, and take us far beneath the ward or district. Doing so need not, as Massey reminded us, risk a hyper-localism but, in fact, could enable us precisely to bridge the universal and the particular.

We have used an archive of food insecurity in this article. Arguably, beyond just food, our assessments apply equally to social protection entitlements of diff­erent kinds. As COVID-19 has once again reminded us, social protection systems must be rooted in an assessment of vulnerability and not just poverty so that they may be promotive and transformative and not just bare safety nets against destitution. Doing so requires not just a moral, ethical, and political commitment to equity but appropriate data infrastructures and delivery mechanisms. E-coupon data offers learning for both and it is imperative that we heed its insights so that post-COVID-19 urban futures may hold the possibility of breaking historically entrenched patterns of urban inequality in our cities.

Notes

1 See, as an example of the extensive coverage, Dhillon (2020).

2 As of 14 May 2021, the portal could still be accessed here: ration.jantasamvaad.org. Aadhaar cards were mandatory for the scheme in order to check that applicants did not have a ration card. However, the Aadhaar card could be from anywhere in India—a Delhi address was not mandated, making the programme open to migrants.

3 Overall 17.8 lakh applications for e-coupons were received on the Delhi government’s web portal. They covered between 60 lakh and 68 lakh individuals, depending on data cited in different sources. The Delhi Economic Survey 2020–21 mentioned the figure of 69.60 individuals but affidavits filed by the Government of Delhi in Delhi, Rozi, Roti, Adhikar Abhiyan v Union of India and Ors (WPC 2161 of 2017) mentioned 60 lakh individuals. Copy available on file. In this article, we use the 60 lakh figure to err on the conservative side of measuring unmet demand.

4 Same as note 3.

5 Scraped from Delhi State NFSA portal (https://nfs.delhi.gov.in/Citizen/Householdtobeincluded.aspx) on 19 August 2020 by Saloni Bhogale.

6 We derive the multiplier for ration card by dividing the total number of ration cardholders as maintained by the government by the total number of households scraped from the Delhi government NFSA portal.

7 Performed on R Statistical language using st_nn function from nngeo package. On application of a collection of points in space, nearest neighbor algorithm finds for each point its closest neighbor based on some attribute. In our case, we used straight line distance to find the closest neighbour.

8 This model has certain data and computation related limitations. Nearest shop as identified by the nearest neighbour algorithm might not always be a shop which is closest to approach while walking (that is natural and artificial barriers can distort this estimation). Further, if a shop has no e-coupons that have been allocated, it does not necessarily reflect that there is no unmet demand in its proximity. It simply tells that another shop was closer to the points of demand in the locality. Also, while matching to the closest shop we have not limited the matching to only shops within existing administrative boundaries (ward, constituency, or district). In present practice on field, it is often the case that beneficiaries are allotted ration shops which are not closest in absolute terms, but are closest within the administrative boundary of unit of which they are the residents. Lastly, we have used consistent multipliers for estimating the number of people linked to both individual ration cards as well as e-coupons. This limits the observation of variation in household sizes though across the state, the total counts remain the same.

References

Aajeevika Bureau (2020): “Unlocking the Urban: Reimagining Migrant Lives in Cities post COVID-19,” Ahmedabad, https://www.aajeevika.org/assets/pdfs/Unlocking the Urban.pdf.

Adhikari, Anindita, Navmee Goregaonkar, Rajendran Narayanan, Nishant Panicker, and Nithya Ramamoorthy (2020): “Manufactured Maladies: Lives and Livelihoods of Migrant Workers During COVID-19 Lockdown in India,” Indian Journal of Labour Economics, Vol 63, No 4, pp 969–97.

Baud, Isa S A, Karin Pfeffer, Namperumal Sridharan and Navtej Nainan (2009): “Matching Deprivation Mapping to Urban Governance in Three Indian Mega-cities,” Habitat International, Vol 33, No 4, pp 365–77.

Bhan, Gautam, Antara Chowdhury, Neha Margosa, Kinjal Sampat and N Sohane (2020): “Lessons for Social Protection from the COVID-19 Lockdowns Report 1 of 2: State Relief,” Indian Institute for Human Settlements, Bengaluru, https://doi.org/https://doi.org/10.24943/LSPCL11.2020.

Bhan, Gautam and Arindam Jana (2015): “Reading Spatial Inequality in Urban India,” Economic & Political Weekly, Vol 50, No 22, pp 49–54.

Bharathi, Naveen, Deepak Malghan, Sumit Mishra and Andaleeb Rahman (2021): “Fractal Urbanism: City Size and Residential Segregation in India,” World Development, Vol 141, Article 105397.

Chakraborty, Shiney (2020): “COVID-19 and Women Informal Sector Workers in India,” Economic & Political Weekly, Vol 55, No 35, p 17.

Deshpande, Ashwini and Rajesh Ramachandran (2020): “Is COVID-19 ‘The Great Leveler’? The Critical Role of Social Identity in Lockdown-induced Job Losses,” GLO Discussion Paper, No 622.

Devereux, Stephen and Rachel Sabates-Wheeler (2004): “Transformative Social Protection,” IDS Working Paper 232, IDS, Brighton.

Dhillon, A (2020): “Delhi Divided: If Coronavirus Doesn’t Kill me, Hunger Will,” Guardian, viewed on 24 May 2021, https://www.theguardian.com/global-development/2020/mar/30/divided-delhi-under-lockdown-if-coronavirus-doesnt-kill-me-hunger-will.

Jones III, John Paul, Helga Leitner, Sallie A Marston and Eric Sheppard (2017): “Neil Smith’s Scale,” Antipode, Vol 49, pp 138–52.

Kesar, Surbhi, Rosa Abraham, Rahul Lahoti, Paaritosh Nath and Amit Basole (2021): “Pandemic, Informality, and Vulnerability: Impact of COVID-19 on Livelihoods in India,” Canadian Journal of Development Studies, Vol 42, Nos 1–2, pp 145–64.

Khera, Reetika and Anmol Somanchi (2020): “A Review of the Coverage of PDS,” Ideas for India, 19.

Krishna, Anirudh (2010): One Illness Away: Why People Become Poor and How They Escape Poverty, Oxford: OUP.

Lahoti, Rahul, Amit Basole, Rosa Abraham, Surbhi Kesar and Paaritosh Nath (2020): “Hunger Grows as India’s Lockdown Kills Jobs,” India Forum, Vol 5.

Massey, Doreen (1993): “Questions of Locality,” Geography, Vol 78, No 2, pp 142–49.

Mishra, Khushbu and Jeevant Rampal (2020): “The COVID-19 Pandemic and Food Insecurity: A Viewpoint on India,” World Development, Vol 135, Article 105068.

Rai Chowdhury, Antara, Gautam Bhan and Kinjal Sampat (2020): “Impact of Covid-19 and Lockdowns on the Domestic Workers: First Report,” Bengaluru, https://doi.org/10.24943/ICLD­­W­10­.2020.

Rajpal, Sunil, Julie Kim, William Joe, Rockli Kim and S V Subramanian (2021): “Small Area Variation in Child Undernutrition across 640 Districts and 543 Parliamentary Constituencies in India,” Scientific Reports 11, No 1, pp 1–9.

Singh, Arkaja (2020): “Hunger, COVID-19, and the Indian Administrative State,” State Capacity Initiative Working Paper No 2020, New Delhi.

 

Dear Reader,

To continue reading, become a subscriber.

Explore our attractive subscription offers.

Click here

Updated On : 17th Jul, 2022
Back to Top