All in it Together? Community Food Aid in a Multi-Ethnic Context

Abstract This paper derives from a study of community food aid in a multi-ethnic, multi-faith city in the North of England. The paper begins to make sense of the diversity of types of food insecurity assistance, examines the potential exclusion of certain groups from receipt of food aid, and explores the relationship between food aid providers and the state. Faith-based food aid is common in the case study area, particularly among food bank provision to the most ‘destitute’ clients. While food aid is adopting service responsibilities previously borne by the state, this does not imply an extension of the ‘shadow state’. Rather, it appears reflective of a pre-welfare state system of food distribution, supported by religious institutions and individual/business philanthropy, but adapted to be consistent with elements of the ‘Big Society’ narrative. Most faith-based providers are Christian. There is little Muslim provision of (or utilisation of) food aid, despite the local demographic context. This raises concerns as to the unintentional exclusion of ethnic and religious groups, which we discuss in the concluding sections.


Introduction
Food insecurity in the United Kingdom (UK) has apparently increased dramatically since 2008. Local communities across the UK have responded quickly, developing strategies to address existing food insecurity and prevent future increases (Lambie-Mumford, 2013). Between April 2015 and 2016 over one million people were given at least three days' worth of emergency food supplies from the Trussell Trust's 1 UK network of foodbanks, a forty fold increase on 2008/09 (Trussell Trust, 2016). The dynamics of changing demand and the speed of the response has outpaced academic analyses.
Parallel with this increase in food insecurity, financial support for people who are unemployed or on a low-income has been restricted. The 2010-2015 Coalition government's implementation of a system of lower and less-inclusive social security entitlements (Perry et al., 2014), continued by the current Conservative government, is part of a wider public policy shift involving a reduction in public sector spending and promotion of civil society responses to social problems (Lambie-Mumford, 2013).
This paper uses data from a multi-ethnic, multi-faith city in the north of England (Bradford) to better understand what constitutes community food aid. There are three objectives. First, to understand whether community food aid is a reaction to a recent rise in food insecurity or the re-naming of longstanding charitable provision for the destitute. Second, to interrogate the sector's development in relation to public policy shifts, questioning the extent to which food aid is a component of a 'shadow state'. Third, to explore how community food aid meets the needs of a multi-ethnic, multi-faith population.
Developments in Bradford are considered within the broader political landscape of emergency food provision and community food security initiatives in the UK and internationally. The paper seeks to provide some clarity in understanding UK food aid and, without negating the generosity of many of those involved in its provision, further a critical perspective on emergency and non-emergency food-related assistance.

Background
Public policy context Changes in welfare and service provision since 1997 have altered the relationship between state and civil society and influenced the shape and responsibilities of Third Sector Organisations (TSOs). New Labour's public services agenda exerted pressure on TSOs to deliver services (Milbourne and Cushman, 2015), promoting volunteerism as an essential part of this strategy (Levitas, 2012). As part of a broader emphasis on organisational collaboration and 'new localism', UK public agencies were increasingly required to collaborate with non-state providers to deliver welfare services. The Coalition government's 'Big Society' discourse also placed a strong emphasis on localised, non-state solutions in public service delivery. However the Coalition (and subsequent Conservative) government saw government as 'crowding-out' community action at the local level. Achieving 'social justice' under the Coalition required a smaller state and a greater role for communities (Nicholls and Teasdale, forthcoming). However discourse and theory did not map neatly onto practice. The 'typical' charity experienced a 13 per cent decline in real income between 2008-13, with those providing social services (such as food-aid) in deprived local areas amongst the hardest hit (Clifford, 2016). Although faith-based organisations also experienced public policy pressures to deliver services under New Labour, in comparison with TSOs they were more reticent in formally engaging with the state and delivering services (Rochester and Torry, 2010). It would appear that the Coalition government's Big Society agenda may have opened up a space for faith-based organisations in non-formal approaches to public service delivery.
The structure of some TSOs has transformed in response to policy reforms. Nevertheless, genuine localism that is responsive to local needs, organised communally and developed organically at the local level (Caraher and Dowler, 2014), has been damaged by competitive contracts and the imposition of national performance frameworks (Milbourne and Cushman, 2015). Restricted services and redundancies consequent upon local authority funding cuts have jeopardised the capacity of the local authorities to ameliorate health inequalities (MacLeavy, 2011) and, more generally, has inhibited local government from engaging with the localities it serves. A continuing tension between central and local government, evident in the Third Way and Big Society, has seen central government acting as if the local tier is inhibiting localism rather than approaching local government as a route to more effective community mobilisations.

Community food aid
Community responses to poverty and hunger have long existed in the UK (McGlone et al., 1999;Caraher and Cowburn, 2004). However, rising activity and the growing media profile of the Trussell Trust foodbank network, and the recent appearance of the term 'food aid' within the UK context, fosters the impression that the provision of food assistance to help people access free or subsidised food is new . The difference between the situation today and previous charitable and local community support is in the increased level of demand for emergency and ongoing food from people with nowhere else to turn (Dowler, 2014) and in the growing scale and logistical difficulty of providing food aid via food banks (Lambie-Mumford, 2013).
There is little clarity as to what constitutes food aid and how the rapid growth of some organisational models -notably the food bank -has impacted on prevailing terminology. The UK's Department for Environment, Food and Rural Affairs (Defra) defines food aid as 'an umbrella term encompassing a range of large-scale and small local activities aiming to help people meet food needs, often on a short-term basis during crisis or immediate difficulty . . . relieving symptoms of household or individual level food insecurity and poverty' (Lambie-Mumford et al., 2014: 15). The use of this food aid terminology in recent publications by NGOs and charities (Cooper and Dumpleton, 2013;Sosenko et al., 2013) is not uniform. However there is an identifiable tendency towards a broad conception of food aid, which encompasses emergency food assistance, such as food banks or 'foodbanks' (the latter denotes Trussel Trust foodbanks), soup kitchens and soup runs; and non-emergency provision which includes day centres and 'drop-in' centres, community cafes, and charities that redistribute food from food retailers and wholesalers that would otherwise be thrown away (intercepted food). Non-emergency provision also includes community kitchens, community supermarkets and food co-operatives, and community gardens (Kirkpatrick and Tarasuk, 2009). A table of definitions of varying models of food aid is set out in Appendix 1. The diverse, informal and fluid nature of food aid providers, rapidly developing in response to changing demand, obscures the exact scale of food aid provision and the dynamics of its demand (Sosenko et al., 2013). Nevertheless, emergent evidence on UK food aid has raised questions about its inclusivity and institutionalisation (Lambie-Mumford and Dowler, 2014a;Garthwaite et al., 2015), issues which have been discussed extensively in analyses of food aid in North America (Poppendieck, 1998;Riches, 2011).

Inclusivity
Unlike some forms of emergency food aid -notably the food bank -which formally restrict who can and cannot receive food by using a referral system , few forms of non-emergency provision directly exclude people and many aim for inclusivity (Donald and Blay-Palmer, 2006) or specifically target any low-income people (Caraher and Dowler, 2014). However organisations providing food aid are often in precarious financial positions and, in adapting to funding streams that are available to them, an organisation may have to change its priorities and focus on an area given priority by those distributing finance. One result is that organisations can focus on individual nutrition at the cost of their attention to the structural determinants of food behaviour (Caraher and Dowler, 2014) and, in doing so, inadvertently alienate families living in the most severe poverty (Tarasuk, 2001).
Critical analyses of the food security and the sustainable farming elements of US community food provision identify unacknowledged racism in the sector and suggest that racial exclusion is highly problematic (Alkon and McCullen, 2011). Whites are over-represented among the staff, leadership and users of the sector, and community food organisations have been criticised for adopting colourblind mentalities (Guthman, 2008) and essentialising discourses, for promoting 'white' notions of healthful food and bodies, and for extolling the virtues of community and self-sufficiency in a way that obscures the 'racist, classist and gendered features of the food system' (Slocum, 2006: 330).
The high level of faith-based organisations involved in UK food aid (Sosenko et al., 2013;Lambie-Mumford et al., 2014) raises further questions about the accessibility of provision, however the possible exclusionary consequences of faith-based provision have not been fully explored in the UK.

A 'shadow state'
Emergent evidence suggests that UK developments in food aid echo wider trends in the structure and responsibilities of the UK voluntary sector and may reflect the history of food aid in North America, in which it has increasingly become a 'shadow state' (Mitchell, 2001, Wakefield et al., 2013, that is: a 'parastate apparatus with collective service responsibilities previously shouldered by the public sector, administered outside traditional democratic politics but yet controlled in both formal and informal ways by the state' (Wolch, 1989: 201). As a component of the 'shadow state' in North America the emergency food system has inadvertently facilitated welfare state retrenchment and become institutionalised (see Poppendieck, 1998;Riches, 2002;Tarasuk et al., 2014, among many). However, while UK food aid may be adopting service responsibilities previously borne by the state, in the context of retrenched and limited welfare provision and restricted scope for local government (Lambie-Mumford, 2014), the extent to which it can be described as a 'shadow state' -formally and informally controlled and funded by the state -remains unclear.

Study site and population
The study was undertaken in Bradford (a city in the north of England) with individuals who have experience with anti-hunger/food security programmes and/or food policy.
Bradford District is the sixth largest city in the UK (in terms of population) (Gill, 2015) and has the largest proportion of people of Pakistani ethnic origin (20.3 per cent) in England, which contributes to its large Muslim population (24.7 per cent).

Methods
The study research questions were: • In the context of recent public policy shifts, what community food aid operates in Bradford, and to what extent is it a component of a 'shadow state'? • Is community food aid in Bradford a reaction to a rise in food insecurity or a re-naming of long-standing charitable provision for the destitute as 'food aid'? • How does community food aid interact with the multi-ethnic, multi-faith context?
Desk-based research, using the Internet and local government resources and dialogue with key informants in Bradford's food security programmes, was used to identify community-based responses to food insecurity in the Bradford District. Following this there were two separate phases of qualitative research.
Phase one, conducted in June 2015, consisted of two focus groups and one interview with individuals (n=9) who had experience with food security programmes/policy at a governance level. Ethical consent was obtained from the University of York Department of Health Sciences Research Governance Committee (HSRGC) (Ref HSRGC/2015/98A). A sampling frame was then drawn up by the authors in conjunction with senior members of Bradford District Metropolitan Council Public Health team to include individuals who had experience with food security programmes/policy. These included councillors in Bradford; members of the Public Health team; members of NHS services in Bradford addressing food/health: nutritionists, dieticians and members of local Clinical Commissioning Groups; and TSOs with experience of food-related coordination/policy. Forty people were identified as meeting the inclusion criteria. There was an element of subjectivity in the final (purposive) sample. A number of people did not reply to our invitations to join the study, and others declined to be involved. Reasons given included inability to attend at the time of the focus groups and a perceived lack of their having relevant experience. Those who did not reply or declined to contribute were more likely to work in the NHS or TSOs. Appendix 2 provides focus group and interview details.
Focus groups and interviews were semi-structured and conducted by a White British female interviewer (the first author) and a White British male member of the Bradford Council Public Health team; the data were recorded on a Dictaphone and transcribed. The topic guide was informed by a literature review, discussion with the project's supervisors, and consultation with senior members of the Public Health team.
In phase two, carried out from September to November 2015, 18 interviews were conducted in Bradford involving individuals from TSOs with experience of food security programmes at a community level (n=18). Ethical consent was obtained from HSRGC (Ref HSRGC/2015/712). Sample organisations were chosen purposively from the 67 food aid organisations identified in the deskbased analysis to form a representative sample, which included various types of organisation and multiple religions. The faith-based organisations in the sample were all Christian or Muslim. Interviewees within the sample organisations were also chosen purposefully to capture perspectives that would best represent the organisations' positions. Publically available information was used to draw up a contact list. Invitations to join the study were sent to the appropriate person within an organisation. Organisations which failed to respond were removed from the sample. Others declined to be involved or suggested another organisation in their place. If this occurred, the organisation was contacted only if it was considered an appropriate replacement. Reasons given for declining to participate included a perception of no relevant experience and failure to see the study's value. The final sample was biased towards emergency food aid providers but this was not considered problematic given the current salience of the issue and the high prevalence of deprivation in Bradford (Wright et al., 2013). Appendix 3 provides details of the interview organisations/participants.
Interviews were semi-structured. The topic guide was informed by a literature review, discussions within the project team and themes that arose from phase one. The interviews were recorded on a Dictaphone and transcribed verbatim.
The three-stage analysis approach was used to analyse the phase one and two transcripts (Dwyer, 2002). Each transcript was summarised to understand the narrative. Thematic analysis was used; a coding frame was devised based upon common themes/sub-themes and, using Nvivo 10, this was applied to each transcript. Relevant text was indexed whenever a theme appeared. The appropriately indexed material was transferred to a grid with basic organisational and demographic details about the sample.
To preserve the anonymity of participants and, as required by the Ethics Committee approval, details about the organisations and individuals in the sample are kept to a minimum.

Results
The nature of provision will first be described, followed by the characteristics of food aid users, and the religious dimension of community food aid in Bradford. The coherence of the community food movement and its role in relation to the state will then be reported, drawing upon data from phases one and two.

Community food aid
Desk-based research and dialogue with stakeholders identified 67 community organisations working to alleviate food insecurity in Bradford. This includes service organisations providing food onsite and/or to take away to lowincome, high-needs 'clients' 2 (traditionally know as a 'soup kitchen' or 'soup run'); community centres offering low-cost, healthy food; and environmentallyoriented organisations aiming to empower people to grow their own food. Table 1 orders the organisations according to the need they addressed. Service organisations providing food onsite or to take away for low-income, high-needs 'clients' are the most common.
Interviews with a representative sample and focus groups with key stakeholders allowed for a detailed understanding of community food aid. Study participants are set out in Table 2.  Activities 'Soup kitchens' and 'soup runs' provided free food to be eaten immediately, while organisations describing themselves as food banks provide a parcel of cold, predominantly dry food to be prepared at home. Only one food bank provided fresh food, including milk, fruit and meat. Many organisations performed multiple food activities, such as 'soup kitchens' also providing cold, unprepared food for 'clients' to take away.
Among emergency food providers, there were two distinct approaches to the distribution of food. One was formal, professional and objective: rules of who could receive food and how it could be received were defined and ostensibly were inflexible. The other was informal: policies concerning the receipt of food were flexible and subjective, and the relationship between the provider and 'client' was personal, rather than clinical/transactional. However, in practice, the formal approach did not apply: there was flexibility in the operational arrangements of all types of organisation. While all bar one of the food banks in this study operated a voucher system -a practice informed by the Trussell Trust foodbank model -all food banks would issue a food parcel to a 'client' without a voucher, if the 'client' was considered 'in need'.
Every time you come you have to have a voucher -although if someone does come without a voucher and I think they need food, I will just issue one. Food bank, Manager, P12 Food parcels, hot meals, gardening tips, and food skills were provided in the context of non-food activities, including job and debt advice, a hairdresser and drug and alcohol support services. Organisations providing only food or food skills were exceptions.

Funding
Study organisations relied on a combination of in-kind and monetary resources from private, public and third-sector sources, as well as from individuals. Only one organisation was fully funded by the state, through local authority (LA) grants. A minority were part-funded by the state, for instance through subsidised LA buildings or small grants for specific elements of their food work. This minority included emergency and non-emergency food aid, none of which were controlled by or accountable to the LA. The majority of food aid in the study sample was funded primarily by churches or mosques and a further minority were funded either by donations from both religious and nonreligious sources, or received in-kind and monetary resources from non-religious actors only.

Who uses community food aid provision?
'Client' demography varied within and between projects according to the model of provision. Onsite food providers tended to mostly serve White British men between the ages of 30 and 50, who were vulnerably housed or homeless. Many of this group had a history of alcohol and substance abuse, and mental ill-health.
It was reported that the 'clients' of onsite food provision were becoming more diverse as demand for assistance with food insecurity from the wider community increased. Providers reported that they were increasingly seeing people in parttime, low-income work, and people with children. Organisations attempted to accommodate children but also expressed concern that serving vulnerable adults and children in the same building posed safeguarding issues: Food bank 'clients' were more likely to be in housing, although this could be insecure. The majority were White British people in receipt of social security, with children. While most 'clients' were experiencing an acute financial crisis and required immediate assistance with food, some visited the food bank out of loneliness, while others could not afford a balanced diet and visited the food bank to improve the quality of their food: The 'clients' of non-emergency food aid were more diverse than those of onsite food providers and food banks. These included people who were homeless or vulnerably housed and people who were financially secure in equal proportion.
All types of organisations explained that meeting competing emotional, financial and health needs could pose challenges: It's about the food, but it's also not about the food for us because it's a person in front of you, and these people are complex, multiple levels of brokenness and dysfunctionality (sic). Onsite food provider, Manager, P5 Projects attempted to overcome these tensions by providing additional services such as food bank referral services or, alternatively, encouraging community kitchen 'clients' to use their cooking skills to gain employment or voluntary experience.

Food aid and ethnic diversity in a multi-ethnic context
Despite Bradford's ethnic diversity, there was limited ethnic diversity among 'clients' at food aid providers (see Figure 1 for the geographical distribution of Bradford's Pakistani population, in relation to food aid providers). Emergency food providers, in particular, served very few Pakistani 'clients' and Christian food banks had a less diverse client base than those that were Muslim or secular. The one Muslim food bank interviewed had a more ethnically diverse client base than its Christian counterparts, with a large minority of Pakistani female 'clients'. However, this was partly associated with the food bank's main referral partner (a domestic violence refuge). The 'client' base at food banks and soup kitchens was predominantly White British, although staff interviewed described Muslim (n=7). This is unreflective of Bradford's religious demographic: nearly one quarter of the population (24.7 per cent) identify as Muslim and just over one fifth of the district's population (20.7 per cent) describe themselves as having no religion (see Figure 2 for the geographical distribution of Bradford's Muslim population, in relation to faith-based food aid).
In Bradford, organisations providing emergency food -'soup kitchens' and 'soup runs', and food banks -were more likely to have their origins in religious charities than those providing non-emergency food -community cafes, community kitchens, community supermarkets and environmental organisations -which tended to be secular. Of the 19 soup kitchens and soup runs in Bradford, 11 were Christian (Catholic, Methodist, Anglican), four considered themselves to be Muslim and four were secular; six of the eight food banks were Christian, one was secular and one Muslim. In the study sample, five of the six food banks were Christian, three of the nine soup kitchens were secular, one was Muslim and the rest were Christian. In total, 12 of the 20 food aid providers in the sample were faith-based. The scope of this paper precludes full discussion of what it means to be a faith-based organisation providing food, the work of Sider and Unruh (2004) and Jeavons (1997) provides detailed typologies, upon which this paper draws, of faith-based social service and educational organisations and their characteristics.

The expression of faith in community food aid
Christian projects operating 'soup kitchens' were mainly part of longstanding Christian charitable organisations. The provision of food fitted within a tradition of charitable services for 'the poor' and, among many, included a commitment to evangelising: 'clients' were informed they were in 'God's house'; they were required to pray or listen to a sermon before eating; Bible classes were offered and religious music was played. The charitable response to hunger was considered to be a form of prayer: Interviewer: What motivates you all here to do it? It is religious for sure. We don't see it so much as a social work as much as like a response to a spiritual call, just our faith, and serving the Lord and the poor.

Onsite food provider (Christian), Manager, P16
Food banks were also mainly affiliated to Christian charities, but were less likely to have evangelical roots and, while the charity could be a few hundred years in age, no Christian food bank interviewed was more than four years old. The relationship with the Christian charity to which the food bank was affiliated tended to be formal and detached; and faith, if expressed at all, was subtle and unforced. The one Muslim food bank interviewed operated from a reception room in a mosque. The building was located in an industrial park and neither the exterior nor the reception room gave any indication that it was a religious building.

Faith-based motivation
Among all Christian food aid providers interviewed, faith was an important motivation for provision. Food aid was a means of expressing Christian charity: I'm learning and beginning to understand grace. By operating in this way we also encourage others to do so.

Social Food Charity (Christian), Manager, P7
Faith-based motivation was less discussed by staff in Muslim charities. Staff spoke of motivation stemming primarily from unmet need, rather than Islamic teachings: Now in my case when I joined (this organisation) I was not a practicing Muslim. I am now, but that had nothing to do with faith when I started. I wanted to serve the community, give back to the community.

Food bank (Muslim), Manager, P2
The high-level religious involvement in community food aid provision was attributable both to the availability of resources and volunteer capacity in religious charities and to religious teachings. Christian and Muslim charities provided sustained finance, buildings and volunteers in a climate of limited, short-term, third-sector funding.

A joined up system or a disparate shadow state?
Study participants were aware of the importance of partnership work and many described collaboration with private-sector organisations, and with food and non-food TSOs. Food banks were most likely to describe links with local and national retailers. Public-sector involvement extended no further than the provision of buildings by the local authority; indeed there was apparent hostility to top-down management from Councillors in their 'ivory towers'. Many organisations providing emergency food had formal or informal links with third-sector social services. Food banks which adopted the Trussell Trust voucher system were reliant on the quantity and quality of their relationships with referral agencies in order to receive 'clients' and distribute food and, for many services working with low-income, high-needs 'clients', collaboration with social services was fundamental to the achievement of their long-term goals. 'Success' involved changing long-term outcomes by tackling 'clients' -their physical and mental ill-health, debt, long-term unemployment and, for some, insecure housing.
Collaboration also occurred between organisations within the community food sector, such as directing 'clients' to emergency food services that could help them throughout the week or in the evening. Partnerships formed around the co-ordination and distribution of intercepted food could be with nation-wide organisations, such as FareShare, or could operate at a local level through personal relationships with retailers. Christian organisations providing emergency food were particularly well-networked. Although a shared faith was an important bond, the basis of partnerships tended to be pragmatic.
The reliability and sustainability of partnerships was questionable. Study participants were aware of the importance of partnership work, but there was a lack of formal co-ordination. Food organisations tended to have little knowledge of projects, activities and issues unrelated to their immediate remit and, while there was some awareness of city-wide coordinating bodies, involvement in these formal networks was limited.

Self-identity and the politics of power
Concern was expressed that community initiatives should not replace the welfare state. Stakeholders questioned whose responsibility food insecurity should be, civil society or the state:

I am aware of criticisms of food banks, that they let the government have a lot more breathing room. Whose responsibility is it to feed the poor? Food bank, Manager, P12
These concerns were tied to questions of social identity. The citizen involved in food aid was characterised as the 'Good Samaritan' and the provision itself was reliant on the 'willingness' of volunteers. This characterisation was associated with the self-identity of some of the stakeholders. It became clear in the latter part of the focus groups that some of the participants perceived themselves not only as 'good citizens' but as different from their 'clients', because of their comparative affluence and superior 'relationship' to food: Rather than dancing (around the issue), we have to deal with it and say, 'the truth is, the people who do fall down we need to provide for them, but there is a whole bunch of people above that who just have this bad relationship with food'.

Food aid coordinator, FG1
None of the soup kitchens or soup runs interviewed for this study excluded 'clients' from receiving food on the basis of deservingness. However, the inclusion of 'clients' could be contingent on their behaviour in the venue, on their acquiescence to religious preaching, or on their responsiveness to staff questions about their circumstances: We always ask for that, especially with the Poles -'have you got your papers?' if it is someone new, because if they are coming down for food, why are they coming down for food? They are either coming down for food because they are not working, or they are coming down because they are suffering because their documents are gone and they are not able to work, because without the relevant documents these days it is difficult. Onsite food provider,Manager,P10 In this study, a majority of food banks used a voucher system, partnering with statutory and third-sector social services. Staff preferred that an expert decided whether or not the client should receive food. Nevertheless, many food banks served 'clients' who turned up without a voucher, if the client was deemed to be in sufficient 'need'. Among food banks who described doing this, none had a set criterion of need; concepts of 'deservingness' could be informed by opinions, prejudice, and religious beliefs: We had this agency sending us people from Poland, and then we had to say, 'Enough is enough' because they used to come up here totally pissed (intoxicated) out of their brains, and collect a food parcel. So I put a stop to it, I said, 'hang on a minute if the guy can afford to drink then does he deserve a food parcel?' Food bank, Manager, P2

Discussion
In 2002, an analysis of food banks and food security in Canada concluded, 'Food banks are confirmation of the re-emergence of the residual welfare state and sit at the interface between critical questions of public health, welfare reform and social policy' (Riches, 2002: 661). This paper underlines not only the extent to which these critical questions straddle multiple policy domains but also the complexity and nuance demanded in academic and policy debates on food aid and food security. This is illustrated by three themes, the multifaceted and fluid character of UK food aid; the ambiguous relationship of food aid to the state and the extent to which it can be considered a 'new' reaction to recent policy developments; and the unintentional exclusion of ethnic and religious minorities from some faith-based food providers. In this section we discuss each of these themes in turn.

Community food aid: multiple models and terminology
This paper identified two types of food insecurity assistance, emergency and non-emergency, and seven types of organisations undertaking food securityrelated work: service organisations for low-income, high-needs 'clients' providing free prepared food ('soup kitchens' and 'soup runs'); service organisations for low-income, high-needs 'clients' providing free food parcels ('food banks'); community centres offering multiple food activities, such as low-cost food to consume onsite, cooking classes and horticulture; social food charities using innovative methods to provide low-cost or free food for communal consumption; environmental organisations offering horticultural training and low-cost fresh food; community supermarkets or co-operatives/box schemes; and food recovery organisations. Most study organisations operated at least one food activitypossibly simultaneously offering emergency and non-emergency provisionand the majority also included a non-food-related activity. In line with other studies (Sosenko et al., 2013;Dowler and Lambie-Mumford, 2014), this paper found that, while the majority of provision is run by charities, projects can vary considerably in their activities and size, in their funding, staffing and commercial arrangements and in their aims. The division of the community food sector into two movements, community food security and the anti-hunger movement, suggested by previous literature (Allen, 1999), was not evident in this case. Nonemergency providers -environmental organisations, community centres offering cooking classes and social food charities -did ascribe to a community-based and prevention-orientated framework that focused on immediate and longterm needs. However these organisations also had sympathy for the concerns of the anti-hunger movement: the eradication of destitution. Similarly, emergency food providers addressed the long-term needs of their 'clients' with, for example, cooking classes, or debt and employment advice.

Food aid and the state
Food aid in the case-study area could not strictly be described as a 'shadow state'. Only one study organisation fitted the definition of a 'shadow state' organisation: a TSO undertaking service responsibilities previously shouldered by the public sector, fully funded and informally controlled by the state through service contracts (Wolch, 1989). This organisation was distinct from other study organisations: it was professional and detached in its engagement with clients, accountable to the state and financially sustainable. There were no systematic differences in operations or organisational autonomy among organisations partfunded by the state and those with no state support. While the voucher system, implemented by food banks, could be seen to open up the possibility of state accountability frameworks, emergency food providers, not the state, retained full control over both the process of issuing vouchers and receiving 'clients' and the collection and analysis of 'client' data.
Food aid primarily reflected a pre-welfare system (Harris, 2004) of food aid, predominantly funded by religious institutions and/or secular philanthropy. Staff, unrestricted by accountability frameworks, could distribute assistance according to 'deservingness' or adherence to standards of behaviour. In her analysis of food in Finland, Silvasti (2015) similarly found that 'the principles and practices of giving aid can, and do, vary widely between charitable organisations. Clients cannot presume they will be treated as equals or equally' (Silvasti, 2015: 478). In this study, food aid continued a Christian history of charitable services for the 'poor' (Wakefield et al., 2013) and was a physical expression of Muslim principles of charity: Zakat (the Islamic doctrine of alms giving for practicing Muslims, which is obligatory for adults who are financially able) and Sadahaq (voluntary giving for practising Muslims) (see Montagné-Villette et al., 2011).
In line with this, only a minority of the organisations interviewed were founded post-2008. Newer organisations were more likely to be food banks and social food charities, as suggested by Lambie-Mumford et al. (2014) and Lambie-Mumford and Dowler (2014a). Some older organisations had recently expanded their food activities into emergency food provision, while some longstanding emergency food providers discussed serving a new type of in-work client. Nevertheless, community food-assistance for people in need was longestablished in the case study area and even organisations that appeared new tended to be affiliated to older organisations, often churches, with a long history of charitable work. As previously suggested (Dowler, 2014), the 'new' element of community food aid relates to the people served. All types of organisation were assisting a new type of client: people whose needs were urgent, who felt desperate and who had nowhere else to turn.

The exclusionary implications of faith-based food aid
The high prevalence of faith-based food aid in the case-study area (48 per cent) is in keeping with previous literature (Sosenko et al., 2013;Lambie-Mumford et al., 2014). The majority of faith-based food aid was Christian, and food banks were more likely to be faith-based than any other model, although soup kitchens were also predominantly faith-based. Muslim food aid was in the minority, despite the local context, and non-emergency food aid was largely secular.
Perhaps surprisingly given Bradford's ethnic diversity, 'clients' of food aid providers were predominantly 'White'. In particular, Christian food banks and soup kitchens reported serving very few Pakistani and/or Muslim 'clients'. Although the reasons for this were not stated, the data intimated possible forms of inadvertent exclusion, also found in analyses of food aid in North America. Representations of 'White' food predominated and few non-Muslim organisations were able to cater for cultural diets; 'Whites' were over-represented among the staff and 'clients' (Guthman, 2008); and a large minority of food aid required 'clients' to engage with Christian doctrine or symbols.
However, this finding may also be explained by possible lower levels of food insecurity among Pakistani Muslims. The ethnic density hypothesis suggests that some ethnic minority groups, despite often being in a low socio-economic position, have better health outcomes than expected due to support within their social networks (Pickett and Wilkinson, 2008). Lower levels of food insecurity among Pakistani Muslims would be in keeping with this. The finding may also be explained by the existence of alternative, hidden forms of food assistance among the Pakistani Muslim community surrounding mosques, as suggested by Klas and Gerdner (2011). The latter found that, in Sweden, Muslim congregations are not only religious meeting places, but also social meeting places and centres for the organisation of a range of social welfare services. There are also non-Mosque based arrangements in Pakistani Muslim communities for sharing savings in mutual aid groupings and distributing according to need. While the particular forms of Islamic finance in terms of banking and associated 'loans' have been examined in the literature (Warde, 2000) the more informal, albeit widespread, expressions of mutuality are relatively hidden (see ethnographies of contemporary life by Alam (2011) and narrative collections linking life for British Pakistani men in Bradford with contemporary policy issues (Alam and Husband, 2006)). Further research looking at food insecurity and networks of food aid among the Pakistani Muslim community in Bradford is required to understand the extent to which current food aid excludes people in need.

Conclusion
This study identifies UK food aid as multifaceted and dynamic. Although food aid is adopting service responsibilities previously borne by the state, this does not imply an extension of the 'shadow state'. Instead, it is reflective of a pre-welfare state system of food distribution, adapted to be consistent with elements of the 'Big Society' narrative. In line with this, community food-assistance for people in need was long-established in the case-study area; the 'new' element concerned the type of 'client'. Faith-based food aid is common and most faith-based providers are Christian. There is little Muslim provision of (or utilisation of) food aid, despite the local demographic context.
The study provides an insight into the different types of organisations responding to food insecurity at a time of high political and media interest but limited academic analysis. By situating the case-study analysis within the broader landscape of and debates on food security provision, it brings some clarity to food aid terminology and illuminates points of tension within and between organisations of different types. It is the first academic study in the UK to look in detail at the faith-based arrangements of Christian and Muslim food aid providers and explore how faith-based food aid organisations interact with people of other faiths. As such, it raises concerns about the accessibility of community food aid.
Nevertheless, this is a small-scale study of community food aid in one city. The sample of community food aid providers is biased towards emergency provision and includes only secular, Muslim and Christian organisations. The study did not interview users of food aid and, therefore, findings about project inclusivity and the experience of 'clients' are the perceptions of staff (note interviews with low-income South Asian and White British women are ongoing at the time of writing; the results of this research will be published by these authors in due course). Comparison of this study's findings with a demographically similar city would be a fruitful line of future research.

Notes
2 'Clients' is used to describe the people accessing food aid. The term was chosen for two reasons: it was considered preferable to 'recipients', which may connate passivity, and to 'users'; and it was used by interviewees to describe people using their services. The term has many flaws -not least its association with systems of market-based production; however there was, unfortunately, no scope for critique.
Wright, J., Small, N., Raynor, P., Tuffnell, D., Bhopal, R., Cameron, N., et al. (2013) Food bank Provides free food parcels often on the receipt of a voucher. This is the main function.

Foodbank
Provides similar services to 'food bank' but registered as a name by the Trussell Trust.

The Trussell Trust
The Trussell Trust is a charity founded on Christian principles with a mission to end hunger and poverty in the UK. From 2000 it began providing foodbanks, creating the UK Foodbank Network in 2004.

Soup kitchen
Provides onsite emergency free food provision to low-income, high needs clients.

Soup run
Provides mobile emergency free food provision to low-income high needs clients. Day centre and 'drop-in' centre Offers various forms of food provision, free or subsidised, as part of wider support, which can be targeted at particular demographic or socio-economic groups. Community cafe Provides low cost or subsidised food, often with very low overhead and staff costs.

Social Food Charity
Offers home-cooked food made from surplus and locally grown ingredients, to be eaten communally, for very low cost or on a Pay As You Feel basis. Pay as You Feel (PAYF) A participative pricing mechanism which delegates the price determination to each customer and requires the seller to accept any price. Also known as 'Pay What You Want'. Food recovery organisations A blanket term that captures organisations that supply food from corporate donations and orchestrate large-scale food collections for individual food banks, and those that focus on gleaning fresh produce from farms and other sources. FareShare A charity which redistributes fresh, quality and in date surplus from the food industry and other charities. Intercepted food Term used by study participants to describe food used by charities that would otherwise be thrown away by wholesalers and retailers. Community kitchen Community-focused cooking-type programmes providing an opportunity for a small group of people to meet regularly in order to communally prepare, and possibly also eat, a meal.

Community supermarket
Food is sold as groceries at below market prices, also known as social solidarity stores and often targets or is restricted to low-income clients. Food co-operative Community owned and operated food distributors selling low-cost, often organic, food. Community-focused, and also often community-initiated, horticultural programmes. Aims to increase access to organic healthy food and train disadvantaged groups in horticulture. May also help to support biodiversity and improve green spaces. Food desert An area where cheap and varied food is only accessible to those who have private transport or are able to pay the costs of public transport if this is available. Access to a cheaper and wider range of food, for some of the groups who need it most, is extremely restricted.