Living on the edge: welfare and the urban poor in 1930s Beijing

Abstract This article examines poverty and welfare provision in early twentieth-century Beijing as dialogue and transaction between the city government and the urban poor. Earlier studies of the Chinese urban have tended to emphasize the material aspects of urban development, and the efforts of planners and city governments to modernize China’s cities, rather than the human experience of the city. This article draws on the extensive archives of the Beijing Municipal Government Social Affairs Bureau to extend our understanding of the experience of poverty and the agency of the poor. The archive confirms that Beijing’s growing and increasingly formalized welfare institutions were designed to discipline the poor as they alleviated poverty. However, the correspondence between applicants for welfare and the Social Affairs Bureau also reveals that the poor often approached these institutions instrumentally and assertively. Recourse to the welfare institute became a livelihood tactic, a claim on the authorities in pursuit of which certain sub-groups within the poor mobilized intangible assets, from social networks to understanding of the intended terms of the system, to sway the terms of their engagement with the authorities in their own favour.

By the 1920s, official control over welfare work was rising dramatically. Most of Beijing's private charitable institutions -with the exception of the Xiangshan Orphanage, founded in 1917 and seen as a model of welfare provision -were neither large nor well documented. 6 In 1928, when recorded, state-run provision of welfare across China was low, the SAB inherited three public welfare institutes (two for men, and one for women which also accommodated children, including boys below the age of ten), two training factories (again, one for men and one for women) and a beggars' rehabilitation unit; most were located in the poorer areas of the inner city. 7 In the 1930s, some of these units were reorganized, and some (notably the women's welfare institute and women's training factory) amalgamated to cut costs and pool staff expertise. 8 By the mid-1930s, most Beijing institutes housed between 250 and 400 inmates, though numbers fluctuated seasonally and turnover was high. 9 'The poor' are here defined not by income or by occupational categories ('rickshaw-puller' , 'factory worker' , 'beggar'). Instead, they fall into two groups of those admitted by the authorities to the welfare institutes (many of whom were orphans, vulnerable women, beggars or addicts), and those who applied for admission. The latter were a self-selecting sub-group of those who lived in poverty, characterized by their own judgement that their future survival depended on official support. It is hard to allocate them to conventional categories applied elsewhere -the 'middling poor' , for example -as their stories reveal the ease with which one could slip from one category to another, from telephone operator or small butcher to recipient of municipal welfare. They include some of the working poor (typically small traders, low-level government employees, apprentices and servants), those among the unemployed who had previously worked in similar occupations, and their dependants. These were families on the very edge of economic security, rather than the utterly destitute, and their experience may therefore not reflect the full complexity, or many of the depths, of urban poverty in pre-war China.
The letters and petitions of these city people highlight a hitherto-neglected negotiation between government and people over the meanings of poverty and the status of the poor. As well as offering insights into the shocks and trends that affected Beijing livelihoods, their communications with the city authorities are inflected by their understandings of their own needs and eligibility for support. They reveal the flexibility shown by poorer families in their pursuit of economic security, and they sketch an implied contract between the poor and the city authorities. They point also to the assertive use of welfare by the poor as a livelihood strategy and not as a submission to the state project, and reveal that the state's ambitions were subject to challenge, deflection and appropriation. They thereby reveal more of the texture of urban poverty, and a more dialogic relationship between local state and individual than earlier studies have suggested. This enquiry runs somewhat against the grain of existing China scholarship. Most urban studies to date suggest a reading of the early twentieth-century Chinese city as a site of technological and social experimentation, in which the natural and social sciences, from engineering and medicine to law and sociology, were assertively applied, on terms set by self-consciously modernizing elites, to the betterment of cities and their people. In this context, poor relief was part of a wider, scientifically authorized disciplinary system that ran through public institutions across the modernizing cityscape. These studies have mined very productively the mass of resources on the reformist projects of officials, businessmen and civic activists that were designed to create more rational, efficient and orderly cities, to demonstrate China's development, and to foster loyal and compliant behaviour among city people as they channelled emerging notions of citizenship. They thereby reveal elite fascination with the urban environment as a testing ground for new approaches to infrastructure and construction, economic growth and social development. 10 The meticulous documenting of this work by elites has ensured that their voices and preoccupations predominate in the archive. While a handful of studies, most notably David Strand's ground-breaking study of Beijing in the 1920s, illuminate the lives of petty urbanites, most reveal cities with greatest clarity as they were managed from above. 11 This scholarship suggests that, as the city became an object of display, it was often imagined also as a site of discipline, in which the urban poor appeared as a blemish upon the modernizing urban environment. Scholars such as Madeleine Yue Dong and Charles Musgrove have highlighted the educative ambitions of China's city planners. 12 Tong Lam's analysis of the mid-century passion for social surveys points to their dual functions of supporting the diagnosis of national ills and demonstrating China's progress towards rational, empirically founded governance. 13 While a narrowly Foucauldian reading of this literature is undoubtedly an oversimplification, its cumulative effect is to underline, far more emphatically than other facets of cities' complex development, the assertive deployment of new forms of knowledge in the service of state and social power. In this context, poverty and associated problems were constructed in the official mind as deviancy. This complicates -but does not fundamentally challenge -established narratives that presented the poor as essentially powerless in the face of punitive and exploitative treatment by the authorities, and that gained currency in part at least for their political utility after 1949. As Aminda Smith demonstrates, these narratives of the dark pre-revolution were central to the legitimation of 'thought reform' work and the Beijing poor under socialism. 14 10 J. Esherick (ed.), Remaking the Chinese City: Modernity and National Identity, 1900-1950(Honolulu, 2000 Thus the poor did not 'suffer' poverty or disease; they were, to borrow Janet Chen's arresting phrase, 'guilty' of it. 15 This position -adumbrated in the work of Ruth Rogaski and Zwia Lipkin -is explored more fully in Chen's study of poverty and 'pauperism' . 16 Rogaski analyses the emergence of 'hygiene' (weisheng) in the north China port of Tianjin as a 'powerful discourse of Chinese inadequacy and an essential "skill" necessary for joining the ranks of the modern' , and argues that it was animated by anxious comparisons between China and the powers -European, North American, Japanese -with rights and interests in the city. As Rogaski demonstrates the interpenetration of hygienic regulation and the social and political hierarchies of the city, she observes that the weight of official scrutiny and regulation fell disproportionately on Tianjin's poorer inhabitants. 17 Lipkin highlights the role of state modernization in constructing groups among the poor such as beggars, rickshaw-pullers and prostitutes as 'deviants' and negative social exemplars, but notes also important dislocations between notions of poverty and the divisions between, for example, the 'deserving' and 'undeserving' poor. 18 Chen points to the conflation of poverty and criminality and the rising use of 'charitable detention' that dumped the vulnerable poor in the 'punitive institution' of the poorhouse. 19 Chen's analysis is most compelling where it highlights the criminalization of young poor men and documents the collision between urban modernizing projects and socially marginalized (but none the less organized) groups such as beggars or shanty town-dwellers, but is less persuasive where it addresses the cases of women, children and other recipients of welfare. She flags, but does not resolve, important contradictions within official understandings of poverty and its implications; while showing that the authorities recognized a distinction between 'undeserving' and (relatively) 'deserving' groups of poor people, she suggests little variation in the institutional arrangements within which these groups were managed. 20 As Chen notes, poverty was recognized as a pressing issue in early twentieth-century China by traditionally minded and modernizing intellectuals alike. Official responses to poverty were enmeshed with shifting understandings of social and state rights and obligations, and these were shaped both by extensive borrowing, mostly from Japanese and North American sociological thinking on poverty, 'pauperism' and 'social parasitism' , in pursuit of national development and self-strengthening, and by the conflict between these and 'Confucian' state paternalism. 21 However, in underlining the ways in which these tendencies combined to support punitive approaches to some groups within the poor, and by framing these within a language of 'guilt' , 'parasitism' , 'interrogation' and 'detention' , Chen somewhat obscures both the diverse routes that led Beijing's poor towards the welfare institutes and the ambivalence that continued to mark elite thinking. 22 This ambivalence was revealed in works such as sociologist Ke Xiangfeng's 1937 text On Poverty. While Ke warned of the risks that poverty would lead individuals into crime or dependency and increase the burden on the state, he identified causes of poverty that were in 15  most cases beyond the control of individual poor families, from the death of a breadwinner through inadequate labour rights to famine and civil war. While he noted factors such as the costs of family rituals and illiteracy, he treated these as failings of collective practice and custom, and not as individual inadequacies. 23 Between the anxieties over 'pauperism' and Ke's understanding of poverty as socially situated, therefore, we see two potentially quite distinct understandings of 'poverty as social problem' . Was 'society' in the abstract threatened by the personal failings of the poor? Or were the poor the victims of a wider social dysfunction?
In his concluding recommendations on poverty relief, Ke suggested strategies that combined the 'scientific' and disciplinary with the 'compassionate' and paternalistic. Effective poor relief must rely on detailed investigations of the poor, be managed by experienced staff, transform the minds of the poor through 'teaching' and 'nurturing' and support self-sufficiency without encouraging dependency. 24 This ambiguous understanding of poverty and the efforts of scholars to distinguish the deserving from the undeserving left ample scope for officials to apply punitive policies to ostensibly compassionate ends. Additionally, they created opportunities for those among the urban poor who were willing to conform to the normative frameworks of welfare institutions to make instrumental use of disciplinary systems. By failing to recognize the prevalence of such strategic, transactional behaviour, earlier studies have understated the agency of the urban poor in Republican Beijing.
Where the agency of the common city people has been explored, it has typically been attributed to larger groups with access to greater economic or social resources than our subjects here. Analyses of early labour activism have offered rich insights into political and economic hierarchies and into modes of workers' resistance and organization. 25 Studies of beggars and prostitutes have highlighted the importance of robust trade networks and practices that operated at several removes from official notice and control. 26 Other works have charted emerging civic engagement, shifting living standards and subaltern consumer cultures. 27 However, the transactions discussed below, between the local state and the Beijing poor over access to the welfare institutes, show us families and individuals who felt their economic and social resources were so depleted that recourse to state support offered their best hope of survival, yet they remained determined to exercise agency and maintain an appearance of autonomy.
De Certeau's notion of tactics, the 'arts of the weak' , urges greater attention to the 'popular procedures [that] manipulate the mechanisms of discipline, and conform to them only to evade them' , and invites us to consider how the discursive and institutional framework of poor relief was received, appropriated and made habitable by the urban poor. 28 Recent work in development studies on poverty and livelihoods highlights the project of making a living as one in which the poor may deploy a range of capabilities, resources and tangible and intangible assets in pursuit of family or personal security. 29 It thereby offers an understanding of poverty that is not limited to official preoccupations, draws our attention to the use of official support as a livelihood resource, and allows us to consider how the Beijing poor understood and represented their poverty as they navigated between official policy and uncertain economic conditions.

Poverty and the Beijing welfare institutes
Urban development in China was negotiated through national networks of officials and researchers, while remaining subject to local variation in understandings of 'the city' and city administration. Local government in Beijing was shaped by tensions between past and present. Beijing's history as imperial capital had left a built environment that was impressive to look at, expensive to maintain and difficult to adapt to modern economic needs. Its present was marked by the loss of political status after 1928, rising unemployment, falling consumption, and growing external pressure between the Japanese occupation of 'Manchuria' -the provinces north-east of Beijing -in 1931, and the outbreak of war in 1937. 30 In these years, Beijing's people suffered entrenched poverty, new disruption to livelihoods and downward social mobility. Although a 1932 official survey opened optimistically, boasting '1.4 million consumers, over 30,000 commercial establishments, 14,000 small businessmen, 700-odd factories, 70,000 workers, 60,000 independent labourers and innumerable family workshops…' , comments on Beijing's decline and demands for its 'revitalization' were far more common. 31 Other studies emphasized local structural economic weaknesses, over-reliance on consumption and neglect of manufacturing, leading to a 10 per cent fall in the number of tax-registered stores between 1928 and 1931, and to declining turnover in surviving businesses. 32 Employment in key commercial sectors fell dramatically, by 36 and 49 per cent in the food and drink and garment guilds respectively, and also affected former government workers, whose numbers have been estimated at around 80,000 in the mid-1920s. 33 By 1933, over 40,000 Beijing households (around 15 per cent of the total, and 170,000 individuals) were classed as 'poor' . 34 Fewer than half of these households achieved the minimum annual income -around 125-140 yuan -required for essential food and clothing. As the typical wages of a servant, shoemaker, mason or junior policeman barely exceeded this figure, we may assume that many lived on the very margins of survival. 35 Poverty alleviation and economic regeneration did not feature prominently in municipal government planning before the 1930s and, despite rising poverty and the proliferation of welfare facilities, the budget allocations for these activities were dwarfed by those for policing, education and public works. 36 In 1936, as welfare and regeneration work expanded, a recorded 34,000 jin of maize noodles, 100,000 jin of coal and 1600 padded garments were distributed to a total of 8400 poor households, and 1.6 million meals were served in soup kitchens. To stimulate the economy, loans totalling nearly 360,000 yuan -double the Bureau's annual working budget for the early 1930s -were made to agriculture, manufacture and commerce between April 1935 and February 1936. 37 The institutes admitted unemployed adults of working age and their families, the elderly or infirm, orphans and lost children. Some applied for entry, while others were referred by charitable organizations or by the police, whose responsibilities extended beyond criminal enforcement into the promotion of health and welfare. Turnover of inmates was high, though some, notably children and young women, might spend several years in the institutes. 38 Many senior welfare institute staff were relatively young, and a handful -such as Dong Shuhui, director of the children's division -were women. Many were educated to university or college level in disciplines such as law, economics or education, rather than in the pre-1905 Confucianized curriculum, and these were relatively well paid, with monthly salaries ranging from sixty to one hundred yuan for managers by 1934. 39 It is important to distinguish the Beijing institutes from European or North American workhouses. Although the image of the nineteenth-century British workhouse was 'fixed … in the popular imagination as a place of unparalleled dread' , any assumption that the Beijing institutes of the 1930s were consistently comparably punitive could be challenged on two fronts. 40 First, a more complex picture of the European institutions is emerging from recent scholarship: understandings of European workhouses as 'total institutions' are eroded by studies that point to inmate resistance and to the periodic rejection of the moral authority of the workhouse regime by charitable and civic institutions on whose co-operation the state depended. 41 Second, while the Beijing institutes' diverse missions consciously bridged disciplinary, preventative and protective functions, the balance between these functions was not fixed. Chen has noted that China's first public 'workhouses' were created in the first decade of the twentieth century as correctional facilities for petty criminals, addicts and vagrants, only later admitting the non-criminal poor on a voluntary basis. Characterizing those later admissions to the Beijing institutes also as a form of 'charitable detention' , she has argued that the institutes' mission conflated welfare and punishment, and in this she echoes some contemporary critics of the institutes. 42 However, what we can see of the institutes' workings by the 1930s reveals that the terms of that mission might be challenged or used creatively by the poor to serve their own ends.

Refuge and rehabilitation: the welfare institutes and the vulnerable poor
Elite ambivalence over poverty and the poor, as articulated above by Ke Xiangfeng, was reflected in ambiguities in the welfare institute project, particularly in the treatment of vulnerable inmates such as women and children. The epigraph to the municipal Women's Welfare Institute's 1932 handbook emphasized its mission of reforming and rehabilitating the poor as it offered them refuge: Protect the people's rights, encourage women's work Support the old and weak, relieve poverty Show kindness to widows and orphans, treat children with virtue Prohibit prostitution and slavery, rectify public morals. 43 That dual mission pervades the text: the main sections of the handbook reflect a preoccupation with the creation of good citizens through order, discipline and practical training for self-sufficiency, and the appendices record cases in which the institute had exceeded its formal obligations to inmates, in securing emergency medical care or escorting women back to their native places. 44 A woman might enter the institute if she had no family or income, wished to leave a 'mean' occupation such as prostitution, was enslaved or mistreated, was referred by her own family or by the police, or was homeless. Others were to be excluded if they were deemed incorrigible, were opium addicts or suffered from a contagious illness. Women were released only on receiving a credible offer of marriage, employment, or support by family or friends. 45 Institute regulations were designed to produce compliant behaviour and well-regulated bodies. Supervisory staff were empowered to 'train, guide, supervise and exhort, and punish' the women where necessary, and penalties for sub-standard work and infractions of institute rules varied from oral and written reprimands to restrictions on food for up to three days. Mealtimes and bathtimes were regulated as collective activities, and hairstyles were to be chosen for 'cleanliness, not self-expression' . Visiting by family and friends was limited to two twenty-minute chaperoned meetings each month. 46 Annual celebrations too were occasions for formative interventions: the handbook records speeches delivered by institute staff and inmates at New Year 1932 on the importance of renouncing self-seeking, and finding happiness within poverty. 47 There is little to show how effective those regulations were in practice: the Beijing institutes spent less time on documenting their internal affairs than many of their European counterparts. 48 Moreover, while normalizing and protective approaches were often closely linked, the balance between the two shifted between cases, and while the normalizing aspects of the project were readily accepted by those with fewest independent resources beyond the institutes, others were more assertive in exploiting its protective dimensions.
The institutes' treatment of young women highlights the importance attached to fitting inmates for conventional roles within conventional family units. Young women admitted to the institute as a refuge from domestic abuse or from work in the sex industry were expected to receive training for return to the community. Marriage between institute women and men from outside was actively promoted. Marriages took place at the rate of two to three per month, and this was one of the commonest exit routes for single women. Marriage arrangements were heavily bureaucratic. Any woman wishing to marry completed six months' training in household affairs, after which her photograph was displayed in a designated room in the institute open to men in search of a wife. Once a man had chosen a prospective wife from this gallery, the institute authorities carried out checks on his background. The final consent of both parties to the marriage was given after a single meeting in which they communicated with each other via a member of institute staff. 50 Official marriage approvals typically recorded that successful male applicants had a good standard of living and appropriate financial guarantors, and were found to be 'honest and sincere' . They offer far more detail and colour on family and financial circumstances than on the matter of character, which is typically dealt with in that single, formulaic phrase. Files from spring 1935 show that marriages were arranged between women of the institute and a mix of widowers and bachelors. Orphan Wang Xiaoduo, aged 19, was chosen by policeman Yan Shixun, 32, who earned fourteen yuan per month and lived with his father, younger brother and sister-in-law but 'had no-one to manage the household' . Tailor Bian Yuncheng, 32, matched to Zhang Jindi, aged 20, and paper merchant Du Shenglan, 23, paired with orphan Zhang Xiufang, 25, had both 'reached adulthood but never married' . Widowers Miao Xiuxin and Wang Hongjian, who both lived with elderly parents but 'had no-one to care for them' , married respectively 22-year-old Zhou Xiaoxi, who had fled her home to escape the cruelty of her adoptive mother, and 18-year-old Liu Xiaofeng, whose mother was imprisoned after attempting to sell her into prostitution. 51 The authorities here acted as matchmakers and as champions of social order as they trained women for marriage on terms set by the institute. Leaving aside the economic benefits for the city -as marriage transferred the costs of their upkeep from the institutes to their families -marriage mattered. Marriage was both a life-cycle event and a substantial expense. The inability of families to afford marriages for their children was commonly cited as a marker of poverty, and the importance of marriage in reinforcing social networks made single status in itself a form of 'social poverty' . Thus the procedures that assigned women to marriages in which their domestic labour was very explicitly exchanged for economic security also served a wider bureaucratic vision of social good.
We see few details of the women who entered these marriages, beyond their names and ages, a sketch of their circumstances before admission and the procedures attending their release, concluding most often with a fingerprint marking their consent on the marriage documents. A handful of wedding photographs survive, showing unnamed couples posing stiffly for the camera, most dressed in a traditional Chinese gown, often accessorized with a fedora for the men and a white, western-style veil for women. 52 The frequency with which marriages took place suggests that they thought even an unequal marriage preferable to the alternatives, and the archive does not reveal whether many women refused specific marriage proposals, or the option of marriage as an exit route, or how often the institute authorities denied applications for training. At a time when marriages among the urban poor might be concluded after the introduction of strangers by acquaintances, and dissolved for pragmatic, financial reasons, it is possibly not surprising that these details should have received scant attention. 53 However, the checks made on prospective husbands, the requirement that marriages take place within the institute and the formal commitment to support any woman abused by her marital family suggest a degree of ongoing official obligation towards those women.
Other records, detailing the experiences of women who were married before entry to the institutes, highlight the uneasy balance of constraint and protection implied by the institutes' paternalism, and reveal that women were not invariably the passive objects of institute procedures. While women with no family or friends beyond the institute might appear effectively powerless, women with outside connections might turn these to their own advantage, and others might use their understanding of the workings and underlying assumptions of the system to exploit limited resources more effectively.
The case of Mrs Yan (née Jin) Yuqing highlights the dependence of many women on connections beyond the institutes. 54 Mrs Yan was admitted in May 1932, because she had lost contact with her husband after he left Beijing to find work. Her entry application was sponsored by a Norwegian church in which she had been brought up and through which she had been introduced to her husband. In December 1933, missionary Olga Schult applied for Mrs Yan's release, detailing various work opportunities available to her. Although the application was approved, Mrs Yan remained in the institute and, in January 1935, Ms Schult withdrew her support without explanation. In July 1935, Mrs Yan's husband, policeman Yan Shoukang, applied for her release, explaining that he had sent his wife to the welfare institute as he had been unable to support her, hoping that she would receive some education or training there, but that now he was financially secure enough to take her back. The authorities agreed that she should be allowed to leave.
The experience of Wang Shuxian reveals the ease with which that dependence on external support might be exploited. 55 Ms Wang fled north-east China in 1931 after the Japanese invasion, leaving behind her husband, Li Ruifeng, and entered the Beijing welfare institute in 1932. When communication between Beijing and the occupied north-east was re-established in 1934, Ms Wang contacted her husband and his brother, Li Kuilin, asking them to arrange her release. Brother-in-law Kuilin, a member of the Beijing gendarmerie, accordingly approached the institute, but husband Ruifeng wrote to the SAB to say that Kuilin was not to be trusted with Ms Wang, and that she should stay in the institute. Kuilin retorted by forwarding to the SAB a letter in which Ruifeng confessed to remarrying, and berated his brother for trying to liberate his first wife, thereby exposing him to embarrassment in front of his new in-laws and the possibility of legal action for the second marriage. After some deliberation, the authorities concluded that Kuilin, being of good character, should be asked to return Wang to her home town 'to protect her rights' .
At times, the protective mission of the institute became the source of more explicit claims by the women housed there. In 1935, Dong Fushan applied for the release of his daughter-in-law, Ms Yan. Although his application was accepted, Ms Yan refused to leave, stating that she had applied of her own accord for admission because her husband Dong Ji, a junior official, had treated her cruelly and taken a former prostitute as a secondary wife. She was afraid to return to him, and would rather stay in the institute to receive training. Her request was approved. 56 The marriage cases suggest that some women received support only where they accepted the institutes' priorities, moral authority and understanding of their best interests, and other cases too show them as largely voiceless. While Mrs Yan's declaration of her willingness to work with the Norwegian church is filed with the original release application, her views on her sponsor's change of heart, on the eighteen further months that she spent in the institute, and on her husband's later intervention are not. Others were denied release for more mundane reasons. Telephone operator Wang Zhengguan found himself temporarily unable to extract his mother and sister from the institute when the financial guarantor for their release application was thought inadequate, and the departure of Mrs Cui, admitted in 1935, was delayed while the authorities checked out inconsistencies in the recording of her husband's name. 57 Other women, however, used intangible assets, from social networks beyond the institutes to understanding of institute principles, in pursuit of their own interests. For some of these women, officials appear to have made an effort to weigh up conflicting evidence and consider a woman's future before handing down a decision. The social or economic status of the parties involved might be taken into account, but was not always the decisive factor. On the one hand, institute recognition of Wang Shuxian's rights independent of (and in defiance of) her husband's wishes, and the decision to allow her release, were secured only by proxy, through the intervention of a brother-in-law who was himself of good standing in the eyes of the institute authorities and was able to reveal Li Ruifeng's ulterior motives. On the other hand, Dong Ji's status as a government employee and his family's financial security counted for less than his wife's accusations of abuse, and her determination not to return to him. 58 The institutes' handling of cases involving unattached children further underlines assumptions of official obligation towards vulnerable citizens. Many of those admitted to the institutes were minors, including orphans, abandoned or lost children, and runaways, with no known relatives in Beijing. It is often recorded that they were found 'wandering destitute' -sleeping rough, or begging -before being referred to the institutes, typically by the city police. Official handling of these cases appears to have been relatively sympathetic. Their education and training followed the standard primary curriculum, and the institutes' roles at times included mediation between conflicting interests and understandings of family responsibilities.
In some of the most straightforward cases, the institutes assumed responsibility for young people affected by crises such as war, civil disorder or bereavement. By the time Zhang Zhu was referred to the SAB in 1934, she had seen her father killed by bandits and her mother abducted. She had fled with her older sister from their home in Henan, completing the journey of several hundred miles to Beijing alone when her sister died of starvation. On reaching Beijing, Zhang supported herself by begging, but was later admitted to hospital with a leg injury and thence referred for welfare institute admission by a doctor, who wrote, 'She has no friends or family here, and it would be terrible for her to lose her chance of education while so young. If the mission of the welfare institute is to alleviate poverty and hardship, surely something can be done…' 59 Other children were admitted to the institutes to escape neglect or exploitation by their families. Orphan Gao Yi' er was rejected by his uncle after losing his job tending pigs. Yu Lianfang was sent to the women's welfare institute by the municipal courts after her father tried to sell her and five other girls to a theatre troupe. Zhu Pengqin was sent by his mother to seek an apprenticeship with a Beijing joinery that already employed his uncle, but was abandoned after the uncle absconded with all Pengqin's belongings and apprenticeship fee. Feng Bingguang's adoptive father refused to reclaim him from the police, declaring that the boy, then aged twelve, had repeatedly stolen money from home, and deserved to be detained. When ten-year-old barber's apprentice Liang Shunyi lost both his parents, his relatives seized control of the house and farmland he had inherited, took him to Beijing and left him there, threatening to kill him if he ever returned home. 60 Material and economic conflicts predominate in these stories, to the extent that when one child, Ma Xiaodai, told the authorities that his widowed father and new stepmother had abandoned him because they 'did not love him' , it appears as something of an anomaly. 61 More commonly, the orphans and wayward children referred to the institutes had been resented as burdens or as competitors for limited family resources, or treated, like Yu Lianfang, as commodities or sources of income. It was expected that the children of poorer families would be economically active. Gamble noted apprenticeship as a common experience for the young in his survey group, particularly for boys, and also pointed to the role of invisible labour, for example, in scavenging for fuel by younger children. 62 Yet the reliance of families on the labour of the young created potential conflicts of interest, even where this was framed within legal arrangements. Guo Xiaoyang and Wang Wengui were admitted after walking out of apprenticeships where they were beaten. Wang, aged thirteen, told the police that he was afraid to go home in case his mother -a widow with other children -sent him back to his former employer. 63 While Yu Lianfang's case appears as a classic illustration of the abuses associated with the commodification of family members -particularly women and children -Wang Wengui's predicament alerts us to a subtler form of the same problem.
Generally, the authorities recorded efforts to trace relatives or to pursue alleged crimes. While they had some success in restoring lost children to their families, they frequently drew a blank in criminal cases: investigations located neither the joinery where Zhu Pengqin had sought apprenticeship nor his original home, and although Liang Shunyi was able to supply the names of his relatives and home village, and details of the disputed land, local officials denied all knowledge of the family and the case. However, this did not visibly affect officials' assumption that they had formal obligations in such cases, and even children who had run away from home over apparently minor quarrels, children who had strayed carelessly and become lost or children whose stories could not be verified in any detail were admitted, at least temporarily. Nor did the authorities readily release minors, even where they had friends or relatives willing to take responsibility for them. Orphan Zhang Dalu was admitted with his brother and sister through the efforts of colleagues of his late grandfather, a bank employee. Two years later, when an aunt was located in Nanjing who was willing to take the children in and educate them, the authorities delayed their release while cross-checks were made on the aunt's background and the personal details of the children's sponsors. 64 More often than not, inmates and their families complied readily with official directives: official support, after all, might often be preferable to sleeping on the streets or returning to an abusive family. Yet even among the families of highly marginalized groups, such as beggars and opium addicts, some tempered their compliance with requests for leniency, or with assertions that family needs should outweigh official prohibitions of begging, while others more openly resisted official judgements and labels. As well as underlining the agency of poor urban families, these cases also point to the variations within these categories: not all 'beggars' were homeless, or vagrant, and not all opium addicts were as useless to their own families as they appeared to the state.
For example, when the son-in-law of elderly Mrs Guo -referred to the welfare institute for begging -applied for her release, he appealed for compassion, declaring that she was old, ill and hungry, and 'had not understood that this was against the law' . He undertook to manage her behaviour so that she would not re-offend. 65 Nor did nineteen-year-old Zhang Diangui deny that his aunt had been begging, but he implored the authorities to release her to take care of his sick brother, who 'cannot eat because of his illness, but weeps all day' . Zhu Fuliang, applying for the release of his son Zhu Fengming -admitted as an opium user -undertook to supervise his son and find farm work for him, but also insisted that his family needed Fengming: 'I lost my parents as a child, and my wife when I was in my prime. I have only one son, and we depended on each other as life itself … after he left for Beijing we heard nothing. His wife weeps all day, and I am bedridden and cannot work, so we cannot survive without him. ' 66 Ma Yuting's wife alleged that the police had mistaken him for a beggar as his clothes were tattered, and the release order appears to acknowledge this as an error. The same misunderstanding caught up the 78-year-old Mrs Qing, who was found 'apparently begging' in dirty clothes. 67 These responses point to the mix of interests at stake: was family dignity better defended by policing one's own relatives, or by challenging official accusations against them? And which approach was more likely to secure a relative's release? Was the alleged beggar or addict a burden on the family, or a source of domestic support? Was compliance with official prohibition of begging or narcotics use an obligation, or was it an observance to be traded in the service of a family's practical and affective needs?
Thus official support -with its associated constraints -was readily accepted by those with fewest alternative resources: by orphans, refugees or young women who could not rely on their own families for support or protection. While there were surely solid practical reasons for compliance in these most vulnerable groups, their apparent conformity is also a product of official recording (or non-recording) procedures that solicited inmate testimony only at specific times and in specific forms. Others, however, were more assertive, or had resources outside the institute, and it is striking that, where their voices were heard, they were often animated by a lively sense of personal interest as inmates and their families judged and commented on the welfare institute project and measured it against their own needs.

'It would be a truly virtuous act': the welfare institutes as livelihood resource
If we shift our attention from referrals to applications, the balance within the transactions between local state and the urban poor also shifts. While the normalizing project of the institutes remains visible, it becomes even clearer that applicants were determined to turn that project and its underlying assumptions to their own ends. Recourse to the welfare institute became a livelihood tactic, a claim on the authorities in pursuit of which certain sub-groups within the poor mobilized intangible assets, from the social networks that supplied references and guarantors to their understanding of the terms on which a successful application might be made. While de Certeau has pointed to the 'joyful dexterity' inherent in such tactics where they are deployed to foil a constraining system, the concern of our subjects here was not to subvert or evade the system, but to appropriate it, and to use its very solidity as a source of support. 68 Many such applications came from adults -most often from mothers or from widowed fathers -seeking admission for young children or for whole families; a sizeable minority came from adults as individuals. These applicants were not destitute: many had a fixed address (as noted above, even those admitted for begging might have families in more stable circumstances), some limited income and a place in some social network. For these people, recourse to the welfare institute did not represent a simple failure in livelihood. Instead, the institutes were used purposefully to supplement normal livelihood resources and claims where these proved inadequate. Their applications appear designed to highlight both their exposure to common economic and political shocks and trends, and their resourcefulness in the business of making a living as a qualification for access to official support.
Application procedures involved some financial outlay, and made demands on intangible resources such as social networks. The SAB archives contain relatively few unsuccessful applications, and it is hard to tell how many were rejected, or how many possible applicants were deterred by the costs and procedures involved. Applications were submitted on paper, and applicants unable to write for themselves might have had to pay to have them written by someone else on appropriate terms. Application forms had to be paid for, at ten cents for the first hundred characters and one cent for a guarantor's form (this for a family that might be living on fifty cents a day). A more substantial hurdle was that applicants also needed a guarantor: a proprietor of a reputable, tax-assessed business who could affirm that the application was made in good faith, and offer compensation if the applicant failed to comply with regulations. The role of officials appears somewhat ambiguous here. While the city police acted as gatekeepers, making home visits to verify applicants' claims of poverty and their guarantors' credentials, Chen notes that SAB clerks served as scribes for some applications, acting also as mediators or facilitators between applicants and government. 69 This restricted voluntary admission to the institutes to those who were willing to accept the procedural discipline involved, and who could call on a taxpayer to vouch for them. The cases noted above of Mrs Yan (née Jin) Yuqing, and Wang Zhengguan's mother and sister suggest that anyone whose connections lapsed during their time in the institute would have found it difficult to get out. Thus voluntary entrants (unlike the beggars and addicts who were forcibly admitted) typically came from families that lived on the edge of social or economic security, but who were much further from the margins of absolute survival than Beijing's very poorest families. These families were of far lower status than the genteel 'shame-faced poor' who received discreet charity in the nineteenth-century Netherlands. 70 Yet their treatment by the authorities reflected a similar assumption: that there were some whose past status entitled them to official support. Their efforts underlined the distinction between the struggling and the destitute. Consciousness of that distinction informed many applications, and applicants worked to appear deserving -a good investment for the city government, so to speak -as well as desperate.
The majority of applicants embedded their account of personal crisis in a narrative that covered -sometimes minimally, sometimes in great detail -not only the straits to which poverty had reduced them, but also those episodes that showed them to be resourceful, industrious and therefore worthy of support. They evoked years of hardship weathered and efforts to secure a sustainable living before resorting to welfare. Applicants detailed their reliance on family and friends, their willingness to move to find work or sell family possessions to remain afloat, emphasizing that their difficulties were the product of forces beyond their control, of regional economic decline, or of external pressures such as civil and other wars that had disrupted their access to personal networks and resources. 71 Some petitioners vividly described their poverty, writing of cold, of their inability to provide one good meal a day for their children, of their fear of starvation, of families sitting weeping at their plight. When Mrs Wang (née Liu) applied for admission with her two children, she declared, 'My husband is away and we cannot contact him. My children are young, life is hard, and we are cold and hungry. If you would do us this great kindness, it would save us from certain death. ' The police report on Mrs Wang's case stated that the family was 'in a most pitiful state' , and reports routinely confirmed that applicants were indeed as poverty-stricken (literally, 'poor and cold') as they claimed. 72 However, poverty was complicated. Having a job or business was in itself no guarantee of an adequate living, and the decision to apply to the welfare institutes did not always mean that a family had reached the end of the line. Applications addressed a mix of economic and welfare concerns, aiming to build future resilience as well as remedying present crises. It might be more precise to suggest that, for some, the exploitation of welfare offered hope of evading destitution, rather than escaping it, and that it might make the difference between security and insecurity, rather than between life and death. Many families, for example, sought institute places only for their middle children. These were old enough to cope with the austere institute regime, but too young to contribute much to family income. Their admission relieved pressure on family resources while leaving open the option of their return once the family's economic position was more stable. Thus Zhang Jing'an applied for admission for his nephew, whose father ran a small vegetable business but was unable to support his family of four. Li Fengqi's market stall did not make enough to feed his wife and five sons, so he sent the second and third boys to the institute. 73 After Mrs Song (née Gao) lost her butcher's business, she asked the welfare institutes to take her sons, aged six and eleven, but not herself. Mrs Yan (née Song)'s husband was a small trader 70  whose business failed. He left Beijing to try his luck elsewhere, and Mrs Yan applied to have three of her five daughters taken in, keeping her oldest and youngest girls at home. 74 In many cases, families specifically asked that their children receive education or training which they were unable to provide. Widower Meng Xiahe, who fled Japanese-occupied north-east China in 1931, applied for admission after failing to find work for himself or an apprenticeship for his son in Beijing. Forty-year-old former official You Shangjun asked that his sons, aged eight and ten, be taken in for schooling. Li Tuyi, a joiner, sought a place in the training factory for his eight-year-old son -business was slack and the family of five 'could not afford one square meal a day' , even though his wife took in washing. Mr Li, a farmer, asked for training for his eleven-year-old son: 'He is eager to learn a trade, but we don't have the means to support him in this…' Mr Li was careful to frame this modest ambition within a desire to serve public interests, continuing '...we would be most grateful, and this might do something to remedy our nation's problems' , but the longer-term implications for family economic security were equally striking. 75 Petitions from former officials underlined the impact of economic instability on groups that had once been relatively secure. Fei Jiandong was a Ministry of Transport clerk until 1928 when the Beijing government was dissolved and he lost his job. In 1934, unable to support his three sons, he asked to have them admitted to the welfare institute. 76 An Dingxin had the same experience, finding temporary security in the Changlu Salt Bureau, until Japanese intervention in the salt administration forced a reorganization and cost him that post too. In 1936, aged thirty-seven, tired of wandering the streets and unable to afford even one square meal a day, he applied to enter the institute to work and to 'avoid starvation' . 77 Shen Delin, aged twenty-nine, was a government clerk with a monthly salary of over ten yuan, which he thought adequate for his family's needs until he too lost his job. His wife and child fell ill and died, and he asked to enter the institute for shelter and to learn new skills. 78 These men might once have found work as teachers -a traditional secondary occupation for ex-officials -but competition in these alternatives was becoming more intense. Gao Jinghan, an experienced teacher aged fifty-five, applied for admission to the institute in 1936. He had lost one job, lacked the formal qualifications that schools now demanded, and was unable to find another. 79 Those who had been lower-level public employees appeared more anxious to show themselves as deserving. When Gao Yufeng applied for admission he emphasized that, despite his poor background, he had managed to secure both an education and a police job -respectable work, though not well paid. When he found himself unemployed, he made a living performing 'cross-talk' comedic dialogue, a popular genre, but a not prestigious one. Only when he fell ill and lost the use of his hands and feet did he appeal to the SAB for support. 80 Another ex-policeman and graduate of Fengtian Police College, Wang Fuguang, filled two closely written pages detailing his efforts to support himself, selling goods of his own production across China, after losing his police job. 81 These applications emphasized the gulf between past and future prospects. Employment histories were cited as a marker of entitlement, rather than as evidence that applicants should continue to fend for themselves. Evidence of past self-sufficiency was translated into a claim for admission to the institutes, and treated less as a disciplinary imposition than as a qualification for official support.
While many petitions speak most directly of men's contributions to family income, women's work mattered too, and women worked purposefully to exploit the opportunities offered by official support. Mrs Zhang (née Zhang) came to Beijing to find work; when she asked to have her two sons, aged eight and eleven, admitted to the welfare institute, she had had no news of her soldier husband for three years and had given up hope of supporting her three children and mother-in-law on her own wages of three yuan per month. 82 In August 1934, Mrs Fu (née Cao) asked for a place for her daughter so that she could leave Beijing to look for a job elsewhere; her unemployed husband had already gone to Tianjin to find work. This was one of a number of applications made by working mothers, their relatives or -occasionally -their employers, for the care of children. 83 Mrs Fu's application was not a declaration that she was unable to make a living, nor does she appear as the passive recipient of official handouts or official control. This was rather a deliberate move to release her own earning power by delegating care of her daughter to the authorities.

Conclusions
daughter-in-law of Dong Fushan, who refused to return to an abusive husband. Here, the welfare institutes offered social as well as material support.
Third, the archive invites us to consider poverty as a genuinely social problem, not simply in that it occurred 'in society' and stood between officials and their understandings of an ideal social order, but also in that it was shaped in part at least by diverse social expectations, experienced as social impairment or exclusion, and might be (for the more fortunate) masked or mitigated by social resources. For the urban poor, as we see them in the archive, poverty meant shame, disappointment and anxiety as well as hunger and cold, and this social understanding of poverty was an accepted strand in the dialogue between the poor and the local state. While this was undoubtedly a formulaic dialogue, its formulae were more complex than can easily be accommodated within a simple binary of guilt or deviance on the one hand, and control or charitable incarceration on the other. While it would be an overstatement to suggest that the dialogue was governed by shared notions of social contract between government and people, correspondence on entry to and exit from the welfare institutes none the less reflects understandings on both sides of obligation and entitlement. The correspondence between families applying for access to the welfare institutes cannot show us definitively how far the poor had embraced the normative basis of welfare. However, it does suggest that they recognized formal observance of those norms as a necessary part of the transaction.
The welfare institutes operated within a web of social relations that extended beyond the immediate negotiations over access to support. Application procedures required that individuals deploy a range of intangible resources in order to qualify for admission to the welfare institutes, ranging from understanding of the 'proper' terms on which application might be made (in the display of resilience, flexibility and the will to self-sufficiency), to the mobilization of appropriate connections who could act as guarantors. This produced a dynamic of inclusion and exclusion -or a form of triage -among the urban poor, managed by the officials who pronounced on admissions to the institutes.
Underlying those processes of qualification and triage, however, was an assumption that the authorities were obliged to provide support to those who proved themselves deserving. The references to the 'virtue' and 'generosity' that would be shown in approving an admission application appear deferential but, in evoking traditional images of benevolent, paternalistic officialdom, they set a moral standard against which bureaucratic decisions would be measured. They therefore stood both as declarations that the applicant stood on the right side of the lines of inclusion and as assertions of moral community between petitioner and bureaucrat. It is striking that, while many petitioners were craftsmen, farmers or servants, others had much more in common -in social, educational and occupational terms -with the officials to whom they appealed for relief.
Thus the archive reveals the Beijing poor as purposeful and instrumental in their pursuit of welfare, working to understand -and to represent to officials -their own poverty on terms that would qualify them for support. The welfare institutes became a potential resource, relieving families of the burden of feeding younger children, and training those children to be more productive as they sheltered them. Beyond their practical functions, they also offered opportunities for the maintenance or restoration of individual and family status and dignity. Those concerns for dignity mark the self-presentation of applicants, in some cases as they assured the authorities that they were worthy recipients of official support, in others -as we see with the families of those admitted for begging or narcotics use -as they accepted or resisted the labels applied to relatives by the police and city authorities. Returning to de Certeau's suggestion that the 'consumer' conforms to disciplinary procedures in order to evade them, we may propose that the Beijing poor who sought welfare were more ambitious than those 'consumers' , despite their greater vulnerability, in that they conformed to local state discipline in order to benefit from it.
A study of the welfare institutes therefore reveals not only the fragility of livelihood but also the creative deployment by some poor families in Beijing of all possible resources in the quest for greater security. The framing of admission applications and the apparently quite instrumental expectations of some applicants suggests that, where new paths towards a more secure livelihood were perceived to exist, they were used assertively by those in need, and that -for sub-groups within the poor at least -welfare institutions were more than just a port of last resort. Admission applications reveal a tentative negotiation between families and the local state, and complex relations between power, agency, and access to material and social support.