Most policies for the regeneration of urban neighbourhoods raise questions about social balance, particularly when, as is the current situation in Britain, social inclusion is a key theme of government policies. In practice, however, concerns about social segregation are mostly manifest in relation to social housing estates (meaning estates managed by local authorities, housing associations, and other types of non-profit- making housing agencies). Social housing accounts for about 23% of the housing stock in Britain (Wilcox, 1998, page 95). It has experienced a long history of `residualisation', meaning a tendency to house only those types of households öthe poor, unemployed, those experiencing a relationship breakdown, people in debt, and people with a history of mental illness^who have little choice in where they live (Cole and Furbey, 1994, pages 82 ^ 85). At the same time, in the absence of countermeasures, residualisation is generally accompanied within social housing by a sifting process that forces the most vulnerable households into less attractive accommodation (Willmott and Murie, 1988, pages 40 ^ 43). Despite its significance, the literature on social balance in social housing is scattered. It comprises a mixture of government statements, surveys of housing practice, and local case studies. It deserves clarification and synthesis, as is the aim here. In the first section we reflect on the use of discourse as an analytical method. The examination of discourse, in turn, establishes a framework for the remaining sections. Discourse analysis and its application Assessments of social balance and related terms such as social mix presuppose that the terms can be easily defined. Therein lies a difficulty. Social balance and mix cover numerous overlapping indicators of the characteristics of a population, including class, income, employment status, age, and ethnicity. They also apply variously to different spatial scales öthe street, the neighbourhood, and groups of neighbourhoods. Social `mix' is merely a description that may apply to virtually any urban neighbourhood. No neighbourhood has a completely homogenous population. Social `balance' is intangible Social balance and mixed neighbourhoods in Britain since 1979: a review of discourse and practice in social housing Barry Goodchild, Ian Cole Centre for Regional, Social and Economic Research, Sheffield Hallam University, Sheffield S1 1WB, England; e-mail: B.Goodchild@shu.ac.uk Received 6 July 1999; in revised form 14 March 2000 Environment and Planning D: Society and Space 2001, volume 19, pages 103 ^ 121 Abstract. In Britain, current interest in social balance has arisen partly as a response to increased management difficulties in social housing and partly in response to the concepts of the underclass and social exclusion. Social balance is now endorsed in some, but not all, the relevant urban policy statements of the New Labour government elected in 1997. A modified form of poststructural discourse analysis offers the best way of understanding the term and its implications. The approach focuses on different levels of social reality öthe level of a national policy, the level of management and estate upgrading, and the level of the social experience of residents. In doing this, the approach reveals the varied and occasionally contradictory meanings of the term `social balance'; it highlights the increasingly common attempts by social housing agencies to control the characteristics of their tenants; and it enables a sceptical assessment of the practical significance of mixed housing estates for poor individuals and households. DOI:10.1068/d39j