1. 1 Why is it so difficult to create really good places? Recently built affordable housing in England Sarah Monk, Centre for Housing and Planning Research, University of Cambridge Rebecca Tunstall, Department of Social Policy, London School of Economics Paper delivered at the ENHR conference, Toulouse, 5-7 July 2011

2. 2 Outline of paper Aims Methods Policy context 2000-10 National impacts 5 local authority and 19 site case studies Whether the sites were ‘really good places’ Conclusions and implications

3. 3 Aims This paper asks, after a decade of policy on new housing, how many of the new affordable housing developments in England 2002/04 to 2007-08 can be described as ‘really good places’? If some developments are not ‘really good places’, why? We define ‘really good places’ as housing developments that: 1)Were popular with current and potential residents, and 2)Were in ‘convenient’ locations, and 3)Were not in ‘very deprived’ neighbourhoods, and 4)Were on sites with no obvious environmental flaws, and 5)Had ‘good’ design and layout, and 6)Had a mix of tenures and home types.

4. 4 There were three main sources: 1) Analysis of the Homes and Communities Agency database on all the affordable housing schemes it funded in England 2002/03-2007/08 2) Case studies of 5 local authorities, and interviews with local authority planning staff: -Swindon, South Shropshire, Croydon, Maidstone, Sheffield 3) Within these areas, case studies of 19 affordable housing schemes built 2002/03 to 2007/08, using desk research, interviews (as above) and site visits. Methods

5. 5 A decade of policy on new housing development, 2000-10 The key policies were: 1)Targets for the use of ‘brownfield land’ 2)Increased minimum housing densities 3)Reduction in funding for affordable housing and increased use of developer contributions 4)Housing quality standards and emphasis on design 5)Policy to encourage mixed communities (la mixite sociale) 4, 5 directly aimed at making ‘really good places’ to live 1, 2 aimed to improve urban areas more broadly 3 aimed to reduce public expenditure on affordable housing while maintaining supply, alongside encouraging mixed communities

6. 6 These policies were unusually effective: 1)The target of 60% of all new housing on brownfield land was reached 8 years early and by 2007 more than 77% of new housing was either on brownfield land or conversions 2)While in 1998-2001, the average density of new housing was 25/hectare, by 2006/09 it was 43. There was also a sharp shift from houses to flats and from 3-bedroom homes to 1 and 2 3) By 2009 more than 60% of all new affordable housing was delivered at least partly through the planning system. Impact of policies 1, 2, 3 on new housing at national level

7. 7 Amount and type of new affordable housing 2002/02-07/08 at local level NumberProportion social rented Croydon1,48865% Swindon1,07952% Sheffield 87187% Maidstone56620% South Shropshire 12878%

8. 8 Substantial but varied impact: Sheffield and Croydon – major cities, focus on regeneration, most affected by national policy, very big change in density, high proportion of flats and small homes, most affordable homes for social rent Maidstone and Swindon – growing towns, focus on additional housing supply, substantial low cost home ownership South Shropshire – rural, least affected by national policy, little change in densities, mostly produced social rented housing, houses and larger homes Impact of policies on new affordable housing at local level

9. 9 Increases everywhere – but variety Density of all new homes (homes/hectare): 1998-20012006-09 -Sheffield 23 96 -Croydon3479 -Swindon2548 -S Shropshire 1624 -Maidstone2140 (Source: CLG) Impact of policy 2 on (all) new housing at local level

10. 10 High proportion everywhere – but variety Proportion of new affordable homes 2002/03-07/08- which were flats, not houses: -Maidstone89% -Sheffield 78% -Croydon75% -Swindon44% -South Shropshire 37% (Source: HCA, some missing data) Impact of policy 2 on new affordable housing at local level

11. 11 Proportion part-funded via s106: -Maidstone36-44% -Swindon30-54% -S Shropshire 15-41% -Croydon17-33% -Sheffield 3-27% (Source: HCA, missing data; excludes those fully funded by s106 Proportion purchased by housing associations ‘off the shelf’/as a ‘package deal’ - Swindon21% - Croydon18% - Maidstone16% - S Shropshire 15% - Sheffield 4% (Source: HCA) Impact of policy 3 on new affordable housing at local level

12. 12 The 18 case study new developments Eg. Croydon

13. 13 Eg. Sheffield

14. 14 Were the new developments ‘really good places’? Most of the 18 sites definitely met several of the criteria None of the 18 sites definitely met all the criteria and were ‘really good places’ Some data was missing

15. 15 In detail… 1) 12/12 with data were ‘popular’ 2) 9/16 with data were in ‘convenient’ locations (a mean of <1km from shop, post office, school, doctors) 3) 10/16 with data were not in the 20% most deprived neighbourhoods (LSOAs, IMD 2010) 4) 11/16 with data had sites with no flaws (judged by visits) 5) 5/13 with data had ‘good or very good’/’average or good’ design (using CABE criteria) 6) 10/18 had tenure mix on the site, 17/18 had a size mix.

16. 16 Further detail… 1.Were the 18 new affordable housing sites popular with current and potential residents? 12/12 with data appeared generally popular compared to other existing affordable homes Planners reported high demand and low turnover 4 sites which replaced older local authority units were markedly more popular than what was there before However, interviewees said that high demand meant any extra affordable housing would be popular

17. 17 2) Are they in convenient locations? 13/16 with data were 1km+ by road from a food store 12/16 were 1km+ from a family doctor 8/16 were 1km+ from primary school 8/16 were 1km+ from post office 9/16 were <1km on average from all four key facilities 14/16 schemes were 2km+ away from the housing association office (as the crow flies), 7/16 were 20km+ away, 3 were 40km+ away

18. 18 3) Are they in less deprived or more advantaged neighbourhoods? 10/17 with data were not in the 20% most deprived neighbourhoods (LSOAs, IMD 2010) 4/7 in the most deprived 20% of areas replaced demolished homes in deprived social housing estates

19. 19 4) Did the sites have obvious flaws? 5/16 sites were constrained in access, shape or neighbouring uses, or ‘blemished’ in some way Eg. ‘backland’ sites; between big box retail, builder’s yard and railway, surrounded by railways, with utilities on site

20. 20 5) Did the sites have ‘good’ design and layout? Using the CABE criteria (HCA 2009). 3/13 with data were ‘good/very good’ 2/13 were ‘average/good’ 6/13 were ‘poor/average’ 2/13 were ‘poor’ Some criteria could not be applied post-occupancy These results matched a similar, larger study of affordable housing design (HCA 2009) The results were similar to those for market housing (CABE 2004, 2005, 2007)

21. 21 6) Did the sites have a mix of tenures and home types? 10/16 with data had a mix of affordable housing tenures within them, with from 16% to 72% social rented, and the remainder low-cost home ownership 6/16 with data were solely social rented Schemes with a mix of tenures were generally ‘tenure blind’ design, but tenures were usually at different ends of the site 2/18 had a mix of flats and houses; 10/18 were solely flats; 6 were solely houses. 17/18 had some size mix: 6/18 had some 1 bed homes; 17/18 had some 2 bed homes; 5/18 had some 3 bed homes

22. 22 Croydon

23. 23 Croydon

24. 24 Sheffield

25. 25 S Shropshire

26. 26 Why were these schemes not ‘really good places’, despite active housing development and affordable housing policy? Inherent tensions between national policies 1, 2, 3 and 4, 5, combined with success in national policies 1, 2, 3 (which requires its own explanation) None of the 5 local authorities prioritised policy 4 and 5 ‘making really good places’; policies 1,2,3, housing supply and regeneration were higher priorities Policy 3 meant reduced influence for local authorities and housing associations on site selection and layout Slow process of site materialisation and development means some sites developed 2003-2008 reflect pre- 2000 policies and partnerships.

27. 27 Conflicts between policies 1, 2 and 4, 5 1) Of case study sites, greenfield sites were least convenient, but most brownfield sites did not have easy access to key facilities either, and brownfield sites were most likely to be in deprived areas and to have blemishes 2) Sites dominated by flats (of higher density) were most likely to score poorly on design and layout However, some brownfield sites had attractive natural features, and even rural greenfield sites could present problems. Some lower density sites also scored poorly.

28. 28 Conflicts between policies 3 and 4, 5 3) Policy 3 meant reduced influence for local authorities and Housing associations on site selection, tenure and home mix and layout Eg. In Swindon the local authority felt that they got what they were given in terms of quality and design, except where a housing association was involved early Case study sites developed via s106 or bought ‘off the shelf’ with little/no housing association involvement in site or design decisions were less likely to be in deprived sites, but more likey to score poorly on design and lauoyt Local authorities were inexperienced and could have asked developers for more – eg. in Sheffield the S106 story is one of lost opportunity and Swindon also had an outdated policy for a long time However sites developed by hosuign associations could also score badly

29. 29 Conflicts between other goals - housing supply and regeneration - and ‘really good places’ at local level Sites ‘materialised’ rather than were selected e.g. Much development has been that kind of opportunistic fashion… both social and private…and We operate very opportunistically to get what we can High land costs mean that affordable housing gets the cheapest sites in the more deprived areas which may not be good for accessibility or services or mixed communities Brownfield sites often on regenerated local authority estates and replicated the poor access that the former council tenants had experienced in that area

30. 30 Summary and conclusion Despite national policies 4 and 5, none of the case study sites met every part of our definition of ‘really good places’ Location, design, tenure/home type/population mix and management arrangements may risk future maintenance, management or popularity problems There were tensions between goals 1, 2, 3 and goals 4, 5 National policy, local authorities, housing associations and developers did not prioritise 4, 5 over 1, 2, 3 and other goals Local authorities and housing associations did not prioritise achieving 4, 5 alongside other goals and/or did not have enough influence to do so

31. 31 However, greenfield sites, low density sites and those developed by housing associations also produced affordable housing sites that were not ‘really good places’ Higher density, brownfield sites and those designed largely by developers can achieve the elements that make up ‘really good places’, but this data suggests it is less likely and may be more difficult in practice The slow process of site materialisation and development means some sites developed 2003/03- 2007/08 reflect past policy and partnership, and interviews saw improvements in sites developed after 2007/08 If we want ‘really good places’ we are simply going to have to prioritise quality over other considerations, and to monitor and manage it effectively However, after the 2010 election policies 1, 2, 3, 4 and 5 and all forms of regulation of new development are all in flux.

32. 32 References This paper reports part of a wider project: Department of Town & Regional Planning, University of Sheffield; Cambridge Centre for Housing & Planning Research, University of Cambridge; Department of Social Policy, London School of Economics (2011) New affordable homes: What, where and for whom have Registered Providers been building between 1989 – 2009? Department of Town & Regional Planning, University of Sheffield; Cambridge Centre for Housing & Planning Research, University of Cambridge; Department of Social Policy, London School of Economics http://www.homesandcommunities.co.uk/new-affordable-homes CABE (2004) Housing audit – assessing the quality of new homes – London the South East and East of England London: CABE; CABE (2005) Housing audit – assessing the quality of new homes – North East, North West and Yorkshire and the Humber, London: CABE; CABE (2007) Housing audit – assessing the quality of new homes – East Midlands, Wst Midlands and the South West London: CABE HCA (2009) Affordable housing survey: A review of the quality of affordable housing in England London: HCA

33. 33