Public accountability: reforms and resistance in social housing

ABSTRACT The paper illustrates and discusses the changing nature of public accountability relations in public services that are transferred to a mutual organization. The paper contributes to the literature that studies resistance to the neo-liberal restructuring process, highlighting civil society campaigns as important actors. A social housing case study in Britain is developed, combining a critical realist and dialogical influenced analysis. The case study identifies the role of private finance, illustrating tensions between democratic-styled and NGO-styled public accountability relations. A conclusion that housing mutuals are in danger of being apologia for private capitalism is established.


Introduction
Housing continues to be a major political and policy issue in Britain (Dorling 2014;The Lyons Housing Review 2014). House prices are rising faster than real wages (HM Treasury 2015), with the average house costing eight times the average wage (The Lyons Housing Review 2014). The adoption of neo-liberal policies towards social housing was predicated on the withdrawal of public funding being replaced by private investment. This has resulted in council house building falling from over 150,000 in 1975 to just a few hundred each year throughout most of the past two decades (Wilcox and Perry 2014, Table 19g). In contrast to the hypothesis advanced by the Conservative government of the 1980s that the state had crowded out private sector construction, the withdrawal of public funding has not been replaced by the private sector, 'in fact total new construction [has] declined' since 1975 (Murie 2012(Murie , 1032. This lack of supply has led to increasing house prices, inequalities in access to housing and an inability to provide decent affordable housing for all (including the poorest and most vulnerable) in our society (Dorling 2014).
The housing association (HA) sector has grown significantly due to the government policy of transferring council housing stock to it, 1 allowing borrowing to be taken off the government's balance sheet. 2 HAs are considered to be one of the successes of the not-for-profit sector. In 1981, they held 2.2 per cent of the total housing stock in Britain; this had grown to 10 per cent by 2012 (Wilcox and Perry 2014, Table 17d).
Council housing is woven into the fabric of British society, with democratic accountability relations at its heart via elected councillors. However, the neo-liberalization of the service has changed public accountability relations, creating an (public) accountability deficit (Eckersley, Ferry, and Zakaria 2014;Shaoul, Stafford, and Stapleton 2012;Smith, Mathur, and Skelcher 2006). This paper contributes to the debate over the nature and causes of this public accountability deficit. The reforming process analysed below concerns a housing stock transfer from local government to a private not-for-profit housing provider as a means of securing the resources required to remedy a backlog of housing disrepair (Pawson and Mullins 2010). In this process, the public accountability relations are reformed from democratic accountability with local government to accountability relations based on a non-governmental service organization (NGSO) (Ebrahim 2003) which has more in common with private sector governance and accountability approaches (Collier 2005;McDermont, Cowan, and Prendergrast 2009). Significantly, this process has also generated resistance in the form of civil society campaigns that contest the changing nature of public accountability (Smyth 2012).
The new NGSO is a community housing mutual (CHM), where the housing stock is owned by tenants through a mutual organization (Pawson and Mullins 2010;Waite 2008). The use of a CHM is designed in part to overcome a public accountability deficit in non-mutual stock transfers (Smyth 2013a). In analysing this public accountability deficit, the perspective of the tenants is adopted. This leads to three lines of analysis covering the competing conceptions of public accountability (an 'accountable landlord'), the governance/accountability structures of the CHM and the role of finance (both public and private) in the restructuring process.
The remainder of this paper is structured as follows; the next section discusses the nature of public accountability starting from a basic definition of accountability that is then contextualised in democratic societies, before exploring the relevant elements of Critical Public Accountability (Smyth 2012(Smyth , 2013a to understand the restructuring/resistance dynamic evident in the competing conceptions of an accountable landlord. This is followed in the section 'Housing stock transfers and NGSO' by an analysis of the stock transfer policy in Wales and how the use of a CHM attempts to develop new governance/accountability structures, while attracting private capital investment. The section 'Research design' outlines the research design, covering the use of a critical realist theoretical framing and dialogical data analysis. The section 'Case study: a CHM' sets out the case study addressing the three lines of analysis identified above, before a final section discusses the findings.

Public accountability
Accountability is a totem of the neo-liberal age. It is elusive, chameleon-like and everexpanding (Dubnick 2002;Sinclair 1995;Mulgan 2000). Accountability is mobilized to justify public services reforms (Broadbent, Dietrich, and Laughlin 1996;Dubnick 2002); yet, it is notoriously difficult to define, with synonyms (e.g. transparency and answerability 3 ) regularly used in its place (Ferry, Eckersley, and Zakaria 2015;Smyth 2012). 'Accountability has become an icon of good governance' (Bovens 2005, 184) but is also a garbage can, containing ill-defined concepts and good intentions (Bovens 2005). This section starts by defining and contextualizing accountability in the public services, before exploring the restructuring/resistance dynamic as conceptualised in Critical Public Accountability (Smyth 2012).
Accountability is more than accounting but not reducible to individual (or simple aggregations of) experiences (Cooper and Johnston 2012). The central elements of accountability are an obligation to present an account of and answer for the execution of responsibilities to those who entrusted those responsibilities. On this obligation depends the allocation of praise and blame, reward and sanction so often seen as the hallmarks of accountability in action. Gray and Jenkins (1993, 55) Bovens (2005) details the elements of public accountability which, in addition to those stated above, includes public accessibility (transparency) to the account giving (answerability). Further, the most important function of public accountability 'is democratic control' (Bovens 2005, 192, emphasis in the original). Therefore, the 'reward and sanction' element, when contextualized in modern democratic societies, sees power emanating from the people and the people holding their representatives to account through periodic elections (Bovens 2005;Mulgan 2003). Or as Michels and Meijer (2008, 169) put it, for public accountability the 'crucial question . . . is whether accountability arrangements ensure popular control over government'.
Building on the above, this paper utilizes the concept of Critical Public Accountability (Smyth 2012) to frame the case study below. Critical Public Accountability seeks to conceptualise the impact of neo-liberal restructuring processes and the resistance they generate from civil society. Critical Public Accountability stands at the nexus of the public accountability and the neo-liberalization of public services. Moreover, it captures the role that civil society (Lavalette and Ferguson 2007;Sader 2002) campaigns have in holding the state (Miliband 1967;Catchpowle, Cooper, and Wright 2004) to account and influencing the nature of public accountability relations (Smyth 2012). In the case study below, civil society is represented by a local campaign group (encompassing tenants, local councillors and anti-privatization campaigners). The group, Mumbles Defend Council Housing, sought to illustrate (among other issues such as increased rent levels) the impact the transfer would have on public accountability relations. These two aspects of restructuring and resistance are explored in the following sections.

Restructuring public accountability
In Britain, the post-war consensus drew on the representative democratic tradition to exercise public accountability. Public accountability was, and remains, expressed through the ballot box in periodic elections for both national and local governments. However, the election of the Conservative government in 1979 brought a challenge to that consensus, with state-provided public services seen as a barrier to the efficient operation of resource allocation through market relations. This led to a range of neo-liberalinspired policies that moved public services away from a direct (vertical) delivery model. Instead, the replacement institutional arrangements (contracting out, agencification, privatization) have led to an indirect (horizontal) model of delivery, leading to a mismatch or disconnect with public accountability relations (Michels and Meijer 2008, 171).
In addition, in this neo-liberal restructuring process Harvey (2005) has identified a suspicion of democratic relations and a preference for governance by experts and elites. 4 The case study below highlights this move from democratic public accountability relations to an HA board comprised of experts and elites, creating a public accountability deficit. The concept of a public accountability deficit is adapted from Smith et al.'s (2006, 159) democratic deficit, arising from partnership working between the public and private sectors, where they argue that there is a democratic deficit in terms of the rules and procedures of public governance when measured against a benchmark of elected local government. Partnerships are in, but not of, the community.
As has been argued above, public accountability's main function is democratic control, thus allowing the democratic deficit to be read as a public accountability deficit. Further, the idea that 'partnerships are in, but not of, the community' is also relevant to the case study below, where the CHM is presented as being a community organization with little actual engagement from that community.
HAs occupy a quasi-public sector, i.e. 'public sector organizations in private sector clothing' (Collier 2008, 933), with a blurring of the boundary between public and private sectors (Shaoul, Stafford, and Stapleton 2012). This quasi-public sector is a product of the rolling back of state funding and direct provision, coupled with a rolling out of privatization solutions (Peck and Tickell 2002;Poulsen 2009).
In social housing, this restructuring dynamic has brought into existence a new form of NGO, the transfer HA. The basic transfer HA is a non-profit-distributing, private company and is incorporated solely for the transfer process. In this context, its main objective is to borrow funds from the private sector to pay for repairs and maintenance of the transferred housing. The funding, governance and accountability relations of transfer HAs are explored in more detail below. At this stage, the accountability relations of these new NGOs are identified utilizing Ebrahim's (2003) work and his differentiation of accountability among NGO types. Ebrahim (2003) identifies three NGO typesmembership, service and network organizations. HAs fall within the category of service organizations (although as the case study illustrates there is an attempt to mobilize aspects of a membership organization). Service NGOs' orientation is 'charitable in the sense that there is no profit motive' (Ebrahim 2003, 205) with their beneficiaries not involved in its creation and remaining external to the organization.
In comparing the two general forms (public service provider and NGSO provider) of accountability relations, there are two major differences evident in Table 1. First is the downgrading of the role of tenants (en masse) who move from being voters who possess the power to change their landlord at each set of elections to being members of the transfer HA. In the process, both their formal accountability (through the ballot box) and informal (through everyday accountability (Smyth 2013a) processessee below) is downgraded or lost completely. Second is the introduction of the private financiers as a forum (Bovens 2005) that demands accountability. This is evident in social housing, where '(n)ew accountabilities have been introduced as a result of the significant levels of private borrowing which have funded stock transfer(s) . . . Lending covenants and credit ratings have become important foci of accountability' (Pawson and Mullins 2010, 166).
These two themes are explored further in the case study below.

Resistance
However, the reforming dynamic has also generated resistance as competing conceptions of public accountability clash and in part lead to the mobilization of civil society campaigns aimed at keeping the housing stock under local government control. Resistance, by public-sector managers and members of the caring professions, to neo-liberal reforms has been identified by others (Broadbent and Laughlin 2013;Kamuf 2007;Laughlin 1996). This is a resistance that adapts to the imposition of increased managerial accountability through buffering techniques or accepting colonization (Broadbent and Laughlin 2013). Ahrens and Ferry (2015) focus on the impact of grass-roots campaigns resistance on accountability and budgeting during the cuts in public funding under the UK's Coalition government . This case study also focuses on resistance from outside of the public organizations studied by adopting Critical Public Accountability as a framing, which identifies civil society as a site of resistance (Harvey 2005, 78) to the neo-liberalization of public services. To fully comprehend the resistance exhibited in the case study, it is necessary to understand the concept of everyday accountability (Smyth 2013a). It was argued above that public accountability is essentially democratic with the reward/sanction element being exercised through elections. Everyday accountability is a corollary to the formal accountability of the ballot box, covering the practices and relations that allow the people (including tenants) to influence their elected representatives to achieve specific aims (such as timely repairs). For example, in 2009 it was everyday accountability relations that allowed tenants and other civil society groups to pressurize the British Housing Minister into reducing a proposed rent rise by half (Rogers 2009). Everyday accountability is an example of direct and informal accountability relations (Romzek 2014), which is undermined and lost through the stock transfer process.
This section has argued that public accountability is essentially democratic in nature, and therefore, it is appropriate to analyse the stock transfer process from the perspective of tenants by utilizing the concepts of Critical Public Accountability and everyday accountability. Further, the restructuring-resistance dynamic also

Housing stock transfers and NGSO
This section analyses the housing stock transfer policy, its roots and evolution in England before being adopted in Wales, before moving on to discuss the lines of analysis identified earlier. The first stock transfer was completed in December 1988, covering 4,650 homes in a southern English local authority (HCA 2012). The policy was developed as a way of leveraging private finance into the sector that has been starved of funds for over three decades. For example, the investment levels in housing by British local authorities fell by over 60 per cent, from £13,887 million in 1979/80 to £5,208 million 5 in 1999/2000 (Wilcox and Perry 2014, Table 57b). Central to achieving political acceptance for the policy is a classification of HAs as non-public bodies, under the accounting rules on what counts towards public sector net borrowing (PSNB) (Perry 2012, 27). Unlike the European Union and other European countries, the use of PSNB by the UK government means that local government organizations with housing stock are heavily restricted from borrowing (Pawson and Mullins 2010). This political decision has the consequence that housing policy became focused on transferring housing stock from local government to the HA sector. Therefore, finance becomes the primary driver in the housing stock transfer policy (Pawson and Mullins 2010;Smyth 2013a). This provides the basis for the line of analysis focusing on the role of finance in the stock transfer policy. However, the model pursued in England of using private limited companies, even if they are non-profit distributing, has been brought into disrepute (DCLG 2007;National Assembly for Wales 2001;Smyth 2013aSmyth , 2013b. This was caused by two factors. First, the transfer process has major power and resource asymmetries between supporters and opponents (DCLG 2007;HOCCHG 2009). Second, it generates a public accountability deficit from the tenants' perspective (Smyth 2013a). Awareness of these problems in England has been an outcome of the years of campaigning by civil society groups such a Defend Council Housing (which encompasses tenants and trade unions) and the House of Commons Council Housing Group (a group of backbench Labour and Liberal Democrat MPs). Significantly, the work of these groups at a local level has meant that in a minority of cases (estimated to be approximately a quarter of all tenant ballots) the tenants reject the transfer and therefore reproduce the existing democratic public accountability relations (Pawson and Mullins 2010).
When adopting the stock transfer policy, the Welsh Assembly Government (WAG) was concerned that there were ownership and democratic deficits involved with the English model (National Assembly for Wales 2001Wales , 2008. This led to the development of a Welsh Solution, CHM, where 'unlike existing models . . . not only is the business conducted for the benefit of the community, but the organisation is also owned by the community' (Waite 2008). The nature of the CHM organization in the case study below shows a combination of both membership and service organization characteristics (Ebrahim 2003). There is a complicated membership structure that involves a small minority of members (who must act in the best interests of the organization), 6 with a board composed of experts and elites. With the existence of the CHM predicated on the need to attract private finance, the implications for accountability are set across a tension between weak forms of accountability to tenants (in comparison to funders) and an organization structure that should enable membercentred accountability. This provides the basis for the line of analysis focusing on the governance and accountability structures of the CHM.
Given the foregoing, it is legitimate to assess to what extent the case study corresponds to this tension, with three empirical questions to the fore: • What competing conceptions of accountability (an 'accountable landlord') are evident? • How does a CHM governance and accountability structure correspond to the differing conceptions of accountability? • What role does (private and public) finance play in this neo-liberal restructuring process?
To summarize, these two sections have defined public accountability, arguing that the restructuring-resistance dynamic leads to contested conceptions of what an accountable landlord is. There has also been an explanation of how access to private finance is a driver for the policy and a recognition that problems with the English stock transfer model have resulted in the adoption of CHMs in Wales. Before addressing the above questions, the paper turns to the research design employed in developing the case study.

Research design
Critical realist ideas underpin the development of Critical Public Accountability (Smyth 2012). Critical realism provides powerful tools for analysing the changing nature of phenomena and specifically the role of individual/collective agency in reproducing or transforming social structures (Bhaskar 2008;Modell 2009). Two such tools used in the case study below are the conception of a stratified reality with our empirical experiences based on the emergent outcome of generative mechanisms (Modell 2009) and Bhaskar's contribution to the structure/agency debate in the form of the transformational model of social activity (TMSA). 7 Bhaskar bases his TMSA upon two triads, where societysocializesthe individual and, the individualreproduces or transformssociety. It is the recognition of human agency reproducing or transforming social structures that captures the role of civil society campaigns' impact on public accountability relations.
Further, empirical social events are emergent from the underlying interaction of competing generative mechanisms. For example, the stock transfer policy is a combination of long-term withdrawal of public funding, with the political decision to classify council housing as 'on-balance sheet' and therefore the need to raise private finance through transfer HAs.
This critical realist approach is coupled with a dialogical (Bakhtin 1981;Sullivan 2012) understanding of the nature of language, as tension-filled and stratified, 8 interacting with the concrete socio-historical environment (heteroglossia 9 ). Sullivan (2012) has developed a data analysis approach where 'key moments are an "utterance" of significance. An utterance is a significant unit of meaning . . . and is defined by its readiness for a reply/reaction' (Sullivan 2012, 72). Applying these ideas to the data collected, the central analysis technique used was to identify the key moments in each document/transcribed interview.
In keeping with a case study research strategy, the data were collected from a range of sources. These included from the local authority (Mumbles); from the new HA (Gower Homes); from the tenants' campaign group (Mumbles DCH); national reports from bodies such as the WAG; clippings from local and regional newspapers, and other documents issued by the independent tenants' advisor, and freedom of information decisions (see Appendix for more information).
These documents were supplemented by observation notes of an anti-transfer public meeting and interviews with key participant/actors involved with the stock transfer process. The interviewees included an anti-transfer campaigner (Hywel), tenant (Lydia) and councillor (Morgan) and a pro-transfer tenant and HA board member (Rees). The interviews were conducted on two occasions; first, during the campaign, and second, 6 months after the ballot result was announced. The emphasis in the interviews was not on their individual experience but on their agency as positioned-practices that endure beyond the individual (Collier 1994). In line with assurances given to the participants, their names have been changed, as has the names of the organizations involved. However, the case study in the main relies upon a dialogical analysis of relevant documents.

Background and context
The HA sector in Wales controls properties to a value of £4.3 billion, of which £317 million have come through stock transfers (Community Housing Cymru 2011, 19). However, the stock transfer subsector has a greater proportionate turnover at £227 million (compared to £373 million for traditional HAs) in 2011 (Community Housing Cymru 2011, 23). Despite the CHM model being advanced as a 'Welsh Solution . . . it should be recognised that only five of the first ten stock transfer associations . . . have been set-up on this basis' (Pawson 2009, 27-8). The Gower Homes case study below is a further example of this Welsh solution.
In 2011, Mumbles County Borough (hereafter, Mumbles) council transferred 9,207 homes to the newly established Gower Homes Ltd, a CHM. The tenants' ballot took place in 2010 with 56.6 per cent of tenants voting in favour of the transfer on a turnout of 61.6 per cent. 10 Mumbles covers two significant towns and numerous small villages. The area has suffered economic decline over the previous two decades as the traditional manufacturing and coal mining industries declined. In general, the housing was amongst some of the poorest in condition in Wales.

Competing conceptions of public accountability: what is an accountable landlord?
The stock transfer occurred after an acrimonious campaign between the anti-transfer Mumbles Defend Council Housing and a pro-transfer campaign group backed by local government, the WAG and Westminster government departments. During the course of the 10-month campaign, the two groups contested a range of issues including what constitutes an 'accountable landlord'. The data analysis highlights competing conceptions, so that when asked to explain what the term an 'accountable landlord' meant, Rees (pro-transfer tenant board member) stated: 'It means being accountable to tenants . . . I mean to transfer to this kind of organisation [a CHM] is a chance for the community actually to have a say in the housing and their area' and that '. . .it does mean involvement. It means tenant engagement'. In this respect, Rees is reflecting a justification for transfer also advanced by the Westminster government and WAG. For example, the Regulatory Framework for Housing Associations in Wales states as one of its principles that The Regulatory Framework puts tenants at its heart by: • Ensuring that housing associations work with their tenants and people who use their services in planning, and assessing the quality of, services. • Providing information that allows tenants and others to compare the performance of their housing association to that of others. WAG (2011, 3) Thus, accountability to tenants is reduced to transparency and answerability (Smyth 2012). It should be noted that the regulatory framework (WAG 2011) emphasizes the need for an effective and well-managed board, but at no point sets out how the board members are to be selected, nominated or elected. Therefore, for tenants the transfer represents a move away from democratic control of their landlord through local elections, to holding a much weaker position in relation to board members. In contrast, the question of an accountable landlord provoked the following discussion between members of the anti-transfer campaign, Morgan (Plaid Cymru councillor) and Lydia (tenant activist): MORGAN: So I believe the accountable landlord is councils. And we're moving away from it unfortunately. LYDIA: Democratically controlled. MORGAN: Democratically controlled by people, they elect members, if a member is not up to the job, out he goes in four years time. If the council is not up to the job, out the council goes in four years time. And that's the way it's got to be.

LYDIA:
And what I find appalling is that you are moving something out of democratic control. And they keep plugging into you all the time, '. . . but we're not for profit . . .', but that's not the point.

(Group interview with Lydia, Hywel and Morgan)
Lydia then outlined how she viewed council housing as an asset owned by the people that was being given away: 'And that, to me, is giving away accountability as well, isn't it? Because you've got no rights through your councillor or anything'. The above illustrates two relevant points; firstly, the importance placed on relations of public accountability as expressed through local elections. This is a common theme of anti-transfer campaigns across Britain (Smyth 2013a). Secondly, Lydia's contribution gives an insight into the motivation of those involved with anti-transfer campaigns; they are attempting to stop the loss of collectively owned assets and the rights they (as tenants) have to control those assets, their homes.
The publicity used by the anti-transfer campaign illustrates a conception of public accountability across two themes (Smyth 2012). First, there is a defence of the existing democratic forms, such that housing 'under Council control . . . is secure and democratic' 11 and that if the stock transfer goes ahead 'direct accountability of council landlords is lost'. 12 Second is a critique of the promised public accountability of the new landlord. This argument had an additional layer in Mumbles as the transfer HA was set-up using the CHM model, which Wales DCH describes as 'privatization with a fancy wrapper'. 13 They also raise criticisms that tenant boards members will have to act in the interests of the HA not as representatives of the tenants (CHMG 2007) and that tenants will only make-up a third of the board.
Wales DCH critiqued the form of control in the CHM as taken from private business with the exception that (1) the tenants are members and (2) the members only have control over a third of the board. 14 These themes were combined in a Mumbles DCH leaflet publicizing a local public meeting, which under the heading '6 Reasons to Reject the Transfer Offer', stated: No accountability. Council tenants can vote for their landlords in local council elections every four years. Registered Social Landlords are simply not accountable to us in the same way. 15 The above examples illustrate that there are opposing and contested concepts of public accountability; one based on democratic ideas of accountability, the second based on a more limited view of accountability based on private sector governance structures.
By drawing on Bakhtin's (1981) work, it is possible to see these two competing meanings as rooted in the concrete social relations of neo-liberalization and the reaction that process has generated from tenants and other civil society actors (Smyth 2012(Smyth , 2013a. Also drawing on Bhaskar's TMSA model, it is clear that the question of changing accountability is one of the reasons that were deployed to motivate and mobilize tenants (and others) to oppose the stock transfer. For both these reasons, accountability becomes a contested arena. The outcome of this contestation then leads to democratic public accountability relations being reproduced, if a 'no vote' is returned. Or, as in this case study, public accountability relations being transformed along NGSO lines (Ebrahim 2003).

Changing governance structures and accountability relations
In addition to other objectives, the neo-liberalization of public services is designed to change accountability relations; in this case, moving from a housing committee with democratically elected councillors to a board running the transfer HA. The following illustrates how the new accountability relations involving the board were designed and have operated in the first two years.
In common with the English model, the board of Gower Homes Ltd is made up of three constituenciestenants, local authority nominees and independents; each constituency has a maximum of the five members. As a CHM, Gower Homes is supposed to have a stronger and closer relationship with its tenants as they become the owners (Waite 2008). However, establishing the CHM has proved to be more complicated in practice. First, tenants did not automatically become members, they had to apply to join and pay a nominal fee of £1. Second, Gower Homes has five different classes of membership; besides tenant members, there are leaseholders, independents, local authority and associate members. 16 The independent members are admitted on the basis of selection criteria set by the board. Mumbles council as the former landlord is the only local authority member. 17 Associate membership is open to anyone who is 16 years or older, lives or works in the local area 'and qualifies for membership in accordance with any membership policy formulated by the Board and approved by the Members' 18 ; this includes corporate bodies.
Third, with five constituencies there are complications and inequities in relation to voting and decision-making by members. For example, to change the rules or amalgamate with another association, there is a weighting of votes in each class of membership, with the total tenant membership receiving 34 per cent, the same amount as the local authority member. 19 For other decisions, each member shall have one vote. 20 Fourth, there is a trigger point where the election of tenant board members moves from a ballot of the whole tenant population to a vote of the tenant members only. This occurs when 5 per cent of all tenants become members of Gower Homes. 21 The actual experience of bringing these rules to life has illustrated a number of difficulties. It took nearly a year after the transfer before tenants could apply for membership. 22 In March 2012, Gower Homes issued a press release celebrating the coincidence of the ninety-ninth member, being 99 years old. 23 At this rate of registration (i.e. 100 members per quarter), it would take over 350 years for all tenants to become members of their landlord. 24 It is clear that gaining acceptance among tenants for the new landlord through mass membership has not been a priority for the CHM management.
Further, there has been a significant turnover of board members, particularly among the tenant members constituency. The initial publicity sent to tenants in January 2010 showed fourteen members of the proposed management board (the Independents constituency had only four members). 25 By the spring of 2012, only eight original board members were left. The most affected constituency was the tenant board members, with the chairperson being the only member still in place. 26 Further, by March 2012 there were only two tenant board members, with three tenants resigning from the board in December 2011. In personal correspondence, the press officer of Gower Homes stated the main reason for the turnover of tenant board members was ill-health.
The foregoing analysis illustrates that while tenant involvement and participation is mobilized as a key benefit of the transfer process, the planned structures are designed to involve a small minority, elite group of tenants. This elite group is combined with independent board members who are appointed on the basis of their experience and expertise in areas relevant to running social housing as a business. In 2014, the four independent board members were an accountant, a solicitor, a chartered surveyor and a retired local government officer. This aptly illustrates Harvey's (2005) conception of governance by experts and elites, where the previous democratic accountability relations are replaced by a group of elite tenants and independent experts. This is not a matter of inadequate policy implementation at a local level but is a common theme across other stock transfers (Smyth 2012(Smyth , 2013b.

A Trojan horse for private finance
The case study also highlights the manner in which finance is used as a generative mechanism (Bhaskar 2008) that drives the stock transfer policy. The pro-transfer publicity repeatedly made statements that given the current funding restrictions a transfer was the only option. The role of finance (public and private) has a dual impact. First, the existing council housing stock has suffered from the roll-back policy of underinvestment for three decades causing a backlog of repairs and maintenance. Second, the roll-out stock transfer policy is made attractive because it is underwritten by public money. Both these aspects are evident in this case study. A stock condition survey carried out for Mumbles council in 2004 estimated that £152 million would need to be spent over a 6-year period to achieve the required housing quality standard. 27 In contrast, the local authority estimated it could only afford to invest £40 million over the next six years. 28 These figures give an indication of the cumulative impact of underinvestment over three decades.
The stock transfer Offer Document also contains justifications for the proposed transfer, including the extent of public subsidy. These justifications include the ability of HAs to borrow outside of the PSNB. 29 In addition, there were two specific and large public subsidies. First, HM Treasury agreed to pay off the historic debt, estimated to be £35 million. 30 Second, even with the debt write off, the housing stock had a negative transfer value (i.e. the cost of the needed repairs and maintenance is greater than the discounted future rent cash flows). In these circumstances, a dowry (additional gap funding) from the WAG was needed. The Gower Homes Business Plan (January 2011) shows an annual dowry of £6.2 million for 27 years from 2012. 31 The extent of the private finance needed was not publicly known at the time of the ballot, but was disclosed in the first set of financial statements of Gower Homes Limited for the period ended 31 March 2011: A loan facility of £135 m has been arranged with The Royal Bank of Scotland, Santander and Nationwide to enable [Gower] Homes to undertake a major improvement programme to our housing stock. £3.5 m of this facility has been drawn down . . . These loans are secured through a Prudential Security Trustee on all the housing stock.
The business plan for Gower Homes shows the total amount of debt peaking in 2016/ 17 at £130.4 million, with nil debt and cash reserves of £47.5 million by 2037/38. Effectively this cash reserve forms another public subsidy.
It was noted above that the stock transfer policy introduces another accountability forum in the form of the private finance providers. There is little publicly available information on the accountability relations covering the £135 million loan facility; however, the financial statements for 2014 include the following: The loan facility is measured against covenants (including net debt per unit, cumulative cash flow deficit and value of loans outstanding compared to the value of housing stock) whose output levels are set by funders' approval of the 30 Year Business Plan that is updated each year.
Users of the accounts are not privy to what the actual calculations are, but must be satisfied that performance against the covenant ratios has been 'audited and signed off'. In this manner, accountability to the finance providers is emergent (Bhaskar 2008) from political and policy decisions and erodes public accountability to the tenants.
The previous sections have presented and analysed the case study addressing the three issues of competing accountability conceptions, their relationship to the governance structure of a CHM and the role of public and private finance. The case highlights the contestation over different conceptions of public accountability among those involved. The new CHM, despite claims of democratic relations and tenant involvement, actually resulted in complicated governance and accountability structures that replaced democratic-styled public accountability. Finally, the driver of this whole process was the deployment of public finance to attract private finance. The time frame for this case study did not extend to the post-transfer period where the planned programme of maintenance and repairs had commenced. If Gower Homes were to follow the experience of most stock transfers, then the new finance will improve the living conditions of the tenants (Pawson and Mullins 2010;Smyth 2013a). This raises other questions such as the desirability of changing democratic public accountability in the pursuit of private finance and whether it is possible to improve the council housing stock within the existing (or even a strengthened) framework of democratic public accountability. These questions are explored in the final section.

Discussion and conclusion
There are three main conclusions that can be drawn from the foregoing case study. First, the neo-liberal reform process is contested; this has also been found by others such as Broadbent and Laughlin (2013) and Laughlin (1996). This case adds to those findings by extending the groups who are resisting the reforms to include civil society campaigns. Second, the reform process has serious limitations when considering public accountability from the perspective of the tenants. The new CHM organization possesses a form of governance by experts and elites rather than democratic public accountability. Third, the major driver of this whole process is the need to access private finance. Each of these conclusions is discussed below.
The resistance found in this case study is consistent with the conceptualisation of Critical Public Accountability, where civil society seeks to challenge and control the state (Smyth 2012). The Mumbles DCH campaign emerged as a reaction to the neo-liberalization of council housing. As argued earlier, the transfer policy is in turn emergent (Bhaskar 2008) from the government's political decisions to use PSNB and the sustained withdrawal of public funding for council housing. Thus, there is a series of generative mechanisms linking together to generate the contestation in the case study. Further, this formation of generative mechanisms, in previous stock transfers, highlighted problems with the policy and in a significant minority of cases led to the tenants rejecting the proposed transfer. In this way, the case study reflects the first conclusion stated above.
The development of a 'Welsh solution', CHMs, was supposed to overcome these deficiencies, including the public accountability deficit. The evidence in this case study raises serious questions about how successful the CHM model has been in this regard. These questions concern both design and implementation issues. Gower Homes describes itself as a CHM but tenants have not automatically become members as the WAG guidance states they should. There is slow progress in implementing the tenant membership structures. This apparent lack of engagement by tenants is not a new phenomenon in stock transfers. McKee (2009), in assessing the experience of tenant engagement in the 2002 Glasgow transfer, found that it 'was not an important priority for tenants at the point of transfer' (McKee 2009, 299). In these circumstances, the use of the CHM appears to be little more than a ploy to deflect criticism away from the undemocratic outcome of the neo-liberal reform process.
Both pro-and anti-transfer campaigns advocated greater tenant participation, although with substantially different views on what it looks like. The lack of tenant engagement with participation structures (such as membership) is a concern for those who advocate greater democratic public accountability. Not on the basis that mutual membership is equivalent to pre-existing forms of public accountability, but that in the absence of an active tenant group who will hold their landlord to account, the New Labour mantra of 'what matters is what works' (Pawson and Mullins 2010) gives way to governance by experts and elites (Harvey 2005).
Also, there is another more fundamental logic to the lack of engagement: community ownership already exists, it is called council housing (Glynn 2007). This mindset was illustrated above by tenant activist Lydia. It follows that NGSO accountability is not equivalent to democratic-styled public accountability. Especially from the perspective of the tenants and therefore represents an undesirable direction for a policy that claims to encourage greater tenant participation and empowerment. The foregoing reflects the second conclusion stated above.
The case study also enables a reflection on the distinction that Ebrahim (2003) makes between service and membership NGOs. Membership NGOs are run by and for members, for the advancement of a particular aim (e.g. agricultural cooperatives). In housing, such cooperatives exist, for example, in Brazil (Fruet 2005) but tend to be small and localized in Britain (Pawson and Mullins 2010). The Mumbles case shows an attempt to graft membership characteristics onto what is a service NGO. This is not to argue that mutuals or cooperatives have no place in providing social housing; they do but not as a substitute for democratically accountable council housing. For example, if Gower Homes had been the product of a group of committed tenants with the view to democratizing housing services by increasing direct control and participation, this would be a different matter.
This leads to a further reflection drawing on the metaphor of a web of accountability (Romzek 2014;Smyth 2013a). Table 1 sets out the changing nature of this web, with two points becoming apparent. First, the tenants as a group become more passive in their relationship with their housing provider; from having a position of power exercised through the ballot box at each local election, they are only able to change a third of the board members if they are members of the CHM. Second, the NGSO has introduced private financiers as an accountability forum. The financial institutions have a legal charge over the assets of Gower Homes, and exercise accountability through debt covenants. In these circumstances, the effective accountability relations are weakest for the tenants/members when compared to the finance providers (Mooney and Poole 2005). This reflects the third conclusion stated above.
This case study also has implications for practitioners and policymakers. For public management practitioners, the Mumbles case highlights a missed opportunity by those in charge of the new CHM in not generating greater participation from the tenant body. Clearly it is not enough just to have the rules in place, there needs to be a comprehensive and resourced participation strategy as well. As for policymakers, the use of CHM in Wales is an attempt to improve a discredited policy design. This has not been a success. First, because only half of the initial tranche of stock transfers used the CHM model; and second, because, as this case study shows, there is still a public accountability deficit. Here, Mooney and Poole's (2005, 38) conclusion appears supported: There is at the heart of all of this an irreconcilable conflict between the needs of private financiers and the pursuit of meaningful tenant empowerment and quality social housing provision that is democratically controlled.
A limitation of this research concerns the generalizability of a single case study's findings, especially where the CHM model has such a low level of adoption. This criticism somewhat misses the key point, the CHM model was specifically developed to overcome identified deficiencies in non-mutual stock transfers. The evidence in the case study shows that this has not been achieved for design and implementation problems. These last two points generate two possible future research questionswhat has been the experience of tenants membership in the other CHM stock transfers in Wales; and to what extent and how has the need for acceptability to private financiers shaped the CHM model?
Finally, based on the experience of this case study CHMs are not vehicles to renew democratic public accountability. CHMs are instead a mechanism for leveraging private finance and are embedded in the neo-liberal reform agenda. In this form, mutuals are in grave danger of being apologia for private capitalism.
One argument against this conclusion is that there was no other alternative to accessing the investment needed. There are two responses to this. First, implicit in the statement is a recognition that something has been lost (e.g. the creation of a public accountability deficit). Some tenants, indeed a majority in this case study, were comfortable with this outcome; even if those involved with the civil society campaign were not. Second, and more significantly, other local authority housing has survived and thrived (relatively) while being retained. For example, council housing tenants in Birmingham City Council (BCC) twice voted against stock transfer in 2002 and 2006, and yet by 2011 they had achieved 99.6 per cent compliance with the Decent Homes standard. Between 2002 and 2009 regulatory inspection ratings had improved from zero stars to three stars (out of a possible four) . . . In the decade since the first rejected transfer, BCC had reduced the numbers on the waiting lists by over 1,000 despite the impact of the global financial crisis post-2008. And significantly in the past four years there has been new council house building in Birmingham on a level not seen for three decades. Smyth (2013b, 21-2) Therefore, the trade-off implicit in the stock transfer policy that tenants must suffer a downgrade of their accountability if they want to live in a decent, affordable home appears to be a false one. Notes 1. This policy is known as large-scale voluntary transfer and requires the tenants to vote in a ballot on whether they agree with the transfer. 2. In October 2015, the Office for National Statistics unexpectedly announced that housing associations would now be considered public organizations and therefore their debt was added to the government's balance sheet, backdated to 2008. However, at the time this stock transfer took place there was no indication this reclassification was a possibility. 3. While transparency (openness) and answerability (account-giving) are necessary parts of an accountability process, they do not exhaust the whole meaning (Ferry, Eckersley, and Zakaria 2015;Smyth 2012). In public management and policy discourse, they are often used to limit the extent of accountability to a relationship less than one of control (or 'reward and sanction' as Gray and Jenkins 1993 put it).

Disclosure Statement
No potential conflict of interest was reported by the author.

Funding
The author acknowledges the financial support of the Irish Accounting and Finance Association (IAFA) in carrying out this research.

Notes on contributor
Stewart Smyth has worked in accounting roles in practice, industry and the public sector. He is currently a senior lecturer in accounting at Sheffield University Management School, specialising in public sector accounting and accountability, and financial reporting.