Exploring the meso-territorialization of third sector administration and welfare delivery in federal and union states: Evidence and theory-building from the UK

ABSTRACT The international trend of state restructuring and the rise of decentralized welfare systems means a key challenge for social research is to systematically explore the breadth of factors shaping the territorialization of third sector welfare delivery at the meso level in federal and union states. We address this lacuna by synthesizing historical-institutionalism and critical realism with Salamon and Anheier’s classic framework on civic infrastructure development to produce an inductive analytical model for wider empirical testing. Its application here to the longitudinal case study data covering Wales shows it to be effective in providing a holistic understanding of the temporal and spatial processes underpinning decentralization. The wider significance of the case study lies in underlining the iterative, reciprocal relationship between governance reforms and territorialization – and showing how territorialization can originate in response to national crises and welfare demand caused by state and market failure in the delivery of public goods.


Introduction
In response to an international trend of state restructuring and devolution a burgeoning literature has sought to analyse the decentralization of the welfare state (Borghi and Van Berkel 2007). Existing studies have tended to examine discrete factors driving this processsuch as marketization (López-Santana 2015), neo-liberalism (Bergh 2012); governance reforms (Pavolini 2015); political pressure from minority nationalist parties (Chaney 2017); and resistance to federal social programmes (Mooney and Scott 2012). However, a challenge for social research is to systematically identify and explore the breadth of factors that shape the territorialization of the third sector welfare delivery at the meso level in federal and union states.
Here we address this lacuna through theory-building and empirical case study analysis. Our aim is to engage with the literature on welfare state decentralization with reference to voluntarismand develop an analytical model for wider application to the empirical analysis of third sector welfare territorialization. In epistemological terms this aligns with the concept of holism (Verschuren 2001) and systems analysis in social science (Bailey 2001).
Specifically, we synthesize neo-institutionalism and critical realism with Salamon and Anheier's (1996) classic framework on civic infrastructure development. We use this to produce an inductive analytical model of contingent factors shaping the meso territorialization of third sector welfare. In the second half of the paper, using a longitudinal dataset spanning nine decades, we apply the model to a case study of Wales. In this way the current study responds to earlier calls to explore the divergent histories and geographies of voluntarism (Crowson 2011). Our core finding is that welfare decentralization is driven by discontinuity at critical junctures related to governance transitions (phases of devolution), national crisis (war) and political shifts (Thatcherite reforms, and the partnership rhetoric of government given expression in the devolution reforms of the 1990s).
The remainder of this paper is structured as follows: first, the analytical model is outlined. Following an outline of the research context and methodology, the model is then applied to the case study. In the concluding discussion we reflect on the model's utility and wider significance as an analytical tool as evidenced by the case study analysis.
An analytical model of contingent factors shaping the mesoterritorialization of third sector administration and welfare delivery in union states (Rodriguez-Pose and Gill 2003). In definitional terms, it should be noted that union states are characterized by the incorporation of parts of [… their] territory at different times … reflected in an institutional infrastructure providing some degree of meso-autonomy to particular areas and asymmetric political and social rights; this may coexist with administrative standardisation over most of the territory. (Kay 2005, 544) In order to develop our analytical model in the following section we first build on Salamon and Anheier's (1996) classic framework on civic infrastructure development to summarize the diverse contingent factors at play in the meso-territorialization of welfare. This is followed by an explanation of the second component of the model, critical junctures. During these historical moments of political and governance reform it is the prevailing alignment of contingent factors that combines with critical junctures to disrupt the path-dependent logic of centralized administration, thereby shaping the emergence of meso third sector welfare.
Contingent factors promoting the territorialization of the third sector welfare at the meso level in federal and union states Salamon and Anheier's (1996)   state. They include the degree of unitary state centralization, intergovernmental relations in multi-level states, and market efficacy in the provision of public goods. Organizational factors include welfare state configuration, the density and strength of associative third sector networks, and the presence of social entrepreneurs. Lastly, socio-economic and political factors include: the strength of national identity/ nationalist mobilization at the meso level, the 'regional' prevalence of social inequalities, and economic prosperity. Further details of these factors and associated literatures are outlined in the latter half of this paper when they are applied to the case study data. Before this, we turn to consider social theory and the mechanism at the heart of the model: critical junctures.

Critical junctures and contingency
In explaining the relationship between the contingent factors and critical junctures in our analytical model we draw on critical realism and historical institutionalism (See Steinmo, Thelen, and Longstreth 1992). These conceptual strands support the model's core thesis: that the interplay and co-presence of contingent factors at key historical moments or critical junctures helps to explain third sector welfare territorialization.
As a social scientific approach the proposed model has its grounding in contingency theory. As Walzer (1990, 32) notes, norms around meaning and understanding of the social world can only be validated when there is cognizance of local, contextual phenomena. It is a view expressed in the literature of critical realism: a theoretical approach to social science that seeks to identify causal mechanisms in social processes that, in turn, enable critique of social arrangements and institutions (Sayer 2000). As Gerrits and Verweij (2013, 172) explain: 'reality is contingent. This means that any explanation [of social processes] is temporal in time and local in place. Since systems are nested within their systemic environments, there is mutual influence between different systems'. Importantly, the evolutionary theory strand of historical institutionalism argues that contingent factors are not static. They should be viewed as evolving drivers of institutional change. The mere existence of contingent factors does not, in and of itself mean that territorialization of the third sector will be actualized. There is no inductive certainty involved and the straightforward co-presence of a core set of factors or preconditions applied across polities will neither deliver universal or uniform results (Sayer 2000). Rather, territorialization is a function of what historical institutionalists dub critical junctures. These are key points in time when contingent factors align to disrupt path dependencyor, 'the dynamics of self-reinforcing or positive feedback processes in a political system … [Involving] mechanisms that reinforce the recurrence of a particular pattern into the future … ' (Skocpol and Pierson 2002, 6).
Applied to the present case, critical junctures break the path-dependent logic of centralized third sector administration in union states and lead to the emergence of new meso systems. Often critical junctures arise from state actions in response to significant external threats (such as war)or major political initiatives (such as those designed to reform governance and democracy). As Thelen (1999, 390) observes, attention to contingency involves, 'close examination of temporal sequences and processes as they unfold … [It] focuses on variables that capture important aspects of the interactive features of ongoing political [and social] processes, and explains important differences in regime and institutional outcomes'.
This approach is relevant to contemporary study of third sector administration for the following reason: it underlines that any attempt to understand social processes in a given civil society setting needs to be cognizant of how the social space for action is at once shaped by the historical development of the polity and also the areal qualities specific to different geographical localities (inter alia, prevailing notions of national identity, patterns and processes of voluntarism, and the existence of inequalities). In ontological terms this has powerful implications for the study of welfare decentralization. It means empirical work must take into account a full range of contingent factors and critical junctures that influence and shape territorialization and re-spatialisation.

Research context and methods
The case selection rationale is founded on a number of factors: 1. in a global context, Wales is typical of other historical nations and regions absorbed into (quasi-)federal and union states, making it appropriate to regional studies and exploring the territorialization of third sector welfare (Royles 2006); 2. The UK features strongly in the empirical literature on welfare state development (Anttonen, Häikiö, and Stefánsson 2012), but much of this continues to promulgate the fallacy of a single welfare state; 3. From an international perspective, Wales is a propitious research context for it has played a key role in the development of social welfare, notably the development of state healthcare; and, 4. The present analysis is facilitated by a rich longitudinal dataset spanning nine decades.
Space constraints do not allow for an exhaustive account of the socio-economic and historical characteristics of the case study polity. Yet key features of this country of three million inhabitants include its language and culture. It is one of three nations that make up Great Britain. With the province of Northern Ireland, they constitute the United Kingdom (UK), an example of a 'Union State'. Wales' legal code and indigenous institutions were largely pushed aside as it was incorporated into the legal and administrative structures of England in the sixteenth century (Rawlings 2003). In consequence, prior to the twentieth century, the Welsh language and culture were the principal markers of national distinctiveness (Morgan 1990). Subsequently, civic nationalism and governance shifts have seen the constitutional 're-emergence' of Wales as a distinct polity. Today it has a national parliament with broad policy responsibilities and tax-raising powers. Against this backdrop, the country's third sector is comprised of 33,000 third sector organizations (TSOs) (NCVO 2018); with international comparative analysis concluding that 'civil society in Wales is relatively well developed, with a strong value base. It operates in a generally supportive environment, and has a high impact on society' (Civicus 2005, 8).
In methodological terms, the current study presents corpus analysis of the annual reports of the principal representative bodyor 'council' of the third sector in Wales 1934-2018. Today it is called the Wales Council for Voluntary Action (WCVA). Its strategic priorities include good governance, influencing and engagement, volunteering and resourcing a sustainable voluntary sector. Membership of the Council is made up of charities, voluntary and community groups, and social enterprises. The Council's annual reports comprise our case study data set. They summarize the activities of the Council over the preceding year. Typically 5,000 words long, they provide a rich social history and chronicle the development of voluntarism in Wales.
The discourse analysis was operationalized using content analysis to examine issue-salience and framing. The latter tells us about the inherent textual meanings and messages in the corpus. Archive hardcopy documents were scanned and converted to searchable text. The corpus was coded using an inductive coding frame based on key frames derived from the literature on third sector representation and administration (see online supplementary file). It was analysed using the UAM CorpusTool 3.3h©.
Case study: exploring the meso territorialization of third sector administration and welfare delivery In this section we apply the inductive analytical model to the case study data. Attention first centres on the contingent factors. As Figure 1 and the earlier discussion shows, these fall into three spheres. The second part of the discussion is based on close textural analysis of the longitudinal data and identification of critical junctures.

Contingent factors
In accordance with the analytical model the following discussion is structured around consideration of each sphere of contingent factors: governance, organizational and socio-economic/historical.

Governance
The political opportunity structures for the emergence and development of meso third sector systems are shaped by the degree of unitary state centralization ( Figure 1, factor 1a). In turn, this links to political ideology (and the extent to which dominant parties in central government are supportive of unionism and centralized administration), as well as institutional structures and legal and administrative practices (and whether these are conducive to reform/decentralization). The present case study evidences how the Council actively pressed territorial claims upon central (unionist) government at Westminster. For example, we make a plea for reasonable recognition … the Council has circulated all Departments in Central Government and all Welsh MPs [Members of Parliament], and it is hoped that the facts presented will help organisations in Wales to continue their work … the Council will persist in pressing the claims of legitimate movements who are doing an immense amount of work for the welfare of their fellow citizens. (CSSW 1979, 6) Allied to the foregoing, a further contingent factor is intergovernmental relations (1b)or, the relationship between different tiers of governance (inter alia, municipal and regional government, and the European Union) and the extent to which they co-operate and accrue power to challenge dominant patterns and processes of centralized government and administration. The present case study data reveal the shifting opportunities for territorialization across the decades. For example, 'the post-war world faces new conditions which demand a reconsideration of the scope of voluntary effort. We should act with the central government and the local authorities. To these ends voluntary organizations must be … co-ordinated. This we conceive to be an important function of the Councilto focus voluntary effort and to be its clearing house in Wales' (CSSWM 1947, 5).
Trust in government (1c) is a predictor of exogenous organizations' policy engagement (Mishler and Rose 2001). It is also key in shaping welfare territorialization for, as Braithwaite and Levi (2003, 240) observe, 'in the case of the levels of government, proximity would appear to be a major factor in helping to determine the differential foundations for establishing and maintaining trust'. Thus, when applied to the development of meso systems, if third sector organizations and volunteers exhibit greater trust in local governance and mistrust central government (e.g. as being alien or neglectful of local interests), then the propensity to press for government decentralization increases. The present case study supports this thesis. For example, 'there can be no doubt that Wales does suffer in comparison with other parts of the United Kingdom in the support it receives from [central] Government for Voluntary Work' (CSSW 1979, 6).
Analysis of the case study data also shows that, as devolved government institutions develop, there is increasing engagement and sectoral scrutiny of government at the meso level. For example: In order to influence how the Welsh Office interprets and implements the report [on third sector funding], the WCVA held a briefing session for voluntary organisations on the content and implications of the scrutiny, in order to stimulate informed debate. It will form the basis of a Welsh voluntary sector response and discussion with the Welsh Office. (WCVA 1990, 40) This is significant and underlines a key transition over time as the council of the Welsh third sector moves from being a recipient, to provider of influence. Hitherto, political influence had almost always come from central (or sometimes local) government. This transition is driven by the strategic leadership of the Councilas well as specific groups of TSOs conducting advocacy activities. This shift is such that statements of the Council's lobbying become routine in its annual reports. This influencing role develops apace with the (re-)establishment of a Welsh legislature in 1999 (see below). For example, the Council has been lobbying for resources to build the capacity of third sector service providers, leading to commitment in the Strategic Action Plan for the Voluntary Scheme, The Third Dimension, to establish an Invest to Serve fund to build the sector's capacity to deliver public services. (WCVA 2008, 19) Heterogeneity theory (Weisbrod 1977) and classical economic thinking argues that both markets and governments may sometimes struggle to provide 'public goods' (e.g. healthcare, housing, education etc.). When the market fails (1d) to provide these (owing to lack of incentives, weak competition etc.) then government may step in and provide such services (1e). However, its propensity to do so depends on demand and the extent of market failure, the political ideology of the party holding power, and the level of taxation/available revenue (which, in turn, determines the level of welfare spending). Crucially, the larger the gap between the general demand for 'public goods' and the extent of welfare provision, the greater the likelihood of third sector organizations stepping-in and providing such services. When state/market failure is evident in particular localities, such as in the provision of welfare services in the heavily industrialized areas of Wales in the 1930s, then it may spur third sector intervention leading to the development of meso systems. For example, In a land and a day in which local income for social services is at its lowest ebb, when the very fabric of social institutions is rotting for lack of paint and repair, and when the flower of our young manhood, with all its potentialities of leadership, is leaving us in a steady flow. The ensuing pages record the work of the South Wales and Monmouthshire Council of Social Service, in the past year, to help in facing these overwhelming problems … Voluntary Social Service is … the nurse of a sick society ministering to such needs as it can, proffering fellowship in an hour of distress, and, through it, planning for the fuller use of new health when at last it comes (SWMCSS 1935, 4).
Allied to the foregoing, welfare state configuration (2a) is a key factor shaping third sector welfare delivery. Traditional welfare state theory (Esping Andersen 1990) gives limited attention to the third sector. When it does, it is often regarded as a residual category. According to this view, welfare state expansion is a (by-)product of economic development (Flora and Alber 1981). As Titmuss (1974, 34) observed, countries that have not extended provision of public goods to the degree that their level of economic development might suggest may be dubbed 'residual welfare states'. Crucially, such states continue to rely both on market and non-profit solutions to social welfare problems. Moreover, when (as in the present case study after 1964) welfare states are administratively configured with an element of decentralization, then there is a distinctive spatial dynamic at work in the mixed economy of welfare. The degree to which welfare is a residual category at the meso level may then act as a spur or contingent factor leading to the development of territorialized third sector welfare provision. This is supported by the case study data. For example: Economic problems which seem to be worsening … The new Government has also said that it is going to encourage volunteers to fill some of the gaps that are created through cut-backs … it is essential to emphasise that this is a time when there can be a meaningful dialogue [… with] the Voluntary Movement on how the citizen who is disadvantaged can be helped in a practical manner. This means that the Voluntary Movement is given its rightful place alongside Statutory Social Service Schemes (WCVA 1979, 6).
'Social origins' accounts of third sector development (Salamon 1970) indicate that the emergence of meso third sector systems is also contingent upon the density and strength of associative networks in a given locality (2b). The present case study supports this and evidences how local TSOs act as catalysts of voluntarism and philanthropy. Notably, this is shaped by strong traditions of trade unionism in the industrialised areas of Walesas well as religious observance (specifically, strong networks of non-conformism). Both act to support the development of the meso third sector infrastructure. For example, 'the Council of Social Service for Wales is now represented by those concerned with industry [and] Trade Unions … There is no substitute for an energetic and organised voluntary movement … In co-operation with all those involved in this work, the newly reconstructed Council is, I feel, ready to play a vital and central role throughout the Principality' (CSSW 1977, 8).
The presence of social entrepreneurs (2c) is a further territorializing contingent factor that aligns with 'supply-side' theory in non-profit development (Salamon and Anheier 1996). It underlines the role of human agency in the territorialization process; specifically, through the presence of 'social entrepreneurs'. As Nicholls (2011, 34) explains, such individuals sometimes conform to the heroic norms and distinctive personality traits associated with conventional entrepreneurs such as risk-taking, creativity, an overly optimistic approach to analysis, and bricolage, but significantly, they are more likely to draw on the communitarian, democratic, and networkbuilding traditions that have always underpinned civil society action.
The present case study data support this and reveal, how third sector directors and chief executivesalongside regional coordinators, fieldworkers and managers of interest groupsall perform a pivotal role in (re-)organizing the sector on meso, territorial lines. For example, the outlook for the next few years will be challenging, but the sector's combination of innovation, agility, and social entrepreneurship, will bring hope to the many who depend on its services and activities to enrich their lives and those of future generations. (WCVA 2009, 2) Meso-territorialization is also contingent upon the capacity of local TSOs to engage with shifting patterns and processes of governance (2d). As the current case study attests, the pace with which this proceeds is dependent on human capital and the presence of fieldworkers, managers and coordinators and their ability to re-shape the institutional practices of sectoral administration and welfare delivery across the meso space. It also depends on the level of available financial resources. This is a constant trope in the case study discourse. For example, Staff Structure: In order to meet the objectives our staff has been considerably changed and their functions re-designated … Now for the first time the Council will have Officers who will liaise generally with the voluntary movement in both north and south Wales … already there is ample evidence that staff are getting to grips with practical problems of both national and local significance. (CSSWM 1977, 9) Anderson's work (Anderson 1991) underlines the role of institutions in the (re-)construction of collective memories. As Bello (2011, 34) notes, 'currently, at the regional level, individuals and citizens' groups have acted together in the formation of civil society, thus contributing to the construction of a regional identity through horizontal dynamics'. The present case study is no exception and underlines that the strength of national identity and civic nationalist mobilization are key contingent factors in sectoral territorialization (3a). As Itçaina's (2010) work attests, they provides a rationale for re-spatialising the structures of voluntarism in a given territory. In turn, this may be conceptualized by drawing on social movements theory (Benford and Snow 2000): specifically, the presence of cultural framings and identity politics, as well as institutional networks and mobilizing structures. For example, The overall aim of the Council is to promote, support and facilitate voluntary action and community development in Wales. In pursuit of that aim the Council will be guided by the following principles: to employ the strength of local identity as a resource by giving encouragement to local associations. (CSSW 1982, 2, see also discussion of national identity below) Meso-territorialization is also contingent upon electoral politics in union states (Chaney 2014). Regional elections determine the political opportunities for respatialising third sector welfare delivery (3b). Inter alia, they present opportunities for third sector interests to lobby and shape the political agenda based on 'sub-national' social policy needs and concerns. In our case study this has been accompanied by third sector bodies publishing their own election manifestos; further driving the development of meso territorialization. For example, 'WCVA's 2011 [National] Assembly [for Wales] election manifesto -A Third Sector Agenda sets out steps a new Welsh Government should take to grow the third sector's contribution to active participation by people in community life and action' (WCVA 2011, 20).
The extent of social inequalities at the meso level is a further contingent factor that can promote territorialization as third sector organizations mobilize at a regional level to fight discrimination and oppression (inter alia, around nationality, (minority) language rightsas well as across 'protected characteristics') (3c). Furthermore, as Esping Andersen's (1990, 31-32) work underlines, this also extends to socio-economic matters. Class relations and patterns of need shape the character of welfare regimes (see factors 1d and 1eabove)including the role of the third sector in the mixed economy of welfare. Here three factors are germane: (1) the nature and extent of working class mobilization; (2) the types of coalitions working class elements are able to secure; and (3) the disposition of new middle class elements towards welfare provision. The contingent nature of inequalities in sectoral territorialization is to the fore in the case study data. For example, in the Articles of Association of the Council its purposes include 'the relief of poverty, distress and sickness … It must be remembered these words were written before the advent of the Welfare State' (CSSW 1964, 7) and, a later annual report states our goal is to 'help build a civil society in Wales that is inclusive and promotes equality' (WCVA 2009, 1).
The territorialization of third sector systems at the meso level is also contingent upon prevailing levels of social capital (3e). The latter term refers to the 'features of social organization, such as trust, norms, and networks' underpinning voluntarism (Putnam 1993, 167). It is a resource for collective action. In this case, this is concerned with revising pre-existing third sector institutions and practices away from centralized to meso-territorial administration. The reason for this lies in the notion of 'embededdness' (Portes 1995, 13). This is the idea that social capital is 'embedded' or linked to local patterns of associative life and is 'spatially-grounded' in a given territory. Ergo, its presence is necessary as a resource for re-scaling sectoral administration along regional lines for it is intimately concerned with territorial integrity and meso-norms of voluntarism and philanthropy (3f), as well as social networks linked to identity and the notion of the (meso) nation or region. It is a trope that is evident in the present analysis. For example, reference is variously made to 'the traditional role of volunteers giving freely of their time and energy … voluntaryism (sic) … This Council stands or falls by this conviction [… and the] activities of those who genuinely desire to serve their neighbours' (CSSWM 1963, 7); and 'the role of the sector is … dependent on having a strong community and vibrant local organizationsnot just to support people, but also to enable them to feel valued and contribute to their own communities' (WCVA 1997, 4).
Economic prosperity is a further contingent factor shaping sectoral territorialization (3g). It underpins the probabilistic relationship between the amount of revenue available to government from taxation and welfare expenditure at the meso level (factors 2a, 1e). Moreover, general income levels amongst the population determine the affordability of commodified welfare and taxation paid to the state (factor 1d), as well as the degree of income inequalities and class-based mobilization (factor 3d). Furthermore, the strength of the economy influences local patterns and processes of voluntarism; not least through shifts in labour market engagement for sections of the population and the concomitant effect on time availability for voluntary action (Warburton and Crosier 2001). The contingent nature of economic prosperity is a key trope in the present case study data. For example, everyone knows that the storm-clouds over the national economy have finally burst in a downpour of inflation and soaring costs. This has been accompanied by a fall in the amount of money available from charitable institutions and the general public [… as well as] a serious drop in the grants on which we have relied in the past, for many important areas of our work. (CSSW 1975, 2) The foregoing analysis affirms the salience of the contingent factors to understanding the territorialization of third sector welfare. We now turn to the second aspect of our analytical framework, critical junctures.

Critical junctures
The following details five critical junctures identified by textual analysis of the longitudinal case study data (Table 1). Each is now considered in turn.

Strategic planning associated with the Second World War
For administrative reasons the Second World War constitutes a critical juncture in the meso-territorialization of welfare. Specifically, the demands of war expose shortcomings in central government's policy-reach. It then uses Table 1. Case study: critical Junctures and corresponding contingent factors in the meso territorialization of third sector administration and welfare delivery in Wales.

Years
Critical Juncture Prevailing Contingent factors (see Figure 1 for key to individual factors).

1939-45
Strategic planning -Second World War central government's use of Wales as a spatial frame for public administrationcrucially, involving the third sector  (Salamon and Anheier 1996, 16). Initially, SWMCSS was a local voluntary organization concerned with the provision of welfare (such as employability schemes, poverty reduction initiatives and social care). It Prevailing Contingent factors (see Figure 1 for key to individual factors).
Political: Sharp rise in social inequalities with unprecedented surge in unemployed (3c, 3d); widespread deprivation (3 g 1999 Devolution: devolved legislature and governmentstatutory Government-third sector partnership operated in industrialized areas affected by the depression and poverty. Its goals rapidly broadened. A miners' welfare committee was established in 1933. The Council's provision of a raft of training, social care, support and advice activities followed. In 1940 the Council reported that The year of the war has presented the Council with many new demands which it has tried to face, without at the same time neglecting those people for whose welfare it has had a special responsibility in the last seven years. (SWMCSS 1940, 3) At this time Wales was firmly incorporated into centralized administration of the British state (Morgan 1990). There were few explicitly Welsh government structures and voluntary sector administration operated on state-wide, British basis from London (Brenton 1985). From these local beginnings, in the parlance of historical institutionalism, the Second World War marked a 'critical juncture' in the territorialization of the sector. As the contemporary record in our dataset reveals: The unpreparedness of voluntary organisations in face of the crisis of 1938 resulted in a determined effort in succeeding months to remedy the deficiency. By the outbreak of war a Standing Conference of Welsh Voluntary Organisations had been brought into being which set itself to ensure that the maximum effort should be put forth in time of war to ensure the greatest economy of voluntary service compatible with efficiency (CSSWM 1940, 3).
Throughout the War the Council sought to operate on an all-Wales basis. This constitutes a milestone in the territorialization of the sector; one that is alluded to in the Council's 1947 report, there is a great need in Wales, both in town and country for an efficient, forward looking, co-ordinating social service agency which will cover the whole Principality [i.e. Wales]. Nowhere are there more formidable difficulties of geography, finance, and organisation to be overcome. (CSSWM 1947, 7) Thus, the all-Wales body (CSSWM) emerges from the British framework of third sector administration hitherto provided by the National Council for Social Services. 1 As noted, the motivation for its genesis stemmed from recognition that London-centred administrative structures did not reach all aspects of welfare, notably the evacuation of children from urban centres, as well as the care of refugees. These became the charge of the Council. Thus, during this formative period, the Council's role was described thus, 'to unite all the County [voluntary] Associations in our Country, so that they can discuss problems common to Wales [and] determine upon a common policy' (CSSWM 1949, 3). Underlining the notion of a critical juncture, the wartime arrangements then extended into the post-war period and became normalized: the Council was seen by Westminster as a Welsh welfare delivery vehicle for aspects of welfare. Thus, for example, in 1951 the Council notes, 'The [Whitehall] Development Commissioners recognize our Council as the liaison [body]' (CSSWM 1951, 8).
Government restructuring: a new, territorial government ministry for Wales During the 1950s and 60s a resurgent nationalist movement in Wales was pressing for home rule (Morgan 1990). In their manifesto for the 1959 general election the left-of-centre Labour Party proposed the creation of a territorial ministry of the British state to be run by a Secretary of State for Wales (Chaney 2013). This would have executive powers and determine expenditure on public services delegated from Westminster. The creation of the Welsh Office in 1964 is the second critical juncture in the territorializing of third sector administration. Its principal significance lies in the creation of the first all-Wales institution of government in modern history with which the nascent third sector body could engage. Government was keen to encourage this. The Minister of Welsh Affairs addressed the CSSWM in 1962 and affirmed that: The growth of bodies like the Council of Social Service with a broad overall view has been of great value in co-ordinating separate efforts. It has done much to bring about a necessary unity, but it is arguable that it has never yet played the full role for which it was designed. [Strikingly, he then posed the rhetorical question …] should the Council of Social Service do more to interpret Wales to herself? (CSSWM 1962, 8).
Over subsequent years repeated reference is made the case study data to the CSSWM's national coordinating role. For example: One of the lessons to be learned in our modern society is the necessity to work together for the common good [… This can only be achieved by] working in concert with Voluntary Organisations. In Wales, the task of uniting these voluntary organisations is the first responsibility of this Council. In all its Departmental activities the Council endeavours to bring those responsible for statutory services around a table with individual groups working in communities in the spheres of health, welfare and education (CSSWM 1966, 9).
In terms of membership during this period, the Council is made up of representatives of local government; representatives of affiliated external organizations (e.g. Workers' Educational Association, the Industrial Association of Wales and Monmouthshire); individual members; and networks of voluntary organizations and clubs representing specific interests (e.g. Drama Association of Wales, and Citizens Advice Bureaux Wales). The post-1964 period is characterized by growing ties between the CSSWM and the Welsh Office. By the early 1970s the CSSWM reported that it was making 'very fast progress' in establishing its national, Welsh role. Reference to its work for disabled people clearly illustrates this. For example, the contemporary record notes that this, 'could only have been undertaken by a Welsh national service bringing together disabled organisations' (CSSWM 1973, 22). Just a decade on from the creation of a territorial ministry for Wales, the CSSWM's assessment is one of 'excellent working arrangements with the Welsh Office … fulfil[ling] one of its main objectives as a consultative and co-ordinating body' (CSSW 1975, 4). Between 1964 and 1974 the case study data reveal the third sector body's transition from welfare body to an institution with growing ambition to become a 'player' in the Welsh political landscape. For example, 'it is to be hoped that in the coming year the new Council will consolidate its position as an important element in Welsh affairs' (CSSWM 1972, 6); and 'the newly reconstructed Council [now renamed the Council for Social Services Wales or CSSW] 2 is, I feel, ready to play a vital and central role throughout the Principality' (CSSW 1972, 2).

Thatcherite reforms
In 1979 the election of a right wing government headed by Margaret Thatcher led to a number of reforms that constitute the next critical juncture. The government's monetarist policies saw a major rise in unemployment accompanied by deep cuts in public expenditure. The following extract from the Council's records illustrates the impact of the reforms: During the last eighteen months, the country has faced a transformation as a result of Government fiscal and economic policies. Latest reports show that one in every eight males, and one in every twelve females, of the Welsh workforce is unemployed. The social implications of these policies, and the serious consequences that face individuals, families and Communities are universally recognised. Naturally, social and Community provisions are affected. Central and Local Government Grants have been reduced … there are signs that most of these Organisations have to curtail spending and programmes' (CSSW 1980, 6).
The Conservative reforms also prompted strong resistance from the Council. For example: we have been taking a firm line on the question of the use of Volunteers in Public Services, especially when those previously working in this sector have been made redundant … Volunteers we recruit must not fill the places of those who have lost their jobs. (CSSW 1981, 4) The financial crises also requires the Council to secure major grant funding from government. However, the resulting contracts to deliver training to the unemployed (which involved a much needed income to the Council in excess of £1M and the employment of 100 new staff) also saw a shift in power relations with government. They now became top-down and transactional. Thus, the Council records note that 1992 was 'the first year that WCVA has ever received a monitoring report from the Welsh Office'. Underlining government's dominant position, it referred to TSOs as 'voluntary sector customers' (WCVA 1992, 4, emphasis added). This subordination is further underlined by the appointment of five 'Assessors to [the Council's] Executive': four were Welsh Office officials.
In 1983 the Council again changed its name from CSSW to Wales Council for Voluntary Action (WCVA). More than symbolic, the change reflected the reciprocal way in which governance shifts (principally, Thatcherite reforms and the strengthening role and policy-reach of meso government) promoted functional change in the sector. In short, a process of politicization was at work as the Council extended its remit, and placed greater emphasis on representation and policy work, whilst continuing with its social service provision. As the following extract illustrates, deteriorating relations with government prompted the Council to assert its independence: We are an independent institution, whose members are voluntary organisations in Wales, and its function is to promote and support voluntary action through development work, through the giving of services [… it is] a national institution which asserts the importance of voluntary action for the health of society: which advocates a consideration in policies and programmes for a proper recognition of voluntary initiative (WCVA 1984, 2).

Third sector compacts
The next critical junctures follow in swift succession. The fourth came in 1997 with the advent of voluntary sector 'compacts'. These were formalized contracts setting out the parameters and expectations of state-third sector relations. They reflected the change of government at Westminster and the New Labour administration's emphasis on partnership working. Their effect was to formalize the territorialization process for the UK government required bespoke compacts from the territorial ministries for Wales, Scotland and Northern Ireland (and Westminster in the case of England). This shift is described in the Council's report of the time as 'a major step forward', noting that the 'government is determined to forge the strongest possible links with the voluntary sector and to support it in every way it can' (WCVA 1998, 3). It went on to note that the new relationship, will be built on integrity, trust and mutual respect. It will commit the Welsh Office to: Clear policies on volunteering, community development and voluntary organisations, and measures to support them; consultation with the sector on policy changes; [and] a range of measures to improve funding mechanisms. (WCVA 1998, 3) Devolution The fifth critical juncture followed shortly afterwards when in 1999 the Welsh Office was replaced by a new elected legislature, the National Assembly for Wales (latterly, the Welsh Parliament or Senedd Cymru). This again illustrates the iterative, reciprocal state-third sector relationship at work in the process of territorialization. Notably, pro-devolution activists in the political elite were keen to encourage the involvement of the sector in the design of the new elected government body for Wales. Thus, at the time the National Assembly's First Minister asserted that: the new legislature 'must recognise the voluntary sector as a key player and partner … at the heart of our work to build a better Wales' (cited in Dicks, Hall, and Pithouse 2001, 148).
The cornerstone of the new arrangements was a statutory partnership between the National Assembly for Wales and the voluntary sector (Government of Wales Act 2006, s.74). 3 The statutory provisions also stated that the key nexus with government was to be the sector's representative body, the WCVA. The significance of these developments was not lost on those writing at the time: there can be no doubt that as the voluntary sector has moved centre stage in national politics in Wales, so has the WCVA. It has been centrally involved in the process of drawing up devolution legislation and preparing [National] Assembly procedures. As a result, the WCVA has entered the era of devolution with expectations of being a significant and key player in the Assembly and its development. (Dicks, Hall, and Pithouse 2001, 156)

Discussion
The wider significance of this study is both theoretical and empirical. In the former regard, it makes an original contribution by synthesizing neo-institutionalism and critical realism with Salamon and Anheier's (1996) classic framework on civic infrastructure development to offer an inductive, systemic analytical model of contingent factors shaping the meso-territorialization of third sector welfare in federal and union states. Moreover, we go beyond earlier discrete approaches to exploring territorialization that centre on single factors such as neo-liberalism and marketization. Instead, we conceptualize decentralization in terms of a raft of temporal and spatial variables. Our core finding is that welfare decentralization is driven by discontinuity at critical junctures related to factors such as governance transitions, national crises and political shifts.
The application of the model to the case study data allows us to reflect upon the evolution of sub-national welfare and the significance of UK devolution to state-third sector relations. The longitudinal analysis reveals how the governance transitions associated with devolution consolidated an earlier process of territorialization when, during the crisis of the Second World War, the Westminster government turned to SWMCSS for welfare coordination and delivery on an all-Wales basis; breaking with the previous British structures of welfare and voluntary sector administration. Indeed, in this regard the territorialization of the third sector in Wales in the 1940s and 1950s can be seen as a precursor to political devolution in the 1990s. Devolution then accelerated the Council's shift from a sole concern with welfare delivery to a more politicized way of working. Notably, the creation of the Welsh Office in the 1960s drew the Council into lobbying government on behalf of the sector in order to shape public policy and services. Thatcherite reforms in the 1980s deepened this process of politicization. Whilst the (re-)creation of a national Welsh legislature in 1999 further transformed matters by institutionalizingand giving legal effect to, the prevailing political discourse around partnership working with government.
At this point it is useful to reflect on how the Welsh case fits with extant work and wider trends. First, it underlines the iterative, reciprocal relationship between governance reforms and third sector territorialization. The emergence of the Welsh Council in the 1930s also illustrates how meso-territorialization can originate as a response to welfare demand in the face of economic depression and pronounced inequalities compounded by state and market failure in the delivery of key public goods. Subsequent development of the sector also underlines the influence of socio-economic factors including national identity, language and culture. This is consonant with the existing body of work on the role of civic nationalism in driving welfare decentralization. The extended time taken to fully establish the Council on an all-Wales basis (completing an organizational infrastructure for the Council in the north of Wales lagged behind institutional development in the south) also has wider significance. It supports the social origins thesis (Salamon 1970) because it reveals how meso-territorialization depends on prevailing levels of social capital, the presence of social entrepreneurs, the strength of third sector networks and organizational density. The developing relationship with the Welsh Office in the 1970s sees the Council drawn into fulfilling government contracts on skills and employability training. This led to the Welsh Office's increasing oversight of the Councila trend that resonates with earlier work on welfare decentralization and the quest for new forms of accountability and performance monitoring (Seabright 1996). Moreover, the Council's dogged opposition to Thatcherite reforms supports earlier work on resistance to central and federal policies as a further driver of territorialization (Mooney and Scott 2012).
The wider contribution of this study to social research lies in highlighting the temporal and spatial factors at work in territorializing welfare in a subnational European polity, thereby extending understanding of how, 'voluntary and community sector groups are agents of a complex process of "scalar manoeuvring" whereby … governance is produced and contested across a range of sites both within and across spatial scales' (Kythreotis and Jonas 2012, 382). This study's analytical framework is designed to inform future empirical investigation and be applied in other contexts. The global trend of state restructuring and international prevalence of federal and union states gives this broad salience and paves the way for further holistic studies of third sector welfare decentralization.