New Public Management Reform ideas and the remaking of the Italian and Danish health systems

ABSTRACT This article explores the role of New Public Management (NPM) ideas in the reform of the Italian and Danish health systems. In particular, it investigates the arguments that have supported changes in their institutional and territorial dimensions. Whereas the institutional dimension is concerned with the allocation of authority between layers of government, the territorial dimension is related to the definition of the size of subnational governments. By adopting an argumentative approach to policy analysis and focusing on the role of ideas in policy formulation, the article shows how Italian and Danish policymakers have shaped the institutional and territorial changes that have taken place in their national health systems. Through document analysis and in-depth interviews with experts and decision-makers, we show that the relationship between changes in size and authority in relation to NPM ideas is not clear-cut. Whereas in Italy arguments about the need to change the allocation of authority were not tied to a discourse about the size of subnational governments and health agencies, in Denmark discourses about size and authority have run in parallel. The present study confirms previous findings and provides additional evidence suggesting that NPM ideas have spread widely through governments following a path of implementation that varies from one national context to another.


INTRODUCTION
From a multilevel governance perspective, public sector reforms consist of two different, albeit intertwined dimensions: (1) an institutional dimension, which is concerned with the reallocation of authority across layers of government; and (2) a territorial dimension, which is related to reshaping the territorial administrative boundaries of subnational governments (Organisation for Economic Co-operation and Development (OECD), 2017).Whereas in some countries reforms have mainly emphasized the need to change the territorial boundaries of subnational governments, in others more emphasis has been put on the redistribution of power and responsibilities between different levels of government, with or without reshaping territorial administrative boundaries (Kjellberg, 1985).
an increasingly decisive role in the functioning of the welfare state since the end of the Second World War.This article focuses on health systems, defined as a set of institutions concerned with the finance, provision and regulation of healthcare (Freeman & Frisina, 2010).In this section we provide a conceptualization of the institutional and territorial dimensions characterizing the overall territorial architecture of health systems.Whereas the former is concerned with the allocation of authority within the multilevel structure of government, the latter is related to the definition of the size of subnational governments and, therefore, to the administrative boundaries of territorial units stricto sensu (OECD, 2017).This distinction between institutional and territorial dimensions has an important analytical rationale in that it allows size and authority to be analysed separately as well as together.
A crucial concept for the analysis of the allocation of authority is 'decentralization', which refers to the transfer of authority in relation to public policies from the central level to subnational levels of government (De Vries, 2000).Decentralization may include policy-planning, prioritysetting, service provision, resource allocation, spending functions and revenue-raising.Authority with regard to these functions can be transferred to either regional or municipal governments, and change within a health system's institutional dimension can lead toward a more or less (de)centralized health system.In the literature, the strengthening of central government control or the transfer of authority in relation to certain functions from subnational governments back to the national levelis referred to as recentralization (Gershberg, 1998).
Decentralization plays a key role in NPM discourse (Pollitt, 2005).Consistent with NPM principles, the crucial arguments supporting decentralization rest primarily on the notion of efficiency, and they are mainly associated with the economic literature on fiscal federalism (Oates, 1999;Weingast, 2009).In essence, because decentralization brings the government closer to citizens, health services can be delivered at a lower cost (technical efficiency) and meet heterogeneous needs and preferences (allocative efficiency) (Hooghe & Marks, 2009).Nevertheless, there are also potential risks of decentralization.Because of differences in the political, administrative and fiscal capacity of subnational decision-making units, decentralization can widen geographical disparities in service provision (in terms of access and quality) throughout the national territory.However, the empirical evidence on the impact of decentralization is ambiguous.(In)equity in health and healthcare is related to a number of contextual factors, such as the presence or absence of adequate equalizing and solidarity mechanisms between subnational governments (Alves et al., 2013).Moreover, decentralization can fuel intergovernmental coordination problems, especially where public finances are concerned.These disadvantages might lead the central government to recentralize political, administrative, and fiscal powers and responsibilities (for a review of the advantages and disadvantages of decentralization, see, e.g., De Vries, 2000).Recentralization is a characterizing feature of what has been labelled 'post-NPM'.Since the end of the 1990s and the beginning of the 2000s, post-NPM has referred to all those reforms that have been aimed at attenuating and counteracting the negative consequences of NPM, ultimately improving coordination and helping the central government to steer the decentralized architecture of the public sector (Reiter & Klenk, 2018;Trein & Maggetti, 2019).
The territorial dimension of multilevel governance reforms is concerned with the redefinition of the geographical boundaries of regional and municipal governments.Over the past few decades, drawing from NPM ideas, reformers have been calling for a restructuring of the public sector in the light of issues of economies of scale (Dijk et al., 2015).This restructuring includes the enlargement of areas through mergers and amalgamations (upscaling).As with decentralization, the rationale behind mergers is that of efficiency.In fact, mergers may make it possible to exploit economies of scale, leading to cost savings (technical efficiency).Moreover, administrative overheads and the risk of multiplication of tasks can be reduced, increasing administrative capacity.However, there are also arguments against mergers, which are in favour of reducing the size of areas through division (downscaling).Such arguments are mainly based on upholding democracy: in small units of government, local decision-makers would be more accountable and responsive to citizens' needs (allocative efficiency) (Blom-Hansen et al., 2014).As well as considering the geographical boundaries of subnational governments, in the health policy field, many studies on scale efficiency have focused on the optimal size of healthcare organizations such as hospitals.A systematic review by Giancotti et al. (2017) shows there is consistent evidence that economies of scale favour the merging of hospitals; thus, policies aimed at establishing large hospitals because they are more efficient than small ones meet with approval.In short, according to arguments supporting the advantages of economies of scale, increasing the size of decision-making units allows for efficiency gains in that large operational units may take advantage 'of economies of scale by producing maximum output per unit of input and reducing the average unit costs of production' (Giancotti et al., 2017, p. 2).
In investigating institutional and territorial reforms, this article points to the role of ideas in policy formulation.As part of the process of designing public policies, policy formulation is 'that stage of policy-making where a range of available options is considered and then reduced to some set that relevant policy actors, especially in government, can agree may be usefully employed to address a policy issue' (Howlett & Mukherjee, 2017, p. 6).It is at this stage of the policymaking process that knowledge and expertise come to the fore.During policy formulation, experts and decision-makers debate ways of addressing policy problems and the feasibility of the proposed solutions.As Béland (2016, p. 739) puts it, policy formulation is 'the stage the most traditionally associated with the role of ideas'.In fact, ideas held by actors participating in the process are a key factor in constructing adequate policy solutions.However, ideas are not only concerned with how to effect a solution with regard to a certain issue, but also its definition.Therefore, we argue that issue interpretation and the elaboration of solutions should not be analysed separately.The way an issue is understood and the articulation of arguments as to how to effect an appropriate solution are both part of the process of policy formulation.Therefore, this article conceptualizes ideas as causal beliefs that specify cause-and-effect relationships, elucidate the issues at stake, and prescribe what to do to effect a solution (Campbell, 2004).As such, ideas provide guidelines for action and serve to justify and 'legitimize the policies in a program through arguments based on their appropriateness' (Schmidt, 2017, p. 172; see also Majone, 1989).Figure 1 provides a graphical representation of our theoretical framework.
In the light of the arguments supporting decentralization/recentralization (institutional dimension) and upscaling/downscaling (territorial dimension) we presented above, we expect reformers to favour decentralization so they may seize the opportunities in terms of efficiency gains.However, if the issue of technical efficiency is understood as a matter of failure to exploit economies of scale, arguments supporting the benefits of subnational amalgamations are expected to feature prominently.
Although this article draws analytical attention mostly to the role of ideas in policy formulation, it does not neglect the influence of contextual (institutional) factors that might constrain or facilitate change.In particular, the article takes into consideration the constitutional design and national legislative framework of the two countries, as well as features of the system of intergovernmental relations.Moreover, because changes in the institutional and territorial dimensions of health systems might be part of a larger reform package involving other policy areas, we also take into account the political-economic environment to obtain a comprehensive picture of the overall reform context.

RESEARCH DESIGN
Health systems can generally be split into two types, national health service (NHS) and social health insurance (SHI), which brings us to the distinction between the Beveridge and Bismarck models (Blank & Burau, 2014).Whereas SHI systems are predominantly financed out of social contributions and coverage is linked to labour market participation, recognizing differences between occupational categories, NHS systems are mainly financed out of general taxation and are committed to ensuring universal, equal and almost free access to health services for all citizens.
This article adopts a small-N research design.The universe of cases we refer to is all those European countries that have adopted an NHS-type system: Denmark, Finland, Ireland, Italy, Norway, Sweden and the UK.The actual selection of cases has been restricted to those national health services presenting a similar territorial architecture organized at three levels: national, regional and municipal (Table 1).Therefore, first, we have excluded Finland and Ireland, the health services of both of which have a two-tier architecture, 1 as well as the UK.The latter, in fact, presents a very peculiar organization comprising four constituent countries each with its own NHS (England, Scotland, Wales and Northern Ireland).This structure renders the comparison with other countries problematic.Second, out of the national health services of Norway, Sweden, Italy and Denmark, we have chosen the latter two for in-depth analysis.In particular, among the Scandinavian countries, we have selected Denmark because it represents an interesting case in which substantial reforms in both the institutional and territorial dimensions took place.Moreover, the type of system and the three-tier architecture are factors that have been constant over time in both cases.In fact, since the two countries adopted a national health service in the 1970s, the type of system has remained the same and the territorial organization is still based on the national, regional and municipal levels of government.As far as the subnational territorial structure is concerned, Italy has 20 regions and 8047 municipalities, whereas Denmark has five regions and 98 municipalities.It is important to highlight that the average municipal size (number of inhabitants) is higher in Denmark (58,155) than in Italy (7545), and the proportion of very small municipalities with fewer than 2000 inhabitants is higher in Italy (44%) than Denmark (1%) (OECD, 2017; OECD/UCLG, 2016).Moreover, since the 1970s, whereas Denmark has witnessed a significant decrease in the number of municipalities and, consequently, the establishment of bigger units through upscaling, in Italy there has been no significant variation (Ebinger et al., 2019).
The sources from which empirical evidence is drawn are as follows: national legislation; government documents; parliamentary debates; the reports of parliamentary commissions and committees; national and international reports; and 19 in-depth interviews with key Italian and Danish experts and decision-makers (see Appendix Table A1 in the supplemental data online for the list of interviews and documents).The latter are mainly representatives from either central government or subnational governments/organizations with relevant roles in national policymaking.Interviewees were selected through snowball sampling and interview questions were open-ended, adjusted to the interviewees' profiles, and covered aspects related to the arguments supporting institutional and/or territorial reforms at the national level.As far as the methods are concerned, we have carried out a content analysis of textual data, systematically interpreting and describing the material in relation to the policy issues at stake and the ideas embraced to effect a solution.

HEALTH SYSTEM REFORMS: INSTITUTIONAL AND TERRITORIAL CHANGE IN ITALY AND DENMARK
Reform content and context Italy In 1978, Italy abandoned a Bismarck SHI system and adopted a Beveridge NHS.The new Italian health system was based on a decentralized organizational structure that consisted of three tiers: central government, the regions and the local health units (LHUs), which operated at the municipal level.Subnational governments enjoyed some political and administrative powers.
Although the regions were responsible for regional planning according to the objectives specified by the central government, LHUs were responsible for providing services and were run by managing committees elected by assemblies of representatives from the municipalities.The central government was responsible for overall planning and it was the only tier with financing responsibilities.
During the early 1990s, decentralization of administrative and fiscal authority was further promoted, and whereas the role of municipalities was significantly reduced, the regions were provided with more autonomy.Moreover, LHUs and major hospitals were transformed into enterprises.These changes of the early 1990s occurred in a particular conjunctural moment and need to be contextualized.In 1990, the overall arrangement of subnational governments was redefined.Functions concerning the establishment of general socioeconomic and territorial objectives were assigned to the regions, and municipal governments were allocated a more consultative and executive role.A few years later, Italy was hit by a profound political and economic earthquake.Between 1992 and 1994, the core parties of the so-called 'First Republic' collapsed due to corruption scandals.In addition to the political crisis, the high level of public debt forced Italy to leave the European Monetary System.The overall economic policy of the governments in charge during those years 2 was aimed at reducing public debt and the budget deficit, with the overarching goal of meeting the Maastricht criteria for joining the European Economic and Monetary Union (EMU).These events opened a window of opportunity for launching major reforms, including measures concerning the territorial architecture of the health system.
During the late 1990s, decentralization was strengthened by several legislative measures approved by centre-left governments during the period 1997-2000 and by a constitutional reform in 2001, all of which consolidated regional legislative powers in relation to many policy areas, including healthcare.These reforms were carried out during a period when other important changes in the system of intergovernmental relations took place.Several bills concerned with the reorganization of public administration were adopted, and in 1999 a constitutional reform introduced the direct election of regional governors, empowering them in their relationship with the central government.Moreover, through the institutionalization of the State-Regions Conference as an instrument for addressing intergovernmental conflict and for reaching joint agreements, a system of 'cooperative federalism' was put in place (Fargion, 2005).
During the 2000s, a new season of intergovernmental coordination with regard to healthcare matters started.This took the form of intergovernmental agreements, and between 2000 and 2014 six agreements 3 were signed between the central government and the regions.These agreements have resulted in a partial recentralization of the health system.In particular, sanctions for regions unable to contain health spending were established.Moreover, regions that did overspend were obliged to provide budgetary recovery plansto be agreed with the central governmentto eliminate (and prevent future) deficits.The introduction of these plans, which have proved to be effective in erasing regional deficits, represented a form of administrative subordination through which the central government gave help to regions experiencing financial difficulties (Bordignon et al., 2020).These changes have to be read in the light of the constraints imposed on public finances by the Stability and Growth Pact enacted in 1997, which forced the EMU countries to control their budget balances and to set rules for financial coordination known as Domestic Stability Pacts (DSPs).In Italy, since 1999, the DSP has set fiscal constraints on subnational governments to ensure fiscal discipline.Therefore, in the health sector, the focus has been placed on expenditure control to monitor subnational deficits.

Denmark
The Danish health system, as well as the welfare state in general, has a strong and long-standing tradition of decentralization.From the beginning of the 19th century, the Danish system of subnational government consisted of 24 counties and more than 1300 municipalities.Counties and municipalities mainly enjoyed powers related to the ownership and administration of hospitals (Pedersen et al., 2005).
During the 1970s, a series of reforms, which included the whole public sector, were implemented and reduced the number of counties to 14 and the number of municipalities to 275.The increase in size and reduction in number of subnational governments were accompanied by a transfer of authority toward the latter, in particular with regard to welfare functions, including healthcare.As far as intergovernmental financial relations are concerned, subnational governments have always had the right of taxation, which was supplemented with central government conditional grants, namely, grants that were allocated as percentages of subnational expenditure for specific purposes.In the early 1970s, conditional grants were replaced by general grants, which could be used in any policy sector.With the shift to general grants, subnational governments enjoyed more autonomy in relation to resource allocation and priority-setting (OECD, 1997).
In 1973, in the context of these reforms concerning both territorial boundaries and the allocation of authority, a tax-funded Beveridge NHS was established, with county and municipal taxes being the most important components of total health funding.In addition to these, the central government provided subnational governments with general grants.Although county councils were responsible for a large part of healthcare delivery (Ministry of the Interior and Health, 2005), they operated within a central government regulatory framework.Moreover, agreements between the national and subnational governments shaped decisions about resource allocation through a system of 'budgetary cooperation'.This system was introduced in 1979 and gradually took shape during the 1980s, becoming a distinctive trait of Danish cooperative intergovernmental relations (Blom-Hansen, 2012).During negotiations, which were mainly concerned with tax rates and expenditure ceilings, subnational governments have always acted as a collective body through the Association of County Councils (representing the counties) and Local Government Denmark (representing the municipalities).This means that the agreements were binding for all counties and municipalities as whole on uniform terms.The commitments were made credible by the introduction of financial sanctions for those subnational governments that did not comply with the agreements (Grønvald & Alban, 1995).
Since the beginning of the 1990s, annual intergovernmental agreements in relation to budgetary allocation have become increasingly detailed, 4 de facto limiting subnational autonomy with regard to healthcare matters.In the light of the EU budget deficit and debt norms and the requirement to provide a convergence programme, the central government gradually strengthened its fiscal policy framework.The centre-left coalition that took office in 1993 focused its efforts on the reduction of public debt with the aim of ensuring fiscal sustainability over the longer term (OECD, 2012).In 2007, a major structural reform abolished the 14 counties, which were replaced by five regions, and the number of municipalities was reduced from 275 to 98.The authority of the central government also increased.In particular, the National Board of Health took over the role of overall planning and coordination for the health sector (Olejaz et al., 2012).Moreover, there were important and significant changes in relation to the financing of the health system.In fact, the power of the regions to levy taxes was abolished.
The role of ideas in policy formulation Italy In Italy, during the 1970s, the need to transfer powers with regard to healthcare matters to the newly established regionsas prescribed by the Constitutiondominated parliamentary debates and there was almost unanimous agreement that this should be done.The reform process was shaped by the idea that Ferrera (1995, p. 279) has called 'full democratic universalism', according to which the organization of the health system should be subject to popular democratic control at subnational level.In line with this notion, subnational governments came to enjoy a certain amount of political and administrative power.
After the 1978 reform, the dividing line between central, regional and municipal authorities remained blurred, giving rise to policy incoherence at both national and subnational levels.This new institutional setting revealed two crucial problems: excessive politicization and the presence of vertical fiscal imbalances (Freddi, 1984).With regard to the first problem, because LHUs were managed by committees formed on the basis of party considerations, they became arenas for political exchange and often venues for corruption and clientelism (Ferrera, 1995).The second problem stemmed from the separation between central financing responsibilities and subnational spending authority (Buglione & France, 1983).In other words, subnational government spending was not financed by regional and municipal revenues.In a context of deep economic crisis, the central government tried to keep health expenditure under control by deliberately underfinancing healthcare and by setting tight budgetary ceilings (France, 2007).However, the intergovernmental financial arrangement created the perception of soft budget constraints: in that subnational governments did not have any fiscal responsibility, they were not encouraged to respect the budgetary ceilings because they knew that the central government would subsidize excessive subnational expenditure (Bordignon & Turati, 2009).Therefore, since its inception and throughout the 1980s, the Italian health system suffered from an inefficient use of resources.The central government had limited power in controlling how funds were spent by the LHUs, and the asymmetry in funding and spending responsibilities triggered subnational deficits, creating a situation of permanent financial crisis (OECD, 1992).
During the 1980s, the debate about how to make the health system more efficient was lively.In particular, experts and decision-makers in central government pointed to the need to improve technical and allocative efficiency.As one interviewee has commented, 'there was the need to empower the [subnational] level [and] create a system of accountability'. 5As also highlighted in an evaluation report, '[t]he goal of building up the premises for having efficient local decision-making is, at the moment, widely discouraged.… The health system has not been able to equip itself with an acceptable administrative and managerial structure' (Perkoff, 1984, pp. 31-33).The Center for Research on Health and Social Care Management (CERGAS) at Bocconi University, Milan, which was established in 1978, was particularly influential in fuelling the debate and in promoting the advantages of fiscal federalism (Terlizzi, 2019b).The director of CERGAS advocated for strengthening subnational administrative and fiscal autonomy.These ideas soon became well established in central government circles.Documents and interview data confirm that the dominant discourse revolved around the idea that decentralization was the right means by which to improve cost-effectiveness (technical efficiency), and responsiveness and accountability (allocative efficiency). 6As a result, decentralization measures were gradually implemented during the early 1990s.The overall aim was to ensure a 'more efficient management of the health system [through the direct involvement and empowerment of the regions]'. 7 Changes in the allocation of authority were accompanied by changes in the management of subnational government bodies, in an attempt to make their internal organizational structure more efficient. 8In line with NPM ideas, LHUs and major hospitals were transformed into local health enterprises (LHEs) and hospital enterprises (HEs), respectively.Therefore, the 659 LHUs were merged into 228 LHEs, and 81 HEs were established.Again, the role of CER-GAS in spreading these ideas was crucial.The managerial approach recognized the separation between politics and management and, therefore, reduced the influence of the former over the latter.In fact, the new autonomous bodies were governed by general managers who were independent of political influence.These arguments were also supported by the central government.
As stated in the 1994-1996 National Health Plan promoted by the then Minister of Health Maria Pia Garavaglia, managerialism was the 'fundamental instrument for managing health care facilities [efficiently] at the territorial level' (Ministero della Sanità, 1994, p. 4).
The number of LHEs and HEs has been progressively reduced during the 1990s and 2000s. 9According to some observers, the rationale behind this is a bid to strike a balance between the need to provide health services according to local needs and an attempt to achieve efficiency gains through economies of scale (Anessi-Pessina & Cantù, 2006;Fattore & Morando, 2016).However, evidence suggests that, particularly during the 1990s, arguments in favour of the benefits of economies of scale were not supported by decision-makers in central government.For example, although in 1992 Legislative Decree n. 502 stated that regions had to reduce the number of LHUs and merge them into larger LHEs, no reference to the advantages of economies of scale has been found in parliamentary debates and committee reports.Moreover, in discussing the crucial issues at stake during the reform processes at national level, interviewees did not mention the relevance of arguments about economies of scale.
In Italy, therefore, the changes in size did not relate to the redefinition of the geographical boundaries of subnational governments, as they did in Denmark, discussed below.Although there have been mergers of territorial agencies for service delivery, this has been a process of change determined by choices made at the regional level.There has also been an increase in the size of healthcare districts, which has been achieved through municipal cooperation (Bolgherini et al., 2019), not municipal amalgamation.In this regard, interview data 10 show that the rationaledeveloped during the 1960s and 1970sbehind the establishment of 659 LHUs in 1978 was to reduce the number of municipalities: the hypothesis was that the LHUs would constitute the optimal dimensions around which to aggregate the municipalities.However, this idea was never implemented and it lost ground during the 1980s.For the purposes of this article, it is important to highlight that, unlike Denmark, the discourse about the (LHEs and HEs) mergers did not run in parallel with the discourse about the redistribution of competences between layers of government.In fact, mergers of LHEs and HEs mostly occurred from the early 2000s, with the pace of formation accelerating in 2013.In this period, no formal change in the allocation of authority between national and subnational governments was implemented, although, as we shall see, intergovernmental agreements were functioning as a tool by which the central government could steer the decentralized arrangement of the NHS.
In the late 1990s, regional autonomy with regard to revenue raising was reinforced in an attempt to reduce vertical fiscal imbalances between expenditure and revenue and solve the soft budget constraint problem.In fact, during the 1990s, there were several instances of bailouts.As Bordignon and Turati (2009, p. 308) point out, 'bailing out of past regional health deficits was endemic to the Italian NHS funding system'.The idea of increasing subnational fiscal power as well as legislative and administrative powers was predominant in the Ministry of Health, Ministry of Finance, Ministry of Treasury, Ministry of Budget and Economic Planning, and Ministry of Public Administration from the early 1990s.For example, as early as 1993, the Committee for the Reform of Regional Finance recommended that the regions' power to raise their own revenue should be increased.In 1994, the Minister of Finance Giulio Tremonti issued a White Paper that pointed to the need to promote fiscal federalism.As stated in the document: the waste of health care resources cannot be effectively stopped by the [central government]; it can instead be effectively stopped by regions, because they are directly involved in the management of health expenditure.… The change needed … is just one: fiscal federalism.(Ministero delle Finanze, 1994, pp. 17, 20, 38) In 1996, the Technical Committee for Public Expenditure issued a statement along the same lines, calling for the establishment of regional taxes to fund healthcare services.Indeed, in 1998, two regional taxes earmarked for the health system were introduced.
However, although the autonomy of the regions was strengthened, concerns about the potential disadvantages of decentralization started to arise, particularly in the Ministry of Health.The need to reinforce the coordinating role of the central government was emphasized, the overall aim being to strike a balance between efficiency and geographical equity in service provision and financing.As stated by the then Minister of Health Rosy Bindi during a parliamentary debate: it is there for all to see that at this moment the main problem of health care in Italy is the gap that exists between different parts of the territory.… I do believe that through coordination … regions can better exercise their powers. 11  During the 2000s, the emphasis on geographical equity and the need for financial discipline gave rise to intergovernmental coordination that was aimed at the reinforcement of central government monitoring, the strengthening of budget constraints, and the prevention of deficits. 12

Denmark
In Denmark, from the late 1950s, concerns about the size of subnational governments, as well as about the actual allocation of authority across the public sector, started to arise.A commission 13  to discuss possible solutions was set up by the central government in 1958 and recommended that territorial boundaries should be redefined by creating larger subnational governments and decentralizing authority.As documented by the Danish Institute: [i]t was generally agreed that the existing local government structure no longer met the needs of the times.This applied not only to the geographical structure but also to the … division of tasks among the State, the county authorities and the local authorities.(Harder, 1973, p. 123) Recommendations by another commission 14 established in 1964 pointed in the same direction, advocating for 'a wide-ranging decentralization.… Proximity … had to be strengthened in public administration and policy implementation had to be flexible and adapted to the locally varied and changing needs' (Hansen, 1997, pp. 50, 57).Accordingly, the number of subnational governments was reduced.
Amalgamations were driven by technical efficiency considerations.The main idea was that larger subnational governments had to be created to save costs through economies of scale. 15 In other words, it was thought that to be provided economically, specialized services in certain policy areasespecially healthcarerequired larger populations.Moreover, as we have seen, upscaling was accompanied by decentralization.In fact, the shared interpretation among national decision-makers was that to facilitate decentralization, larger local governments had to be established. 16The increase in size was, therefore, seen as a necessary condition for decentralization.
During the 1960s, a general agreement developed within the Ministry of Finance: 'the conditional grants had to be repealed or reduced, at least in areas where the local discretion was large [such as health care]' (Lotz, 1998, p. 21).Accordingly, in line with the principle that financial responsibility must follow authority, conditional grants were replaced by general grants.The idea was that general grants would reinforce and empower subnational governments and make them more economically responsible (Ministry of the Interior, 1999).
During the 1980s, in a context of economic recession, the health system proved to be effective in ensuring control of expenditure, which was a priority on the agenda of the centre-right governments that were in power from 1982 to 1993 (Andersen, 2008).In fact, the institutionalized system of intergovernmental cooperation with regard to budget setting ensured tight budget constraints, and (public health) expenditure remained stable during the decade (Christiansen et al., 1999).Moreover, compared with Italy, where large vertical fiscal imbalances were present, given that financing and spending responsibilities were aligned at the level of subnational governments, the latter were aware of the relationship between expenditure and available financial resources.As far as public management at the subnational level was concerned, the principles of professionalism and managerialism were key features during the 1980s.These ideas were introduced and spread as part of the wider public sector modernization programme during the decade in which NPM discourse played an important role.The programme was aimed at boosting 'the strong role of local governments in delivery of welfare services as a way of improving government effectiveness and responsiveness' (OECD, 2000, p. 14).
From the beginning of the 1990s, some of the arguments supporting the advantages of decentralization began to be questioned.Geographical differences in service provision between regions and municipalities, which had not been seen as an issue during the 1980s, became less and less accepted.Moreover, intergovernmental tensions between the national level and the counties started to rise.On the one hand, the central government complained about the fact that it was accountable for performance in the health system without being able to control the counties effectively, and on the other, the counties were frustrated by central government interference in health policy decisions (Vrangbaek & Christiansen, 2005).All of this fuelled a new debate about the need, yet again, to reform the entire structure of the Danish public sector, which in regional and municipal government circles was seen as simply not strong or robust enough. 17 In 1996, a commission 18 was appointed at the national level to analyse the allocation of authority between the different layers of government.It only had the mandate to show what could be gained by redistributing responsibilities within the existing territorial structure and its work did not lead to a reform (Blöchliger & Vammalle, 2012).As explained by the chairman of the commission, Johannes Due, the members were told that they could not propose any change to territorial boundaries, namely, any change in the size of subnational governments: 'if you cannot do bigger sizes you cannot move tasks … and therefore we could not recommend any change because we could not change the structure'. 19Again, the idea was that change in size and authority had to go hand in hand.
The centre-right coalition that came into office in 2001 again presented itself as the 'modernizer' of the welfare state, emphasizing the need to render the public sector more efficient and effective for delivering high-quality services at a reduced cost.In 2002, the government established another commission 20 to assess the advantages and disadvantages of alternative models for the organization of the public sector in general.Within the same commission, an advisory committee specifically concerned with the healthcare sector was set up.In 2004, when issuing its recommendations, the commission highlighted several weaknesses related to the size of counties and municipalities and to the distribution of responsibilities between the different levels of government.It recognized that in comparison with other countries, 'the public sector in Denmark is characterized by a very high degree of political and financial decentralization [and that] decentralization … provides a good basis for democratic control, simplicity for the users, management, coordination [and] efficiency' (Ministry of the Interior and Health, 2004, p. 14).Howeveras with the reforms of the 1970sit was also recognized that subnational governments were not large enough to carry out their tasks.Document analysis and interview data confirm that the predominant arguments were those supporting the benefits of economies of scale for improving technical efficiency. 21In essence, the commission's recommendations emphasized the need for a comprehensive reform of the public sector, involving boundary changes and a reallocation of authority between the national and subnational levels of government.The main message was that the changes should result in fewer and larger counties and municipalities to gain economies of scale.Indeed, 'fragmentation was regarded as a major obstacle to efficient and sustainable management of resources' (Busse et al., 2005, p. 79).
Six months after the commission's recommendations were issued, agreement on a reform was reached.Ideas supporting decentralization were not rejected.Before the reform came into effect, in a statement to Parliament the Minister of the Interior and Health affirmed that the basic goal behind the reform was 'to consolidate and strengthen the decentralized Danish model' (cited in Local Government Denmark, Danish Regions, and Ministry of the Interior and Social Affairs, 2009, p. 8).However, in effect, the reform contained elements of recentralization, because some counties' tasks were transferred to the central government.Again, NPM ideas constituted an important driving force within the commission: with my background as an economist at the Department of Finance I have been a keen supporter of NPM.There has been a big debate in Denmark about NPM.… I think that, for most of the members of the Commission, NPM was the driving force. 22  Contrary to the reforms of the 1970s, in the health sector, the increase in size and reduction in number of subnational governments was not accompanied by authority being transferred to them.Instead, the need to increase the authority of the central government vis-à-vis subnational governments was underlined.
The reform, implemented in 2007, reduced the number of subnational governments and reinforced the authority of the central government, especially with regard to the financing of the health system.The rationale behind significant central government financing was that it gave regions equal opportunity to finance healthcare services.Such an arrangement partly departs from arguments supporting fiscal federalism.Indeed, the authority subnational governments have in relation to decisions on expenditure is no longer accompanied by the authority to finance the expenditure with their own revenues.Danish intergovernmental financial relations have, therefore, become skewed, as was the case in Italy until the late 1990s.However, through institutionalized budget cooperation, the central government has been able to exert a certain amount of control over subnational expenditure.

DISCUSSION
By focusing on the role ideas play in policy formulation, evidence has shown why and how Italy and Denmark have adopted different institutional and territorial reforms in relation to their respective health systems.Ideas have been conceptualized as causal beliefs that determine cause-and-effect relationships, define the issues at stake, and prescribe the course of action to effect a solution.Such a definition encompasses both the interpretation of the issues at stake and the elaboration of solutions.These two processes need to be analysed together within the realm of policy formulation.Without neglecting the role of institutional, political and economic contextual factors that have constrained or facilitated change, the article has shown that different changes have been the result of different ideas in relation to the issues of technical and allocative efficiency.Consequently, different solutions have been adopted.
Since its inception and throughout the 1980s, the Italian health system suffered from an inefficient use of resources.Experts have pointed to the need to improve technical and allocative efficiency.In particular, decentralization and fiscal federalism were seen as the right means by which cost-effectiveness, responsiveness and accountability could be improved.These ideas became well established in central government circles and they were gradually implemented during the 1990s, with the NPM discourse playing a crucial role.In particular, the shared view was that the transfer of authority to subnational governments through decentralization was a necessary condition for implementing managerialism.As already documented by Mattei (2006), the so-called 'enterprise formula', namely, the transformation of LHUs into LHEs, can be considered a NPM feature that seemed to offer the remedy for the pathological politicization of the health system.Arguments supporting the benefits of economies of scale were not supported by decision-makers at the national level.However, although no change in the size of subnational governments took place, a process of amalgamation did occur within regions at the LHE and HE level.Nevertheless, unlike Denmark, this was not connected with arguments concerning the need to redistribute competences between levels of government, but was more related to the reorganization of subnational administrative and executive processes at the regional level.During the late 1990s and early 2000s, concerns about the potential pitfalls of decentralization started to arise.In the attempt to strike a balance between technical/allocative efficiency and geographical equity, the coordinating role of the central government was strengthened.These changes have resulted in a partial recentralization of the health system.Such a path of recentralization, which also emerged in other policy areas, for example, the cultural sector (Ferri & Zan, 2014), has been considered as being part of a post-NPM movement.However, on closer inspection, (re)centralization was also a characterizing feature of NPM in its original formulation, in that NPM ideas simultaneously prescribe both more autonomy through decentralization and more central control through centralization (Christensen & Laegreid, 2011).In fact, in Italy, although a partial recentralization took place during the 2000s, 'the organizational arrangements in local health authorities, driven by NPM-reforms, have shown continuity' (Mattei, 2006(Mattei, , p. 1015)).
In Denmark, during the 1970s, considerations in relation to technical efficiency led to subnational amalgamations.The main idea was that larger subnational governments had to be created to save costs through economies of scale.The increase in size of subnational governments was accompanied by a transfer of authority toward the latter.The shared interpretation was that larger subnational governments had to be established to make decentralization effective.Arguments supporting the benefits of decentralization and fiscal federalism were well established in government.During the 1980s, the health system proved to be effective when facing issues of cost containment.Moreover, contrary to the situation in Italy, NPM principles of professionalism and managerialism were already key features of the public sector as a whole during these years.From the beginning of the 1990s, arguments supporting the advantages of decentralization started to be questioned.Geographical differences in service provision between subnational governments that were not seen as a prominent issue during the 1980s became less and less accepted, and a new debate about the need to reform the entire structure of the public sector started.Again, the main arguments that led to the 2007 reform were that changes in size and authority had to go hand in hand.As with the reforms of the 1970s, the benefits of economies of scale for improving technical efficiency were promoted.However, in the health sector, the increase in size of subnational governments was not accompanied by a transfer of authority to the latter as had occurred in the 1970s.Instead, the authority of the central government increased.As in Italy, recentralization in Denmark has characterized other policy areas as well (Mailand & Hansen, 2016).
To conclude, evidence from Italy shows that (1) the predominant policy issues at stake have been technical efficiency, allocative efficiency, geographical equity and financial discipline; (2) during policy formulation, NPM ideas in relation to the institutional dimension (decentralization/ recentralization) have featured prominently; and (3) arguments have mostly pointed to the need to implement institutional reforms (Figure 1).Evidence from Denmark shows that (1) the predominant policy issues at stake have been technical efficiency and geographical equity; (2) during policy formulation, NPM ideas in relation to both the institutional (decentralization/ recentralization) and territorial dimension (upscaling) have featured prominently; and (3) arguments have mostly pointed to the need to implement both institutional and territorial reforms (Figure 1).Table 2 specifies the key findings with regard to the institutional and territorial reform processes of the health systems in Italy and Denmark at the national level.

CONCLUSIONS
The aim of this article has been to explore the arguments supporting changes within the institutional and territorial dimensions of both the Italian and Danish health systems.Whereas the institutional dimension is concerned with authority and its reallocation between different levels of government, the territorial dimension is related to the (re)definition of the size of subnational governments.The article has shown that the relationship between changes in size and authority vis-à-vis NPM ideas is not clear-cut.In Italy, arguments developed by decision-makers in central government with regard to the need to change the allocation of authority were not tied to a discourse about the size of subnational governments (or health agencies).On the other hand, in Denmark, discourses about size and authority have run in parallel.Moreover, in Denmark, it has been shown that the increase in size of subnational governments could be accompanied by either the decentralization or recentralization of authority.In exploring the linkages between size and authority, issue interpretation and the elaboration of solutions during policy formulation are crucial factors explaining different institutional and territorial changes.
The empirical findings of this study confirm that NPM ideas have penetrated governments and administrations widely following an implementation path that varies from one national context to another (Esposito et al., 2018).Nevertheless, since the 2000s, the traditional NPM model has been in decline and a post-NPM and neo-Weberian model of public sector organization is emerging.However, rather than replacing NPM, this model combines a host of features of the Weberian tradition with traditional elements of NPM (Mazur & Kopycinski, 2018;Zafra-Gómez et al., 2013).

Table 1 .
Health systems in Europe: type and territorial architecture.

Table 2 .
Institutional and territorial reforms in the Italian and Danish health systems: key findings.