Explaining the role of expertise in the state: the case of Italy

ABSTRACT The expert competences of the public bureaucracy are often seen as crucial for the quality and legitimacy of public policies. Yet, the role of expertise in the state varies greatly across countries. How can we explain the position of expert competences in the bureaucracy? The paper examines this question through a historical-institutional analysis of the Italian ministerial bureaucracy. Extending the ‘public service bargains’ framework, it argues that the role of expertise in the state is the result of a bargaining process between civil servants and the politicians that they serve. The paper shows how the peculiar position of expertise in the Italian state – the paucity of technical expertise in the permanent administration, the role of ministerial cabinets as politicized providers of expertise, and the dominance of legal training among civil servants – was the result of a chain of political and bureaucratic strategies and responses.


Introduction
The expert competences of the public bureaucracy are often seen as crucial for the quality and legitimacy of public policies. For Max Weber, expert staff was one of the principal reasons for the superior efficiency of the bureaucratic form of organization (Weber 1947). The analytical capacities of bureaucratic organizations are one of their main sources both of legitimacy with politicians, citizens and other stakeholders, and of autonomy from political interests (Carpenter 2001). Moreover, recruitment to the bureaucracy based on merit rather than connections is key for government performance, such as limiting corruption and inefficiency and improving the quality of policies (Dahlström and Lapuente 2017).
At the same time, the analytical capacities of the government administration vary greatly across countries and organizations, and so do the types of skills and education that dominate (Page and Wright 1999;Peters 2010;Wilson 1989). In comparative perspective, the Italian bureaucracy stands out for the persistent lack of technical expertise among civil servants, the overwhelming dominance of officials with legal training, and for the marginal role of bureaucrats as providers of policy advice relative to ministerial cabinet officials (Cassese 1993;Natalini 2013, 2016;Galanti and Lippi 2022;Kickert 2005). For a long time, observers have pointed to the paucity of specialized expertise in the Italian bureaucracy as a source of poor public policy and implementation, red tape and lack of innovation (Bellucci 1995;Pastori 1997). Yet, this situation has persisted despite numerous attempts to reform the civil service. This paper therefore asks: How can we explain the peculiar pattern of expertise that we observe in the Italian state?
Existing studies of the development of the Italian state have examined related questions, such as the fate of administrative reforms (Kickert 2005;Ongaro 2011) and the growing role of ministerial cabinets within the executive (Di Natalini 2013, 2016). Yet, this is the first study to focus squarely on the position of expert competences within the Italian bureaucracy. We here use the terms 'expertise' and 'expert competences' broadly to denote the whole range of different knowledge and skills that bureaucrats may possess. Expertise can be of different kinds, including specialized/technical expertise pertaining to a specific policy field or issue, generalist skills applicable to any issue, political process knowledge or managerial/leadership competences.
To analyze the role of expertise in the state, we develop a theoretical argument that extends Hood and Lodge's (2006) 'public service bargains' framework. We argue that the position of expert knowledge in public bureaucracies reflects the explicit or implicit agreements or understandingsi.e. 'bargains'between public servants and the politicians that they serve. The notion of a bargain implies that the parties to the agreement get something in exchange for something else. Yet, exactly what is exchanged varies. Bureaucrats may bring either political loyalty or particular competences to the bargain, and in return they may be granted responsibility and discretion or material or immaterial rewards. Moreover, bargains may persist over time but may also be subject to renegotiation and change. Patterns of continuity and change in bargains are the result of both exogenous shocks and endogenous dynamics, which induce politicians and bureaucrats to pursue strategies to preserve or alter the existing bargain.
We apply this argument through a historical-institutionalist analysis of the public service bargains struck between politicians and bureaucrats in the Italian stateand the place of expert competences in those bargainsfrom 1945 to 2010. The analysis focusses on the ministerial bureaucracy. We trace developments from the reconstruction of the Italian administration after World War II and the collapse of fascism, via efforts at competence reform in the 1970s, to attempts to change the place of expertise in the state after the breakdown of the party system in the early 1990s. The historical-institutionalist analysis is based on secondary sources in Italian and English, including social-scientific and historical studies and accounts from involved actors.
The paper makes three contributions. First, it makes a theoretical contribution to our understanding of the role of expertise in public bureaucracies. Our 'public service bargains' argument goes beyond existing principal-agent and institutional and sociological arguments about expertise in the state, by emphasizing the different bargains that can be struck between politicians and bureaucrats and how the bargain chosen conditions the role of expertise in the administration.
Second, the paper contributes to historical-institutionalist scholarship on changes in the state (e.g. Bezes and Lodge 2015;Bezes and Parrado 2013). While several existing studies of state development rely on historical institutionalism as an explanatory approach (e.g. Di Natalini 2013, 2016;Ongaro 2011), they point to generic historical-institutionalist mechanisms and do not offer a theory about the specific institutions and interaction between institutions that shape over-time developments. Our paper combines a historical-institutionalist approach with a theory of public administration, specifying the institutions, actors and strategies that explain change and continuity in the bureaucracy.
Third, we make a specific contribution to the literature on Italian public administration. The novelty lies not only in the focus on bureaucratic expertise. It also lies in the application of public service bargains as a theoretical lens to interpret the Italian case. While several studies speak casually of 'pacts' or 'exchange' between politicians and administrators and of 'negotiated change ' (e.g. Cassese 1984;Di Mascio and Natalini 2013;Ongaro 2011), our study systematically applies a theoretical argument centred on this 'bargained' character of political-administrative relations.
The rest of the paper is structured as follows. In the next section, we discuss existing perspectives on expertise in the state and present our public service bargains perspective and argument about the dynamics of continuity and change. We then apply our argument to the Italian case, by tracing the development over time of the position of expert competences in the Italian public service bargain. We conclude by discussing and reflecting on the implications of our findings.

Theoretical discussion
Existing perspectives on expertise in the state Scholars have long highlighted the importance of the expert knowledge of the civil service for public policy-making. Max Weber saw the expert training of bureaucrats as a key feature of the bureaucratic model of organization and as a primary source of the power and efficiency of bureaucracies (Weber 1947). Expert knowledge is also central to the principal-agent model of public administration: Politicians delegate responsibilities to bureaucrats partly because of their specialized expertise, but this expertise simultaneously makes it difficult to control bureaucrats (Bawn 1995). Likewise, in the reputational theory of public administration, the analytical capacities of bureaucratic organizations are a key basis for their reputation and autonomy (Carpenter 2001). Moreover, the type of skills and knowledge in an administration can have a profound effect on decision-making and the content of public policies (Peters 2010;Wilson 1989).
However, existing scholarship has failed to offer convincing accounts of what determines the role and type of expert knowledge in public bureaucracies, and of the processes that bring about change in bureaucratic expertise.
Principal-agent scholars have argued that the level of expertise in a bureaucracy reflects the technical complexity of its tasks. When delegating responsibilities to bureaucrats, politicians gain access to the agent's knowledge about policy consequences that can improve decisions but incur the costs of bureaucrats potentially deviating from the principal's preferences. When technical uncertainty is higher, the gains from delegating tasks to specialized experts are greater (Bawn 1995). Yet, this argument cannot explain why national administrations with tasks of similar complexity often possess very different levels of specialized expertise (Christensen 2017;Weir and Skocpol 1985;Wilson 1989). Another argument derived from principal-agent models is that bureaucrats' investment in expertise depends on civil service institutions. A civil service that offers job security will induce policy-interested bureaucrats to invest in task-relevant expertise, which in turn will induce the political principal to grant discretion to the bureaucracy (Andersen and Moynihan 2016;Gailmard and Patty 2007). Yet, as our case study will make clear, job security is no guarantee for the development of bureaucratic expertise; under some conditions it may even be detrimental to expertise.
In institutional and sociological scholarship, the position of expertise in government bureaucracies is often attributed to how knowledge produced through external academic and professional training filters into and is shaped by state institutions. The varying position of specific types of expert knowledge in government bureaucracies thus depends on features of administrative institutions, such as the openness of recruitment (Fourcade 2009;Weir and Skocpol 1985) or the institutional ties between bureaucratic organizations and academic disciplines (Chwieroth 2010). Although these contributions highlight important factors, they neglect the specific significance of expertise as an element of political-administrative relationships.

A public service bargains perspective
The paper goes beyond these existing perspectives by proposing a theoretical argument about expertise in government bureaucracy that builds on the 'public service bargains' framework (Hood and Lodge 2006). The core idea of this framework is that public administrations rest on explicit or implicit agreements -'bargains'between public servants and those they serve, that is, politicians, citizens, clients, etc. Public service bargains are institutions: they are both formalcodified in constitutions or civil service regulations or addressed in case lawand informal, based on unwritten conventions or understandings. In a public service bargain, the parties to the agreement get something in exchange for something else. 'Politicians', Hood and Lodge (2006, 7) argue, normally expect to gain some degree of political loyalty and competence from bureaucrats or public servants, and those public servants normally expect to gain some assured place in the structure of executive government, a definite sphere of responsibility and some mixture of tangible and intangible reward.
Yet, the relative emphasis on these elements varies between countries and organizations and over timeand so do the specific loyalty, competences, responsibility and rewards involved.
In public service bargains, bureaucrats can provide different kinds of competences, including specialized/technical expertise relevant to a specific policy field and derived from specialized training and experience, generalist competences applicable to any policy dossier (e.g. analysis and problem-solving skills), knowledge about how to navigate the political process, and managerial competences, i.e. the ability to lead organizations, formulate strategies and achieve objectives (see Kruyen and Van Genugten 2020;Lodge and Hood 2005). The kind of qualifications required for bureaucratic jobs also vary, ranging from academic degrees to training in particular schools or passing entry exams. Moreover, the degree of autonomy granted to bureaucratic experts differs.
Hood and Lodge distinguish between two main types of public service bargains: 'agency' and 'trustee' bargains. In agency bargains, public servants are regarded as the loyal servants of the political principal and subject to the principal's direction and control. In trustee bargains, by contrast, public servants 'possess a domain of autonomy in which they exercise discretion in a way that is not subject to commands or control from elected politicians' (Hood and Lodge 2006, 25) and are loyal not only to the sitting government but also to the public good. One important form of trustee bargain is centred on the expert competences of civil servants, in which public servants are granted considerable autonomy and discretionary power in exchange for providing specific technical expertise.
The public service bargains framework thereby helps us conceptualize the varying role of expertise in the public servicethat is, the 'dependent variable' of our study. But it also offers a theoretical basis for thinking about the dynamics that can explain continuity and change in expertise in the bureaucracy over time. Public service bargains are not static: bargains are forged and can persist over time, but they can also be renegotiated or replaced by new bargains (Hood and Lodge 2006, chap. 8). Existing work has suggested that the dynamics of change in public service bargains can be analyzed using a historicalinstitutionalist explanatory approach (Bezes and Lodge 2015; Bezes and Parrado 2013), but has not fully explored how the perspectives can be integrated.

Combining public service bargains and historical institutionalism
Going beyond existing work, the paper combines historical-institutionalist insights and the public service bargains perspective to explain over-time developments in bureaucratic expertise. These two perspectives are theoretically compatible: While the public service bargains framework takes a strategic-action view on political-administrative relations, it sees these strategic interactions as taking place within institutions that have both formal and informal aspects and that are shaped by history and vary across countries (Hood and Lodge 2006, 14, 195). This fits well with the historical-institutionalist view that institutions vary across countries and time and affect behaviour by shaping the resources, incentives and ideas of actors (Pierson 1993;Skocpol and Pierson 2002). The two perspectives also usefully complement each other: While historical institutionalism proposes generic mechanisms for understanding institutional change, the public service bargains perspective offers a specific theory of political-administrative relations.
The paper makes two main arguments to explain change and continuity in bureaucratic expertise. The first is an argument about endogenous dynamics. Historical institutionalists generally claim that institutional development is characterized by path dependent processes, where past institutional choices shape choices further down the line (Mahoney and Schensul 2006). Under the broad heading of path dependence, a key argument is that institutions produce politics: the institution in place shapes political choices by providing actors with specific resources, incentives and information (Pierson 1993). Applied to public service bargains, we argue that bargains, once struck, will shape the subsequent politics of expertise in the public service. More specifically, the character of an existing public service bargain will condition the strategies pursued by political and bureaucratic actors regarding expertise in the public service (cf. Bezes and Lodge 2015, 145).
This may take the form of what historical institutionalists call a self-reinforcing or 'increasing returns' process: once in place, an institution tends to create positive feedback that makes the institution increasingly difficult to change (Pierson 2000). For instance, actors who benefit from the institution will defend the status quo. Applied to public service bargains, existing bargains may induce political and bureaucratic actors to pursue strategies that further entrench the bargain. For instance, a bargain may empower and benefit certain groups of bureaucrats, who then use this power to preserve their privileges.
But bargains, once struck, may also set in motion processes that lead to change of the bargain, by prompting reactions that eventually undermine or modify the existing bargain. This is what some historical institutionalists have labelled 'negative feedback' effects: Institutions may become so dysfunctional or costly that they provoke a backlash from important actors (Weaver 2010). An existing public service bargains may for instance fail to provide politicians with the necessary expertise to formulate policy, and may therefore lead politicians to pursue strategies to reform or replace the bargain.
The second argument is about exogenous shocks. In historical-institutionalist accounts, major institutional changes usually come about during 'critical junctures', which are caused by exogenous shocks that challenge fundamental aspects of an institution and expand the menu of possible options for change (Capoccia and Kelemen 2007). Building on this, we argue that the role of expertise in the public service may change if external developments undercut the legitimacy of existing public service bargains in the eyes of politicians or bureaucrats. These external shocks can be of different kinds, such as major political upheavals, including regime change, party system change or dramatic election outcomes (Hood and Lodge 2006, 154), economic crisis, intellectual changes such as the emergence of new professions or expert ideas in academia (Christensen 2017), or pressure from international bodies. External shocks can alter the views of politicians and bureaucrats about what kind of competences are needed in the public service, and about the loyalty and autonomy of bureaucratic experts. This may lead political or bureaucratic actors to pursue strategies to alter existing bargains. Yet, the success of these strategies will depend on the response from the other party to the bargain.
Politicians and bureaucrats may pursue a wide range of strategies in response to endogenous features of bargains or exogenous shocks, aimed either at altering or preserving the existing role of expertise. Politicians may for instance seek to change the type of competences administrators possess by reforming recruitment policies (Christensen 2017), set up alternative expert bodies or units inside or outside the bureaucracy (Campbell and Pedersen 2014), or limit the autonomy or responsibilities of bureaucrats. Bureaucrats may for instance seek to strengthen particular types of expertise within their department by altering recruitment and promotion practices, or conversely, to preserve the existing bargain by entrenching specific recruitment practices or ways of working. Importantly, political and bureaucratic action can be met with different responses from the other side, which range from acceptance to superficial adherence to rejection. For instance, civil servants may either embrace or sabotage new recruitment policies.
To sum up, in contrast to previous work, the paper offers a theoretical argument that sees expertise in public administrations as resting on bargains between politicians and bureaucrats. And going beyond existing accounts, it proposes a theoretical argument about the specific dynamics driving continuity and change in expertise in the bureaucracy that combines historical-institutionalist reasoning with a public service bargains perspective. In the next section, we apply this argument to the Italian case.

Administrative traditions
In comparative typologies of administrative traditions or models, Italy is usually described as belonging to the Napoleonic administrative tradition, which also includes France, Spain, Portugal and Greece (Ongaro 2008;Peters 2008), or to a more specific Southern European administrative model, encompassing the same countries except France (Kickert 2011;Sotiropoulos 2004). The Italian state that was established after unification in 1861 waslike other Southern European statesinspired by the Liberal constitutional Rechtsstaat set up by Napoleon I in France in the early nineteenth century. A core feature of the Napoleonic administrative model was its formalism and legalism: law played a primary role in the state and strictly defined the activities of the administration (Kickert 2011). But the Southern European administrations also came to be defined by other common features: strong politicization at the topeither through political appointments of top civil servants or the role of ministerial cabinetsand pervasive clientelism, through the awarding of public sector jobs in exchange for political support (Sotiropoulos 2004). In the following, we will show how these features found expression in the specific public service bargains forged within the Italian state in the post-war era and how they shaped the development of expertise in the administration.
The forging of the post-war public service bargain  The collapse of the fascist state and the advent of the First Republic The end of WWII and the collapse of fascism represented an exogenous shock that delegitimized the prewar Italian state and made way for a new regime (Kickert 2005;Sepe 2018). These events were followed by the 1946 institutional referendum that replaced the monarchy with a republic, the election of the Constituent Assembly, and the subsequent creation of the First Republic.
The entry into force of the Constitution in 1948 also marked the emergence of the first public service bargain analyzed here (Lewanski and Toth 2011;Pastori 1997;Sepe 2018). Most scholars highlight that the Constituent Assembly did not pay much attention to the administration. Few articles in the Constitution address it directly and, when they do, they are vague and abstract (Allegretti 1984;Pastori 1997;Sepe 2018).
Regarding the loyalty of civil servants, the Constitution sought to free the administration from political control. Politicization and relationships of personal loyalty had been the norm in the prewar period (Cassese 1984;Sepe 2018). According to the new Constitution, bureaucrats had to be loyal exclusively to the rule of law and to the collectivity. Separation between politicians and civil service should be the new normal, meaning no political control over bureaucratic activities, political patronage, or favouritism through administrative action (Cassese 1984;Pastori 1997;Sepe 2018). This separation was also expressed in the notion that the administration was responsible for implementing policies and should not be involved in policymaking (Allegretti 1984;Pastori 1997).
In the Constitution and connected legislation there was no mention of the competences that should characterize the bureaucracy. Nevertheless, a novel element was the notion that the administration should be a functional entity: effective, efficient, and result-oriented (Allegretti 1984;Pastori 1997;Sepe 2018). The bureaucracy had the responsibility to deliver results and therefore needed relevant competences and protection from politicization (Allegretti 1984). Thus, the ambition for bureaucratic expertise can be deduced from the principle of functionality.
To ensure independence and competence, additional rules about personnel specified recruitment and career arrangements. Civil servant careers were subjected to administrative law and not to the rules of private employment, which entailed substantial job security. Moreover, civil servants had to pass a public examination to enter the administration: the concorso. Personnel was organized hierarchically, career advancement was based almost exclusively on seniority, and the entire career path had to take place in the ministry in which the competition was won. Therefore, personnel regulations entailed low internal mobility and closure towards the outside. Bureaucrats had a clear path to the top and could be sure about their future income and benefits (Cassese 1984;Lewanski and Toth 2011;Pastori 1997). These rules did not represent a radical departure from previous legal arrangements. The Constituent Assembly favoured a liberal administrative model, that is, along the lines of pre-fascist arrangements (Dente 1977;Pastori 1997;Sepe 2018).
Loyalty and rewards over competence and responsibility However, a profound discrepancy existed between what was stipulated by the Constitutionand connected legislationand actual practices. The new constitutional ideals of competence, meritocracy, and functionality were not followed up in practice. Since entry examinations relied on old and inadequate evaluation techniques, such as open questions, systematic and objective evaluation of candidates was difficult (Giannini 1979;Lewanski and Toth 2011;Pastori 1997). Additionally, knowledge evaluation did not affect career advancement, which was primarily based on seniority (Cassese 1984). Legal education was predominant (more on this below), and the concorso and the career advancement system erected high barriers to entry for experts from other disciplines.
Also, the attempts to provide (prospective) bureaucrats with task-relevant competences were few and unfruitful. In 1957, the government passed a reform to establish a school of public administration to offer training to future bureaucrats. Yet, in the end, the plan was scaled back: the schoolthe Scuola Superiore della Pubblica Amministrazione was only made open to those already working in the administration and only supplied basic knowledge to low-ranking bureaucrats (Sepe 2018, 144). Its contribution to building competences in the civil service was thus modest compared to the major role played by for instance the École Nationale d'Administration and similar schools in France in providing the bureaucracy with expert personnel.
With regard to responsibilities, civil servants enjoyed very limited effective autonomy, only executing decisions and carrying out routine activities (Kickert 2005;Lewanski and Toth 2011;Pastori 1997). In other words, in Italy both key elements of a trustee-type competence-based public service bargainbureaucrats providing expertise in exchange for policy-making responsibilitiesfailed to materialize in the immediate post-war period.
Instead, the public service bargain that emerged was an agency bargain where loyalty was exchanged for rewards. Concerning loyalty, politicization continued to be pervasive and administrative actions were subjected to the influence of politicians, who could exert control over the administration. Clientelism was widespread (Cassese 1993;Kickert 2005;Lewanski and Toth 2011). For instance, while recruitment took place through the concorso, recruiters were often biased (Giannini 1979). The main attitude of bureaucrats towards politicians was one of submission and acquiescence.
In return, bureaucrats gained job security and employment benefits such as career advancement based on seniority (Battistelli and Amendola 1995;Bellucci 1995;Cassese 1984). Bureaucrats did not oppose clientelism, but instead exploited it to get employment guarantees (Lewanski and Toth 2011;Pastori 1997). Rewards and career progression were highly predictable. In fact, predictability and safety were what attracted bureaucrats-to-be in the first place (Aberbach, Putnam, and Rockman 1981;Kickert 2005). On average, civil servants entered the administration at twenty-two and stayed for more than thirty years. Integration of new personnel at intermediate or advanced levels almost never occurred (Aberbach, Putnam, and Rockman 1981;Cassese 1984).
Explaining the post-war public service bargain: political and bureaucratic strategies Why did a public service bargain centred on expertise fail to materialize in the post-war years? The prevalence of loyalty over competence can be seen as the result of the interplay of political and bureaucratic strategies.
On the political side, the conservatism and inexperience of the new political class prevented a clean break with pre-war practices. The Constituent Assembly and the first government of the new Republic promoted continuity with past institutions instead of pushing for change (Allegretti 1984;Dente 1977;Lewanski and Toth 2011;Sepe 2018). Antebellum arrangements fostered legalism and political control over the administration. Personalistic relationships between politicians and civil servants persisted and expertise, autonomy, and functionality were neglected (Lewanski and Toth 2011;Sepe 2018). Continuity was reinforced by the inexperience of the new political class, which was unable to force the adoption of the new norms of the Constitution (Cassese 1993;Putnam 1973).
On the bureaucratic side, several factors worked against change. First, there was considerable continuity in personnel from the pre-war state. Most bureaucrats had already been employed for some time and strived to maintain old practices and modi operandi (Dente 1977;Ginsborg 2006;Putnam 1973). Second, the administration was heavily dominated by officials from southern Italy, for whom '[s]ecurity of posts and careers [were] held at a premium as compared to other possible values' (Aberbach, Putnam, and Rockman 1981;Lewanski and Toth 2011, 7-8). This manifested itself in a quest for employment guarantees, preference for continuity, legalism, opposition to productivity and meritbased promotion, and disinterest in expertise and responsibility (Cassese 1984;Lewanski and Toth 2011). Third, the absence of a public administration school forming the bureaucratic elite impeded pre-socialization into values of professional administration (Pastori 1997). Civil servants instead absorbed preexistent modi operandi and perpetuated attitudes reflecting their social background (Cassese 1984).
Entrenchment of the bargain and attempts at reform  Politicization, legalism and the rise of ministerial cabinets The 'loyalty-for-rewards' bargain formed after the war set in motion a chain of political and bureaucratic strategies and responses, which conditioned the role of expertise in subsequent periods. First, the stronger party system that emerged after the war reinforced the politicization of the administration (Battistelli and Amendola 1995;Sepe 2018). Through political patronage, politicians could take over the powers and functions of higher bureaucrats. Hence, they sought loyalty rather than expertise from the administration (Battistelli and Amendola 1995;Cassese 1984;Sepe 2018).
Second, as politicians took over administrative functions, the autonomy and influence of bureaucrats in policy-making were severely restricted. Their responsibilities became limited to routine activities, making Italian bureaucrats far less powerful than their international counterparts (Aberbach, Putnam, and Rockman 1981;Bellucci 1995). This caused a loss of prestige for bureaucrats, but also led bureaucrats to adopt an obstructionist and highly legalistic approach (Aberbach, Putnam, and Rockman 1981;Cassese 1984). Indeed, further strengthening legalism was the way bureaucrats had to compensate for their loss of power. By imposing strict rule following, they were able to keep some degree of control (Bellucci 1995). For the same reasons, they dragged their feet and sometimes impeded policy implementation (Cassese 1984). The predominance of legalism was also fostered by the dominance of legal corps in the administration. In the early 1970s, around 54% of higher bureaucrats had a degree in law, compared to 37% in social sciences and 10% in natural sciences. This trend was reinforced in subsequent years (Cassese 1984;Putnam 1977). Yet, the legal dominance left ministers without much-needed technical expertise, such as financial and economic expertise in the economic policy area (Pastori 1997).
Third, the ministerial bureaucracy's obstructionism and lack of technical expertise led politicians to look elsewhere for these capacities. Partly, politicians sought to make up for the shortcomings of the administration by empowering external agencies, which were supposed to be more functional and efficient (Cassese 1984;Kickert 2005;Sepe 2018). Equally important was the expansion and empowerment of ministerial cabinets, which allowed politicians to bypass the bureaucracy (Di Mascio and Natalini 2016;Kickert 2005). Politicians transformed cabinets into an intermediary between themselves and bureaucrats. Cabinets came to occupy a more prominent role than the regular administration in the ministerial hierarchy. Their task was to translate political will into administrative action and to oversee the bureaucracy, whose autonomy of action was thus further limited. Furthermore, cabinet officials were personally appointed by politicians; they were personal loyalists who followed their master from ministry to ministry. While cabinet officials did provide some expertise, it was mostly of a legalist kind and politicians often lamented a lack of technical competences (Bellucci 1995;Cassese 1984;D'Auria 1981D'Auria , 1995. Both external agencies and ministerial cabinets were politicized and their appeal was not so much technical expertise as their ability to increase politicians' power. These bodies thus further entrenched the key elements of the post-war bargain: the emphasis on loyalty over competence increased and the autonomy and responsibilities of bureaucrats were further reduced.

Social conflict and attempts at administrative reform
In the 1960s-70s, Italy underwent a period of intense social conflict, including student protests, large-scale strikes and political terrorism (Montanelli and Cervi 1991;Sepe 2018). These events were coupled with growing criticism about the inefficiency, lack of responsiveness, and lack of competence of the state administration (Ginsborg 2006). This delegitimized the existing bargain and set off a push for administrative reform.
In 1972, the Andreotti government established the managerial career within the public service by decree. Managers were assigned different responsibilities than regular bureaucrats. Managers had to ensure the good functioning of the administration and the achievement of goals, and were supposed to be autonomous and shielded against political influence (Pastori 1997;Sepe 2018). Additionally, managerial skills were required. Bureaucrats could not gain access to the role through seniority, but had to be accepted to a public administration course and then sit a competition. An external organ was to train and select candidates. The decree also included the possibility of appointing people from outside the administration, to facilitate the influx of relevant expertise (Cassese 1984). This was thus an attempt to redesign the public service bargain with the goal of bestowing new responsibilities and autonomy upon bureaucrats in exchange for greater competence.
Yet, in practice, the decree was not fully implemented and managers ended up being like ordinary bureaucrats (Cassese 1984;Pastori 1997). Concerning competence, the people employed as managers often had a legal rather than managerial background. With few exceptions, they were recruited among higher civil servants, who for the most part had a law degree. Moreover, even when they were social scientists, they were mostly graduates in Economia e Commercio, a study that is much more humanistic and less technical than economics studies elsewhere. There was an almost complete lack of people trained in management (Kickert 2005;Putnam 1977). Also, the training that candidates were supposed to join before selection was overlooked and competitive examinations never took place (Cassese 1984;Giannini 1979). As a result, managers did not actually possess managerial competences (D'Auria 1995). In addition, a very limited set of responsibilities was actually transferred to managers from politicians. In the end, managers mainly executed routine activities, much like other bureaucrats (Battistelli and Amendola 1995;Cassese 1984). Therefore, the reform 'that [seemed], at first sight, to establish a different relationship between the political class and the higher civil service (the latter less irresponsible and more open to the outside, the former less engaged in [routine administrative activities])' did not bring about the intended change (Cassese 1984, 54).
Explaining the failure of managerial competence reform: political and bureaucratic strategies Why did the attempt to strengthen the role of competence in the Italian public service bargain fail again? Mainly, this was due to the dynamics set in motion by the existing bargain, which favoured further development along the existing path rather than change (Sepe 2018). Indeed, in the 'loyalty-for-rewards' bargain, politicians attained power and bureaucrats attained employment benefits. As a consequence, both were unwilling to radically alter the terms of the bargain (Lewanski and Toth 2011).
Bureaucrats disliked the values of competition and meritocracy upheld by the reform, since they threatened their job security and career prospects. Furthermore, civil servants resisted the inclusion of managerial expertise in the administration since it was incompatible with legalism (Cassese 1984). The limited supply of new forms of expertise also undermined the efforts at reform. The absence of public administration schools caused a severe shortage of people with managerial competencies (Cassese 1984;D'Auria 1995). This led to the continued reliance on the legal corps even though their generic procedural knowledge could not satisfy the broader expertise needs of politicians. This perpetuated legalism and the disempowerment of the civil service (Cassese 1984;Pastori 1997;Putnam 1977).
Politicians also failed to follow through on their initial aspirations to create a more competent and responsible administration. Politicians never used their rights with regard to the appointment and control of managers to implement the reform; they preferred to retain political control and carry out managerial activities personally (Cassese 1984). Moreover, despite the government instability that characterized those years, the same politicians served in many successive governments, which fostered loyalty relationship between politicians and bureaucrats (Battistelli and Amendola 1995).

Exogenous shocks and attempts to modernize the administration (1990-2010)
Party system breakdown and EU pressure A new reform period was sparked by the political upheaval of 1991-1992. The party system collapsed due to new electoral laws and the Mani Pulite corruption investigations. At the same time, economic hardship sparked dissatisfaction with the politico-administrative system (Di Mascio and Natalini 2013;Ongaro 2011;Sepe 2018). The old expertise bargain was also put under pressure by the European Community, which wanted Italy to adopt a more output-oriented public administration model constructed along the lines of New Public Management (NPM) (Natalini and Stolfi 2012;Quaglia and Radaelli 2007).
Political leaderspartly politicians who had survived the party system breakdown and partly new politicians and partiesresponded to these shocks with a set of administrative reforms. This ushered in a prolonged period of instability of the politico-administrative system (Ongaro 2011, 739). Overall, reforms targeted the office of manager, ministerial cabinets, and the personnel system. The goal was aligning with other Western democracies and implementing NPM models. This meant boosting managerial competences, eradicating politicization to encourage competition and meritocracy, and generating a result-oriented and more responsible public service (Capano 2003;Sepe 2018).
First, reforms aimed to empower the office of manager, whose powers and expertise were completely disregarded in the previous bargain (Di Mascio and Natalini 2013). The responsibilities of managers were restated, clearly laid out, and separated from the managerial activities that could be carried out by politicians. In this way, legislators wanted to impede politicians' appropriation of managerial powers (Bassanini 2000;Ongaro 2011;Pastori 2009). With regard to competence, the position of manager could now be attributed to professionals from outside the administration more easily. The position was also made temporary and had to be characterized by high mobility. By reducing barriers to entry and by encouraging competition, the new rules were supposed to bring about a highly competent workforce (Ongaro 2011;Pastori 1997).
Moreover, reform of ministerial cabinets was meant to restore the powers that bureaucrats had lost to politicians through cabinets. Indeed, in the previous bargain cabinets were politicized bodiesseparate from the civil servicewhich exercised administrative functions. The new rules aimed to dismantle this shadow administration and instead turn cabinets into advisory bodies composed of skilled senior bureaucrats and managers which could supply relevant expertise to ministers. These measures were supposed to decrease politicization, give responsibility back to bureaucrats, and pool competence Natalini 2013, 2016). These ambitions were in line with a 'competencefor-responsibility' public service bargain.
Reform of the personnel system also upheld this type of bargain. The aim was to hire bureaucrats on private-sector-like contracts which were not subject to the jurisdiction of administrative courts (Bassanini 2000). This marked the end of bureaucrats' privileged career system. In order to discourage old and inefficient attitudes, job security was diminished. Remuneration and career advancement were to be based more on performance. A more meritocratic promotion system based on examinations was added to the previously seniority-based system. Moreover, to make the concorso more effective in ensuring the competence of civil servants, more specialized and unbiased recruiters were appointed (Pastori 1997). Thus, competence and responsibility were to be central, replacing the focus on political loyalty and rewards.
However, conformity with these rules was mostly ceremonial. Managers in practice continued to enjoy limited autonomy. The separation between managerial and political responsibilities turned out to be hard to apply, and politicians seized the opportunity to gain control over managerial activities (Pastori 2009;Zoppoli 2008). And although managers had more relevant competences than in the previous bargain, managerial skills were employed only to a limited extent (Adinolfi, Botti, and Vesci 2010;Ongaro and Valotti 2008).
Ministerial cabinets continued to be dominated by legally trained officials and did not open up to other forms of expertise. Even if the cabinets provided legal procedural expertise, cabinet officials were not autonomous experts; their role was highly politicized and they used their expertise mainly to substantiate the choices of ministers. Instead of hiring senior civil servants to cabinets, politicians preferred recruiting political personnel to maintain political control. Loyalty to the minister was necessary to be appointed. Furthermore, bureaucrats continued to have little responsibility (Di Mascio and Natalini 2016). Lastly, even though public contracts transitioned smoothly into their new, privatized form, no substantial change in the behaviour of bureaucrats occurred (Ongaro and Valotti 2008;Zoppoli 2008).
Despite the limited success of the broader reform efforts, pockets of expertise did emerge in the Italian ministerial bureaucracy in this period. For decades, the Bank of Italy, which enjoyed considerable autonomy, had had a near monopoly on economic expertise in the Italian system, with a high-powered research department and close links to academic economists (Quaglia 2005). Yet, in the 1990s, the economic expert capacities of the Treasury were boosted, with the appointment of highly qualified economic academics and professionals to senior ministry positions. (Among them, Mario Draghi was brought to the Treasury as director-general from 1991 to 2001). This upgrading of bureaucratic expertise was made possible by the decline of the old political class and the pressure from Europe, combined with the entrepreneurship of ministers of the treasury who had economic training (Quaglia 2005, 552-556). Treasury economists were important driving forces behind reform of Italian economic policies in the 1990s. Yet, this development was largely confined to the Treasury and did not lead to broader changes in the position of expertise in the Italian public service bargain.
Explaining the failure of reform: political and bureaucratic strategies Why did also this reform period fail to establish a 'competence-for-responsibility' bargain in the Italian bureaucracy? Again, the outcome reflected the interplay of political and bureaucratic strategies and responses conditioned by the existing public service bargain.
The deep-seated legalism of the administration shaped both bureaucratic and political responses to the reform efforts. While many countries abandoned the legal monopoly over the government bureaucracy after WWII, the Italian administration continued to be highly legalist (Kickert 2005). Legal procedures and legal training were mutually reinforcing: the importance of administrative rules and procedures increased the need for officials with a legal background, who in turn emphasized rules and procedures in their work as civil servants (Di Mascio and Natalini 2013;Ongaro 2011). Legalism entailed that every responsibility was translated into procedures and formalities and that civil servants' main activity was following them (Ongaro 2011). This hampered the introduction of new types of competences in the cabinets and permanent administration. Politicians continued to seek cabinet officials with legal competences and showed little interest in other backgrounds, since they considered legal skills most useful in a context completely dominated by administrative law (Di Mascio and Natalini 2016). Also, officials with managerial competences immersed in this context eventually adopted a legalist mindset themselves (Adinolfi, Botti, and Vesci 2010).
Moreover, bureaucrats continued to emphasize job security over responsibility and competence. The centrality of rewards in the existing public service bargain and the job uncertainty that bureaucrats experienced in the reform period reinforced this tendency. Core aims of the personnel reform were undermined by civil service trade unions, which had considerable bargaining power and fought successfully to preserve many of civil servants' employment benefits. In turn, responsibility, meritocracy, and competence came to bear little importance (Di Mascio and Natalini 2013;Natalini and Stolfi 2012).
If bureaucrats were reluctant to abandon their rewards, politicians were equally unwilling to give up political control and the emphasis on loyalty over competence. Not only did politicians claw back functions assigned to managers. They also chose politically loyal officials when forming ministerial cabinets, since they could use them to control administrative activities (Di Mascio and Natalini 2013). Politicians were reluctant to incorporate bureaucrats in ministerial cabinets, since they could not be controlled to the same extent. Moreover, politicians predominantly selected cabinet members who had staffed cabinets prior to reform, partly because the new political class needed their experience to navigate the administrative system (Bull and Pasquino 2007;Natalini 2013, 2016). Again, this contributed to continuity rather than change in competences in the public service.

Discussion and conclusion
The role of expert competences in the Italian state has come to be characterized by three particular features: the relative lack of specialized expertise in the permanent administration, the key role of ministerial cabinets as providers of expertise, and the overwhelming dominance of legal training among senior civil servants and cabinet officials. Thus, in the Italian 'public service bargain', competence has been given little emphasis relative to loyalty, expertise has been politicized rather than autonomous since located in cabinets rather than in the permanent bureaucracy, and legal knowledge has trumped other forms of knowledge. This sets the Italian case apart from many other Western countries, where competence has been a leading concern in the staffing and organization of the state, where experts in the bureaucracy have been granted considerable autonomy to formulate public policies, and where other forms of knowledgee.g. economic expertise or multi-disciplinary public administration traininghave become dominant in the civil service (Christensen 2017;Page and Wright 1999;Peters 2010).
How can we explain the peculiar position of expert competences in the Italian state? We have put forward a historical-institutionalist argument about the evolution of public service bargains to account for this pattern. We have argued that continuity and change in the role of expertise in the state are the result of political and bureaucratic strategies adopted in response to endogenous features of existing public service bargains or exogenous shocks. We have traced the limited attention to expert competences in the Italian bureaucracy back to the public service bargain struck after WWII, which was based on political loyalty and rewards such as job security rather than on competence and responsibility. This initial bargain shaped subsequent political and bureaucratic strategies in a way that worked against attempts to increase the expertise of the civil service: bureaucrats resisted competence reform to preserve their employment benefits and career prospects, and politicians were unwilling to put bureaucratic competence ahead of political control over the administration.
The key role of ministerial cabinets as expertise providers can be seen as the result of a similar chain of political and bureaucratic moves. Since the post-war bureaucracy lacked expert competences and took an obstructionist attitude to political initiatives, politicians turned elsewhere for expert advice. This led to the empowerment of ministerial cabinets, which expanded dramatically over time and came to include a large corps of legal experts. The ascent of cabinets as providers of knowledge further reduced bureaucrats' role as expert advisers. Significantly, the fact that this expertise ended up in the cabinets rather than in the regular ministerial administration meant that it was subject to greater political control. Thus, while this process did lead to the development of expertise within the Italian political-administrative system, this expertise was politicized rather than autonomous.
The persistent dominance of legal knowledge and the failure of other forms of knowledge to emerge within the Italian state can partly be attributed to self-reinforcing processes. In the post-war administrative state, civil servants were predominantly trained in law, which led to great emphasis on administrative rules and procedures. This resulted in a highly legalized administrative system, which in turn increased the demand for legal expertise (Ongaro 2011;Pastori 1997). This dynamic also involved politicians, who predominantly recruited legal staff to the ministerial cabinets because they saw this as the crucial competence for navigating the state apparatus Natalini 2013, 2016). Yet, this is only part of the story. After all, other countries were able to break the 'legal monopoly' (Juristenmonopol) in the bureaucracy in the period around WWII (Kickert 2005). This was the result both of active political strategies to put other experts in charge, for instance by systematically replacing lawyers with economists in strategic bureaucratic positions, and of the emergence of new expert professions eager to enter the bureaucracy (Christensen 2017). In Italy, both factors were largely absent: instead of seeking to replace existing bureaucrats with officials with different competences, politicians bypassed the bureaucracy by empowering ministerial cabinets. And the supply of new forms of expertise was limited, in part due to the absence of public administration schools.
Several of the factors discussed in our narrativelegalism, politicization, clientelism are common to other Southern European administrations (Kickert 2011;Sotiropoulos 2004). Does that mean that the dynamics observed in Italy can be generalized to these countries? Some mechanisms may travel well to other cases, such as the mutually reinforcing effect of a legalist administrative system and the dominance of legally trained staff. But other dynamics may be specific to Italy. For instance, whereas politicization in Spain, Portugal and Greece has been expressed in political appointments of top civil servants, in Italy bureaucrats have been more isolated from politicians and politicization has manifested itself mainly in the outsize role of ministerial cabinets (Kickert 2011). This has had a distinct effect on the place of expertise in the Italian state, which will likely look different in other Southern European countries where politicization has taken a different form. However, further comparative research is needed to establish which mechanisms are shared and which dynamics differ across these cases.
The explanatory argument offered in this paper differs subtly but significantly from existing theoretical accounts of expertise in state bureaucracies. Principal-agent accounts of bureaucratic expertise have highlighted the importance of job security for bureaucratic investments in task-relevant expertise and, in turn, for politicians granting autonomy to bureaucrats (Andersen and Moynihan 2016;Gailmard and Patty 2007). We see the opposite dynamic in the Italian case: While job protection was meant to ensure meritocracy by limiting political control, it effectively raised barriers to competence reform since bureaucrats resisted changes that might hurt their employment security and career prospects. This had to do with the social and educational profile of the Italian bureaucratic class inherited from the pre-war era, which was stuck in old ways of working and legal procedures and had limited interest in policy. Since bureaucrats failed to deliver task-relevant expertise, politicians were also unwilling to grant them greater responsibilities; instead they severely limited bureaucrats' role in policymaking. In other words, the outcome was the opposite of what we would expect from the principal-agent argument.
Existing institutional and sociological scholarship, on the other hand, has highlighted how the interaction between academic institutions and state institutions shapes the role of expertise in the state. This account finds partial support in the Italian case, including in how a closed bureaucratic recruitment system and the lack of educational institutions producing graduates with non-legal knowledge impeded the inflow of new forms of expertise (see Fourcade 2009;Weir and Skocpol 1985). Yet, this argument misses how the specific nature of the relationship between politicians and bureaucrats that emerged in Italy conditioned political and bureaucratic strategies that shaped the role of expertise, such as the chain of actions and reactions that resulted in the empowerment of ministerial cabinets as expertise providers.
Moreover, our analysis builds on but also goes beyond existing accounts in the literature on Italian public administration. One contribution is the focus on expert competences, which offers a new angle on the development of the Italian state. While the question of expertise is linked with questions of administrative reform (Kickert 2005;Ongaro 2011), focusing on expertise provides a more targeted approach for thinking about one of the core challenges of any public administration. Our analysis also draws on studies of the growing role of the ministerial cabinets within the executive (Di Natalini 2013, 2016), but offers a broader frame for understanding the peculiar place and advisory function of cabinets in the Italian system.
A second contribution is the application of a public service bargains perspective to understand the Italian case. While existing studies have mentioned in passing the pacts or exchange between Italian politicians and bureaucrats (Bellucci 1995;Cassese 1984;Di Mascio and Natalini 2013;Ongaro 2011), our analysis shows the utility of thinking systematically about the bargaining between politicians and bureaucrats over the loyalty, competence, responsibility and rewards of civil servants.
Lastly, we contribute to historical-institutionalist scholarship on changes in the state (Bezes and Lodge 2015), by combining a historical-institutionalist approach with a substantive theory of public administration. There is already a considerable literature on how administrative legacies shape trajectories of public administration reform, including in countries with a Napoleonic administrative tradition (Bezes and Parrado 2013;Kickert 2011;Ongaro 2008). Yet, to explain processes of institutional change over time, it is not enoughas many studies doto point to generic historical institutionalist mechanisms such as self-reinforcing processes or critical junctures. We also need a theory about the specific institutions, actors and strategies involved. With our historical-institutionalist argument about public service bargains and political and bureaucratic strategies we have attempted to offer exactly that. This explanatory framework can be applied widely in future comparative-historical research on the role of expertise in government administrationand on changes in the state more broadly.