Advancing the theory and practice of public sector reform through the analysis of social mechanisms

Since the early 2000s, a strand of research works in public management has systematically resorted to the analysis of social mechanisms (basically as defined in Hedström and Swedberg, 1998 and Gambetta, 1998 – more on defining issues below) to improve our understanding of the dynamics of public sector reform processes. This chapter appraises the adequacy of the analysis of social mechanisms as conceptual building blocks for an improved theoretical understanding of organizational reform processes in the public sector. The main questions can be formulated as follows: are social mechanisms ‘adequate’ building blocks for a form of theoretical understanding of public sector reform processes, complementary to other theories used to study such processes? And if so, how can they be employed for bettering our understanding of the dynamics of reform processes: what is the ‘added value’ of the analysis of social mechanisms? And finally, in a more prescriptive fashion, how can such knowledge be used for public sector reforms to be ‘successfully’ carried out by policy-makers? The key thrust of the research agenda outlined in this chapter, which is aimed at improving our understanding of the dynamics of public sector reform by resorting to the systematic usage of social mechanisms, lies in identifying developmental patterns (Van de Ven, 1992) of public sector reform processes: patterns that can be understood in terms of concatenations of social mechanisms, which may be manipulated – triggered of defused – so as to achieve desired public sector reform goals. In terms of methods, this work is partly theoretical-speculative and partly it proceeds by revisiting already published works (e.g., Ongaro, 2006, 2013; the Special Issue of Governance guest edited by Barzelay and Gallego, 2010a,

2010b; Asquer, 2012) in the light of the research questions and frame of analysis worked out in this chapter.
The chapter unfolds by, first, addressing issues of definition, then by reviewing the scientific literature on the analysis of social mechanisms for the study of public sector reform processes. The adequacy of social mechanisms analysis is then gauged, and the usage of this approach examined by contrasting social mechanisms-based explanations and institutionalist accounts of public sector reforms (the latter being the mainstream in the field). A discussion follows and the main conclusion is that social mechanisms-based explanations and institutionalist accounts chart a largely different terrain and are highly complementary to each other: their combined usage may be beneficial for the advancement of the knowledge about the dynamics of public sector reform processes.

DEFINITIONS
Social mechanisms can be defined as conceptual tools that can be employed in the analysis of complex change processes, like those triggered by reforms of the public sector, tools that -it is argued -are capable of revealing the multiple causes of change (Pettigrew, 1990) and specifying the social 'cogs and wheels' of the phenomenon investigated (Elster, 1993, p. 3).
Social mechanisms can be defined (Hedström and Swedberg, 1998, p. 7) as unobserved analytical constructs that provide hypothetical links between observable events: 'a social mechanism is a plausible hypothesis, or set of plausible hypotheses, that could be the explanation of some social phenomenon; the explanation begins in terms of interactions between individuals, or between individuals and some social aggregate' (Schelling, 1998, p. 32). Elster (1989) and Stinchcombe (1991) interpret social mechanisms as 'the building blocks' of an advocated middle-range theorizing that, in their opinion, can provide an important contribution to revitalizing the study of social phenomena, by getting beyond contenting oneself with merely establishing systematic covariation between variables or events, rather aiming at investigating the causal texture of social phenomena. This approach has strong roots in sociology, as illustrated by Boudon (1991, revisiting the work of Merton, 1968, pp. 43-4 in particular).
Social mechanisms are generally used in combined ways: concatenation of mechanisms is central in understanding change processes (Gambetta, 1998, p. 105), as '[e]xplanations of most concrete social events or states require resort to several elementary mechanisms; one is not enough. Sometimes, these mechanisms counteract one another, and sometimes they work together' (Hedström and Swedberg, 1998, p. 21). Social processes (McAdam, Tarrow and Tilly, 2001, p. 24) are in this perspective seen as regular sequences (concatenations) of such mechanisms. In particular, purposive organizational change processes Edoardo Ongaro -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:08:40AM via free access (as a subset of the broader social processes), such as those that can be detected in public sector reforms, can be seen as developmental patterns (Van de Ven, 1992) from a certain state of affairs (generally, the current status) to another state of affairs (interpreted as the resultant of the reform itself).
As to the notions of public management reform, administrative reform, and public sector reform respectively, there are various definitions, each emphasizing different aspects and profiles. For the purposes of this chapter, which aims at discussing the employability of social mechanisms for improving knowledge about the dynamics of organizational reforms, also in a 'design science' perspective (i.e., intending knowledge as applicable for intervening on a social system, i.e., for 'managing' a reform), we adopt the following definitions (Kickert and Ongaro, forthcoming; see also Ongaro, Ferré and Fattore, 2015;Di Mascio et al., 2017;Ongaro and Kickert, forthcoming, 2019): • Public sector reforms are deliberate attempts to reconfigure the public sector, or significant portions of it, broadly intended -the regulation and organization of public services at large (thereby including social security, taxation, economic regulation, etc.) -to make the public sector work better, according to given criteria of 'betterment' and 'improvement'. • Administrative reforms are deliberate attempts to reconfigure the public administration of a given jurisdiction (mostly but not necessarily only in the executive politics order), to make it -in some sense -work better. • Public management reforms are deliberate attempts to change the structures and processes of public sector organizations with the objective of getting them (in some sense) to run better (see Pollitt and Bouckaert, 2017).
In a nutshell, public sector reforms encompass a larger phenomenon; administrative and public management reforms concern subsets of the broader public sector reforms. Both are change processes that, under certain conditions, may be conceptualized as a policy process. 1 Throughout the chapter we refer mostly to public sector reforms, unless differently specified. Finally, it may be noted that the analysis of social mechanisms may be employed more broadly beyond the investigation of processes of reform and change (e.g., for the study of the routine, everyday delivery of public services); crucially, the analysis of social mechanisms may represent a key component in framing theoretically one way in which the field of public policy, on one hand, and the field of public management, on the other hand, may be more closely interconnected through an agentic and event-centred theoretical framework that emphasizes how actors may perform in interconnected ways both functions of the policy process and public management functions (Asquer and Mele, 2018); however, this chapter has a more focused scope and the emphasis is here on the analysis of social mechanisms for the study of processes of reform (change management).  Barzelay and Campbell (2003) is a first attempt, with an organizational-level focus, to produce a thick account of an organizational change process embedding, albeit mostly implicitly, an analysis of social mechanisms at work (mostly, the mechanisms of actor certification; see Chapters 3 and 4 in particular). The studied cases are two exercises of strategic envisioning process in the US Air Force. Chapter 5 in particular delineates a procedure for extrapolating practices from case analysis for replication elsewhere of similar patterns of change (which we may conceptualize as developmental patterns aimed at producing, as their 'end state of affairs', an enhanced capability by a public organization to envision the future and adapt to it). This approach aims at making knowledge derived from the analysis of social processesseen as concatenations of social mechanisms -utilizable for 'design science' problems, that is, transforming explanatory knowledge -often originally generated from intrinsic case studies -into actionable knowledge, replicable and applicable for enabling the designing of 'solutions' to tackle extant social problems (Barzelay returned more theoretically on this issue in his 2007 paper, and Ferlie and Ongaro elaborate on this in Chapter 8 of their 2015 book). A more systematic employment of social mechanisms can be found in the Governance Special Issue on the dynamics of public management policy change in selected Southern European countries edited by Barzelay and Gallego. The introduction and conclusion papers Gallego, 2010a, 2010b) provide an overview also pointing to concatenations of social mechanisms that may be the explanation of dynamics of public management policy change. The paper by Mele (2010) is an application to the public sector innovation policy, and in a later paper Mele and Ongaro (2014), albeit primarily interested in profiling the traits of public leadership in reforming the public sector under conditions of frequent government turnover, also resort to the analysis of social mechanisms in their comparison of reforms in the public personnel policy and the innovation policy in Italy throughout 1992-2007. Asquer (2012) picks up these conceptual tools for the study of an organizational-level intervention of redesign of public agencies in the agricultural policy field. In a similar vein, Ongaro (2006) applies the analysis of social mechanisms to the dynamics of devolution processes in legalistic countries (that is, in jurisdictions where administrative law has become a dominant cultural paradigm around which an epistemic community wields a quasi-monopoly of the policy; Capano, 2003 What mechanisms do these case studies detect as being in action (or at least purport to have been in action)? The set of mechanisms that can be encountered in these pieces of research include the following. First, the mechanism of actor certification: it refers to the validation of actors, their performance and their claims by external authorities (McAdam et al., 2001, p. 121). Second, attribution of opportunity and threat: defined as an activating mechanism responsible for the mobilization of previously inert social groups (ibid., pp. 43 and 95); it involves (1) invention or importation and (2) diffusion of a shared definition concerning alterations in the likely consequences of possible actions undertaken by some 'political' actors. Third, threshold-based behaviour about whether to accept or resist (some form of looming) change: it is a mechanism belonging to the class of rational imitation mechanisms that is based on the idea that 'an individual's decision whether or not to participate in collective behavior often depends in part on how many other actors already have decided to participate. . . An actor's threshold denotes the proportion of the group which must have joined before the actor in question is willing to do so' (Granovetter, 1978, elaborated in Hedström and Swedberg, 1998, p. 19 (referencing Granovetter 1978). Fourth, brokerage, which can be defined as 'the linking of two or more previously unconnected social sites by a unit that mediates their relations with one another. . .it can become a relational mechanism for mobilization ' (McAdam et al., 2001, p. 26). Fifth, appropriation of mobilizing structures refers to social spaces put at the service of interpretations of situations and objectives that may be employed to mobilize actors towards certain courses of action (ibid., p. 102).
Interestingly, the probably most famous social mechanism has not been encountered in the (so far limited in number) studies of public sector reforms based on this theoretical perspective: self-fulfilling prophecy, 2 which is a mechanism in the class of rational imitation mechanisms whereby actors are induced to join a certain course of behaviour by the observation of other actors behaving in a certain way (the common example being rumours on a bank's insolvency leading to larger and larger portions of bank account holders to withdraw their savings from the bank, hence ultimately making the bank go bust, even if in the first instance there was no substantiated motive for fearing losing one's savings). It has been debated whether a catalogue or inventory of social mechanisms would be useful (Hedström and Svedberg, 1998;also Hedström, 2005), and if so what criteria should it meet -for example, how do we know it is exhaustive, or should it include pertinent classifications whereby certain mechanisms may be attributed to bear explanatory power 'especially' or 'more often' for certain categories of phenomena?
It may be noticed that other mechanisms are drawn mainly from economic analysis, and their categorization as 'social mechanisms' may be questioned, though they seem to fit, at least if broadly intended, the definition of being Edoardo Ongaro -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:08:40AM via free access capable of 'providing hypothetical links between observable events': this includes the feedback mechanism (mainly drawn from the discipline of cybernetics and applied in economics studies) and the mechanisms of decreasing or increasing marginal returns (widely used in economics), whereby the relative convenience of a certain course of action decreases (or increases) the more the course of action is pursued. The extent to which learning processes should be encompassed in a catalogue of social mechanisms may also be questioned (Dunlop, 2015;Dunlop and Radaelli, 2017).

SOCIAL MECHANISMS AS BUILDING BLOCKS FOR A FORM OF MID-RANGE THEORIZING ABOUT THE DYNAMICS OF PUBLIC SECTOR REFORMS?
We can now turn to addressing the central question about whether social mechanisms are 'adequate' building blocks for a form of theoretical understanding of public sector (administrative, public management) reform processes, a kind of theoretical understanding that may also guide the design of policies of reform? 'Theoretical understanding' means both supporting theory-intensive explanations of extant reform processes (i.e., knowledge guiding the employment of social mechanisms to conduct intrinsic case studies of reform policies in order to understand the inner dynamics of how the processes unfolded) and developing a form of mid-range theorizing about developmental patterns of organizational reform processes 'more generally', by outlining what forms reform processes take, hence potentially what mechanisms can be activated to facilitate or to prevent reforms from happening. By 'design of policies of reform' we mean tackling the question: how to design a 'successful' public sector reform (by triggering the 'appropriate' mechanisms). In short, we adopt an 'instrumental-rational conception of policy design' -conscious on the one hand that the reform of the public sector does not always take the form of a public policy (see Barzelay, 2001 and subsequent works), and on the other hand that the 'instrumental-rational conception' may be only part of the story, and that designing a policy may serve purposes other than identifying means and connecting them to more or less 'desirable' goals: designing a policy may be a way for practitioners to engage in highly mediated processes of sense-making and sense-giving, beyond a linear conception whereby the problem is given before the solution is sought for through the design or redesign of a policy -in short, that the instrumental-rational conception of the activity of policy design is but one way of conceiving policy design (see recently Turnbull, 2017).
Knowledge about concatenations of social mechanisms is in turn instrumental to 'making a reform happen', or preventing it from doing so. Triggering, or defusing certain concatenations of mechanisms may facilitate or hamper certain courses of events associated with the occurrence of a certain reform. Knowledge about the alleged effects of the triggering of certain concatenations of social mechanisms on the unfolding of certain courses of events can be used in both ways: to make a reform happen, or to hinder it. It should be added that this is not an evaluative claim: if a reform is reckoned to bring about more damage than benefit, then preventing it may well be an 'appropriate' course of action. (Though any claim about benefit or damage begs the question, for whom? And it may hence have different answers.) However, sticking to a logic of instrumental rationality, the evaluation of the adequacy of a reform design is made on the bases of its consequences, that is, the effects it produces (that it is expected to produce ex ante, and it actually produced ex post): and if such is the case, then negative effects are a reason for employing knowledge about the effects of the (un-)triggering of certain concatenations of social mechanisms on the unfolding of certain courses of events to prevent the reform from happening.
The key thrust of this research agenda lies in identifying developmental patterns of public sector reform processes; that is, in detecting how certain configurations of social mechanisms may lead to relatively regular ways of unfolding of courses of events that correspond to the implementation of certain reforms of the public sector (like the introduction of a system for the management of performance, or the organizational redesign of some part of the public sector, for example, by decentralizing service delivery, and so on).
There remains a big question in order to be able to frame the scope and remit of a research agenda on the analysis of social mechanisms for the study of public sector reform dynamics: how do explanations based on the analysis of concatenations of social mechanisms relate to explanations based on 'contextual factors' influencing the dynamics of public sector reforms (e.g., like the ones outlined by Pollitt and Bouckaert in their highly cited work analysing public management reforms in different politico-administrative contexts; Bouckaert, 2000, 2017)? It is to this question we now turn.

REVISITING THE STUDY OF PUBLIC SECTOR REFORMS BY CONTRASTING SOCIAL MECHANISMS-BASED EXPLANATIONS AND INSTITUTIONALIST ACCOUNTS
This section tackles the key question of gauging the contribution brought about by analyses of concatenations of social mechanisms for improving our understanding of the dynamics of public sector reforms, by contrasting it with explanations looking for 'influential factors', like the ones outlined by Bouckaert (2011, 2017). We should notice that the latter accounts -which in a very simplistic way can be labelled as 'institutionalist', as the authors place great emphasis on the influence of institutions on such processes, but Edoardo Ongaro -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:08:40AM via free access actually resort to a much wider array of theoretical sources -are mainstream in studies of public sector reforms. Let us first see an example of how this perspective may be contrasted with the approach of social mechanisms, by reviewing in a paired way two papers dealing broadly with the same theme, and at least partly studying it in the same polity and jurisdiction (Italy, Spain and the UK). Fedele and Ongaro (2008) tackle the same issue, the dynamics of devolution processes in two legalistic countries (Italy and Spain, using the UK as a comparator from a non-legalistic tradition), as Ongaro (2006, focused on Italy only, and notably on one specific episode of devolution reform in the Northern Italian region of Lombardy), with a different theoretical perspective. Ongaro (2006) relies on the analysis of social mechanisms to identify developmental patterns that may lead to relocating tasks and staff from upper to lower tiers of government in the presence of unfavourable conditions that would otherwise tend to hinder and ultimately thwart the implementation of devolution processes. The selected case is an instance of successful implementation of devolution (in the Northern region of Lombardy and in the specific policy sector of agriculture) that contrasts with what was happening nationwide in Italy, where devolution of tasks and relocation of staff had been designed -the devolution law having been enacted -but not implemented. Specific concatenations of social mechanisms are evoked to explain the 'puzzling' outcome. Fedele and Ongaro (2008) set out to explain why devolution of competences to the devolved regions in the UK (Scotland, Wales and -with its distinctive arrangements -Northern Ireland) was comparatively more intense (factual and rapid in its execution) than in Spain or Italy, and why devolution in Spain was comparatively more intense than in Italy. To address the question, they identify a range of factors that may be influential on the dynamics of the implementation of devolution processes. Such factors encompass: horizontal coordination of central government; the nature of executive government; the social status and prestige associated with working at different levels of government; the geographical provenance of bureaucrats; the overall orientation of certain public management systems. In short, Fedele and Ongaro develop an institutional analysis (mostly historical-institutionalist, but with an eclectic flavour) centred on the identification of factors that may enable or hinder the implementation of reforms designed to relocate powers, competences, and staff from a 'central and upper' level of government to 'lower-level' governments by definition larger in number (as aptly noticed by Pollitt, 2005, decentralization -and devolution as a special case of decentralization -is defined by the spreading out of power from a smaller to a larger number of actors).
What are the differences in the findings of the two studies? Basically, they are complementary and pitched at different levels. Fedele and Ongaro (2008) point to a range of conditions broadly acting either in the direction of facili-Edoardo Ongaro -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:08:40AM via free access tating or hampering the implementation of devolution reforms -what we may call 'influential factors' -while Ongaro (2006) analyses what specific concatenations of mechanisms allowed overcoming the hurdles, enabling developmental patterns leading to implementing devolution under unfavourable conditions. So the two explanations are pitched at different levels of analysis, and broadly complementary with each other (it may be noticed the two papers were prepared at about the same time, and only due to the journals' internal editorial process dynamics carry a different year of publication). Let us now turn to what is probably the most known model for the study of public management reforms, centred on the perspective of the analysis of 'influencing factors'. Bouckaert (2011, 2017) point to five main factors characterizing a(ny) given politico-administrative regime that are singled out as especially significant in affecting both the contents and the process of public management reform. Before delineating such factors, it should immediately be noticed that they also introduce a broader, general model for mapping the dynamics of public management reform, intended as a first approximation model providing a conceptual map and a heuristic device to depict the broad forces at work, both driving and constraining change (Pollitt and Bouckaert, 2011, Chapter 2). They then turn to illustrate more specifically the five factors they have singled out as especially significant in affecting public management reform; they are (Pollitt and Bouckaert, 2011, Chapter 3): the basic structure of the state (vertical as well as horizontal dispersion of power); the nature of the conventions of governing (whether majoritarian or consensual); the manifold relationships between elected and tenured officials (thereby including the form the public service bargain takes; see Hood and Lodge, 2006); the administrative and organizational culture; and the sources of policy advice in matters of public management reform.
The range of influencing factors can be enlarged, notably when the cluster of investigated countries/polities is more focused, to make the analysis more accurate: in a study of administrative trajectories of five countries in the 'Napoleonic' administrative tradition (Peters, 2008;Ongaro, 2010;Painter and Peters, 2010) -namely France, Greece, Italy, Portugal and Spain -Ongaro (2009) introduced such factors as: the relationship between centre and periphery in the political party system; clientelism and politicization at the bottom; the geographical provenance of civil servants; the conception of the civil service and the role of public sector unions; the configuration of the system of corps and grands corps; the rise of legalism to the status of a cultural paradigm (Capano, 2003); and the engagement of civil society in politics and public policy as additional influential factors on the dynamics of reform processes.
The influence of 'contextual factors', broadly intended as 'independent variables', is modelled in terms of influencing the following 'dependent variables': how radical the reform can potentially be, in terms of difference and Edoardo Ongaro -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:08:40AM via free access distance from the current state of affairs; how rapid the pace of the reform can be; how government-wide (covering the broader functioning of government) vs sectoral the reform process can be (e.g., performance-related pay across the government or just in some areas); how rapid the reform process can be. Alongside holding sway on process dynamics, contextual features also exert an influence on the selection of the substantive contents of reforms (the second subset of 'dependent variables'): factors like the ones outlined above affect the 'fit' between certain public sector reform doctrines (intended as a generic form of knowledge about 'how the public sector should be organized') and certain local contexts as the recipients of the reform 'impulse' (see also Pollitt, Birchall and Putman, 1998;Pollitt, 2005Pollitt, , 2013. How can this analytical framework, which has become quite 'mainstream' since the appearance of the Pollitt and Bouckaert much-cited book in the year 2000, 3 be combined with the approach of the analysis of social mechanisms? The article by Ongaro (2013) is an explicit attempt to combine the two approaches and integrate an analysis in terms of 'influencing factors', explicitly rooted in the perspective of historical institutionalism, with an analysis of the social mechanisms at work in given reform episodes, notably to provide an explanation of the administrative reform trajectory of the European Commission (see also Ongaro, 2012 and. The paper is built mostly around the Pollitt and Bouckaert framework, then integrating and supplementing such explanation with a second and finer-grained layer of analysis rooted in the investigation of social mechanisms at work, notably during the reform episode also known as the 'Kinnock reforms' over the early 2000s (named after the European Commission Vice-President, Neil Kinnock, who was tasked with administrative reform).
The broad trajectory of reform of the administration of the European Commission is one that is characterized by more than 40 years of striking continuity of the basic organizational-administrative model followed by a critical juncture of radical and relatively rapid reform in the early 2000s, and then by some incremental reforms, mostly as reactions/adaptations to the 'Big Bang reform' and characterized by successive adjustments. The dynamics of this reform are explained by a model of analysis that consists of an adaptation with qualifications of the Pollitt and Bouckaert model, to then explain the finer-grained dynamics of the early 2000s episode of reform (in itself conceptualized as a critical juncture: a notion borrowed from historical institutionalism) by resorting to an analysis of the concatenation of social mechanisms at work to explain why New Public Management-inspired doctrines found their way into a continental European bureaucracy like the European Commission (whose details are summed up in Box 9.1).

BOX 9.1 A FINER-GRAINED ANALYSIS OF THE EARLY 2000S' ADMINISTRATIVE REFORM EPISODE AT THE EUROPEAN COMMISSION
Specifically pertinent to the issue of explaining the reform trajectory is the property of incorporating a significant cohort of staff in key posts for the reform policy process to occur coming originally from a country, the United Kingdom, which had systematically experienced New Public Management (NPM)-inspired reforms over the previous two decades. This feature may have enabled, or at least facilitated, the triggering of mechanisms of displacement via invasion of 'foreign' institutions and practices that supplant indigenous ones. As argued by Streeck and Thelen, such processes not only require that new institutional rules are enacted, which occurred with the formal approval of the Kinnock reform package (an outcome of a public management policy cycle reaching the decision phase - Barzelay, 2001Barzelay, , 2003 -and determining new rules in the areas of public management, like financial regulation and expenditure planning, or personnel regulation): actual change beyond policy formulation and throughout the implementation phase requires the active cultivation by the local actors -acting as enterprising actors -of the 'foreign' institutions and practices (Streeck and Thelen, 2005, p. 21). Such enterprising actors have been made available by the incorporation of staff with a different socialization and previous identity, definitely more receptive towards NPM doctrines. Moreover, and complementarily, beyond socialization and previous identity of staff, it is the direct access to administrative systems and practices at the national level to be significant in endowing the Commission with access to competencies and cognitive resources for making sense of the new managerial rules and practices. The institutional location of the Commission as the 'executive government' situated at the upper level of governance of a multi-level order is a feature of the politico-administrative context in which the Commission operates. Such an institutional location provides the channels for making sense of and setting to work novel, foreign institutions and practices: it is sufficient that such practices have found their way into one or more of the EU member states. This is a major dissimilarity between the politico-administrative context of the Commission and that of the two countries whose administrative model originally provided the pattern for its administration (continental Franco-German model of bureaucracy). Officials in central government at the national level do not have the kind of direct access to multiple administrative systems that its unique supranational position in the executive sphere provides the Commission with.
Edoardo Ongaro -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:08:40AM via free access By way of complementing the analysis, it may be hypothesized that also the mechanism of conversion -'the redirection of existing institutions to new goals, functions or purposes [that] can come about through changes in power relations, such that actors who were not involved in the original design of the institution and whose participation in it may not have been reckoned with, take it over and turn it to new ends' (ibid., p. 26) -has been at work. One component of the Kinnock reform was that public competition has been reasserted as the general rule of recruitment by the Commission, 4 and such recruitment process through public competition came to be executed almost exclusively through a novel organization, the European Personnel Selection Office (EPSO), which conducted recruitment in such a way as to strongly privilege staff with skills more attuned to the new managerial practices -like team working, exercise of leadership, and the like -rather than the traditional 'technical' skills, the expertise in technical areas, or in European history and law, which used to be the main areas of expertise tested by the previous selection procedures. This redirection of the recruitment procedures may have played a role in sustaining the institutionalization of the new practices and the consolidation of change. Source: From Ongaro (2013, pp. 356-7).

CONCLUSION: THE USAGE OF SOCIAL MECHANISMS FOR UNDERSTANDING AND MANAGING PUBLIC SECTOR REFORMS
We can now go back to the main questions: are social mechanisms 'adequate' building blocks for a form of theoretical understanding of public sector reform processes, and how can they be employed for improving the 'successful' management of public sector reforms? To address these questions, we must return to the issue discussed in the previous section of how the analysis of social mechanisms can be combined with the analysis of politico-administrative contextual features interpreted as 'influencing factors'. What is here suggested is that explanations based on the analysis of structural, cultural or functional features of the politico-administrative system -interpreted as 'influencing factors', that is, factors either facilitating or hindering the extent to which reforms are radical, rapid, wide in terms of the process of change, and what substantive contents they tend to favour rather than discard -provide explanations of the dynamics of public sector reform processes both at a higher level of aggregation and of a different kind than explanations resorting to concatenations of social mechanisms. It is not, or not only, that the former tend to be more 'broad brush' explanations, whilst the latter tend to be finer grained -it is that the former tend to be expressed in the terms of enablers or hindrances, Edoardo Ongaro -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:08:40AM via free access conceptualized as either facilitating or dimming the capacity of the system to get a reform pushed through, while the latter tend to be at the level of specific reform episodes (like the 'Kinnock reform' of the European Commission administration in the early 2000s) and to identify specific developmental patterns that, by either leveraging on the enablers or counteracting the hindrances, may be able under specific configurations of multiple conjunctural causes to put into effect specific reform provisions. The two perspectives are broadly complementary, and often the latter may tend to complement the former by providing insights into how certain courses of reform found their way 'notwithstanding' generally adverse contextual conditions. In a certain sense, the former perspective sets what has nowadays become the common wisdom, that is, what reform dynamics can be expected in certain politico-administrative regimes, and the latter explains puzzling exceptions (in the social scientific sense), that is, it explains why reform dynamics follow alternative courses than what would be expected by analysing the influencing factors (e.g., why devolution in the field of agriculture in the Italian region of Lombardy throughout 1998-2002 succeeded while the bulk of the devolution reform more widely across the country did not).
We may also add that social mechanisms analysis brings the vivid fabric of social processes to light. It is more empathic with real-life events. It is often coupled with a conception of time as duration rather than a spatialized notion of time (i.e., akin to the so-called process philosophy and the related notion of time propounded by Alfred North Whitehead, or -albeit from different premises -by the philosopher Henri Bergson: for a discussion of the significance of different notions of time for public administration studies, see Ongaro, 2017, Chapter 4, pp. 99-108 and 140-44 in particular). It is also in this sense that the approach of the analytical narratives, which underpins and constitutes the approach of the analysis of social mechanisms, complements institutional analysis of the influencing factors (Ongaro, 2016). Analyses of factors take the broad picture, the bird's eye view, but they may lose contact with the unfolding of events, which is what analytical narratives bring in.
There is, however, an important, and final, qualification: uncovering concatenations of social mechanisms at work is a kind of knowledge that seems to be applicable mostly if not exclusively to focused, specific, circumstantiated cases or types of phenomena. The scope of this approach is both focused and relatively narrow. Its generalizability resembles more that of casuistry, 5 as a method of treating the varied circumstances in which decisions are taken by human beings, than that of generalizable propositions, let alone encompassing law-like generalizations. In sum, the social mechanisms approach may be used on its own, but it is better used and more meaningful when it represents a zooming in to focus on specific dynamics set within a larger explanatory frame. Edoardo Ongaro -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:08:40AM via free access NOTES 1. In a number of works, Barzelay (Barzelay, 2001, p. 14; Barzelay and Gallego, 2006, 2010a, 2010b introduced the definition of public management policy as: 'institutional rules and organizational routines in the areas of: expenditure planning and financial management, audit and evaluation, organization and methods, labor relations, and procurement that guide, constrain and motivate the public sector'. Such a definition -albeit slightly problematic where it attributes 'agency' to these rules and routines alleged to act towards 'motivating', impersonally, the 'public sector', and might rather read 'public sector officials' or more loosely 'people working in public services' -explicitly conceptualizes the reform of public management as a policy process, thus 'enabling' the application of the conceptual tools of public policy to the analysis of a specific domain: the public management policy domain. Public management is not necessarily a policy domain in every period in any country, but it may become such when changing the rules and routines of the whole government -at the government-wide level -in the areas above specified becomes an issue high on the governmental agenda and in a sustained way. 2. Etymologically, prophecy means 'to speak in the name of somebody else'. It is here used in the common (albeit improper) usage to mean 'forecast (of a future scenario)'. 3. A simple count accessing Google Scholar indicates 8988 citations for the 2nd and 3rd edition of the book, excluding translations (accessed on 8 January 2018). 4. Traditional entry mechanisms included the so-called 'submarine approach', whereby seconded staff or staff initially hired with temporary contracts used to have a privileged access to restricted competitions, and 'parachuting', whereby top positions used to be filled with externals, through processes of political appointment (Stevens and Stevens, 2001). 5. Casuistry, historically developed by the Jesuit Fathers, is mostly concerned with decisions in the moral order under varied circumstances.