Change Management that Works: Making Impacts in Challenging Environments

Achieving better governance has been a central problem for development. When public services are not delivered as intended, reform action becomes necessary and that involves deliberate activities to change laws, structures, and processes to improve public sector performance and benefit public service users. The key challenge is that changes in the design of the institution or its procedures do not necessarily translate into immediate changes in the behavior of relevant actors. A central problem of public sector reform is ensuring that changes in laws and policies also prompts changes in the way that people work, so that service delivery improves. There is no one-size-fits-all approach ensuring that change happens the desirable way; however, experiences from the field suggest that a useful combination of political economy analysis with change management tools can help to maximize positive impacts. Different contexts will require different approaches to change management, and therefore political economy analysis can be used productively to design a targeted change management strategy that builds on existing strengths and opportunities. Greater integration of political economy analysis into change management assessments has been helpful in deepening understanding of attitudes to change within these particular contexts. This has allowed more effective leveraging of the opportunities for reform through the more systematic tailoring of change management strategies to different sets of issues emerging among particular groups of actors. Cambodia and Indonesia, the case studies presented in the paper, help to illustrate this.

Achieving better governance has been a central problem for development. When public services are not delivered as intended, reform action becomes necessary and that involves deliberate activities to change laws, structures, and processes to improve public sector performance and benefit public service users. The key challenge is that changes in the design of the institution or its procedures do not necessarily translate into immediate changes in the behavior of relevant actors. A central problem of public sector reform is ensuring that changes in laws and policies also prompts changes in the way that people work, so that service delivery improves. There is no one-size-fits-all approach ensuring that change happens the desirable way; however, experiences from the field suggest that a useful combination of political economy analysis with change management tools can help to maximize positive impacts. Different contexts will require different approaches to change management, and therefore political economy analysis can be used productively to design a targeted change management strategy that builds on existing strengths and opportunities. Greater integration of political economy analysis into change management assessments has been helpful in deepening understanding of attitudes to change within these particular contexts. This has allowed more effective leveraging of the opportunities for reform through the more systematic tailoring of change management strategies to different sets of issues emerging among particular groups of actors. Cambodia and Indonesia, the case studies presented in the paper, help to illustrate this.
Managing public finance is invariably a highly charged political issue and changing the technical possibilities for moving public money and accounting for its expenditure is likely to be politically contested. In environments where public finances have historically been mismanaged to the benefit of particular groups, there is likely to be significant concern over or resistance to more transparent systems. How does successful change happen in such a highly charged political issue? This paper attempts to answer this question by reviewing the experience of change management in two Public Financial Management (PFM) reform projects supported by the Bank. The case studies chosen for this study both involve the roll out of Financial Management Information Systems (FMIS) in Ministries of Finance in Asia, in Cambodia and Indonesia. FMIS involves computerization of PFM activities to improve the efficiency, transparency and accountability of public financial management. The use of integrated online systems for managing all transactions allows better sharing of information, reduces mistakes, and permits managers to run reports which can assist in analyzing and evaluating performance against a variety of measures. Advocates of FMIS suggest that it represents a transformational change with respect to Ministries of Finance, releasing staff from mundane manual form-filling and opening up new possibilities for critical analysis of financial management and public expenditure. Cambodia and Indonesia implemented very similar FMIS programs, and used similar change management strategies. The different political and institutional contexts of the two countries entailed that the program rolled out differently and with different effects. This allows us to demonstrate how change management needs to be informed by political economy analysis in order to develop appropriate strategies and tools that work in facilitating change.
This paper locates an evaluation of change management strategies supporting PFM reform within a political economy analysis of the institutional functioning of PFM agencies, and the way that this relates to political concerns about state-building, service delivery, and corruption. Cambodia and Indonesia both have historical state-building trajectories that were heavily and violently influenced by Cold War geopolitical dynamics, economic disaster, and international aid dependence. Both underwent partial transitions from authoritarianism to at least formal multi-party democracy during the 1990s, and in both countries significant political parties perceive a pressing imperative to improve service delivery as a means to win elections and to legitimize government. However, this is difficult because Cambodia and Indonesia have both historically experienced difficulties with institutionalization of integrated and accountable PFM systems. This has significantly undermined the integrity of public finances and the ability to fund improvements in public services.
The paper is organized into four sections. Part 1 reviews the literature on change management with a particular focus on how this literature addresses the specific difficulties inherent in PFM/public sector reform. It further lays out some different approaches to political economy analysis used in development operations, and proposes a model whereby political economy analysis can be used to inform more effective change management. Part 2 discusses Cambodia's experience of PFM reform, locating this in the context of Cambodia's historical state-building trajectory and contemporary political economy; the way in which the institutional structure and culture of the Cambodian state emerged from this historical context; and the change management techniques that were used to attempt to alleviate the impact of these concerns on the FMIS process, noting both successful strategies undertaken and opportunities that were missed, and the way in which earlier and more systematic use of change management tools, informed by political economy analysis, could have helped to promote smoother processes of reforms. Part 3 moves on to a similar investigation of Indonesia's experiences of its version of FMIS, known as SPAN. Finally, the paper concludes with a discussion of the lessons to be drawn from the comparison to assist in designing new analytical tools and strategic approaches for more successful change management.

Change Management in the Public Sector
Change management is defined as "the process of helping people understand the need for change, and to motivate them to take actions which result in sustained changes in behavior" (World Bank 2015). This is different from project management or systems design in that it focuses less upon the technical side of reform than on the behavior of people. A central tenet of change management is the idea that organizations must frequently change in response to pressures from the external environment, including availability of new technology, the emergence of new problems, or the threat of external competition. Change is considered to be part of a "continuous, open-ended process of adaptation to changing circumstances and conditions (By 2005)". Much of the change management literature has focused on the response of lower level workers to management initiatives, and the need to overcome resistance to change through spreading information to promote "buy-in" -the realization that change is beneficial.
This framework has been challenged as critics increasingly acknowledge the conflictual and political nature of organizations particularly in the public sector. In practice, any given change is unlikely to be equally beneficial to all employees, and therefore some employees will have perfectly rational reasons for resisting change and that resistance could increase as more information becomes available. Armenakis and Harris (2009) argue that a better approach to conceptualizing employee responses is through the concept of "readiness" and that readiness should be pro-actively promoted before resistance has an opportunity to emerge. According to Armenakis and Harris, recipients react differently to proposed changes depending on their prior positions or beliefs. Anticipating or exploring different employee responses allows for more subtle approaches to change management that address differentiated concerns about the process and content of change (Armenakis and Harris 2009, p. 129).
This kind of approach takes change management away from the underpinnings of rationalism and change as a public good that is evident in much of the early business-oriented literature. Higgins et al identify "organizational culture"-"the pattern of shared values, norms and practices that help distinguish one organization from another"-as a crucial determinant of how change is perceived and how it can be managed. According to Higgins et al (2006), "These values, norms, and practices define 'what is important around here,' and 'how we do things around here.' They provide direction, meaning, and energy for the organization's members." (Higgins et al 2006, p.401).
Transferring change management models to the public sector requires greater attention also to interactions between the administrative and political spheres. For public sector organizations, maintenance of legitimacy in the eyes of external constituencies is crucial to survival, and consequently success for public organizations means consciously pursuing change in a manner that links "technical" and "political" rationalities (Bryson 2007, p. 25). It may be added that in many contexts, the public sector forms one or many political constituencies in its own right. For example, the political power wielded by public sector workers -particularly those who work in highly legitimizing front-line services such as health, education and firefighting-through appeals to the public in support of claims against the government over pay and working conditions can be considerable.
Consequently, despite the fact that "good governance" and the public administration reform have been a central focus of statebuilding and development programs on the part of major international donor agencies for the past twenty years, the record of success of donor-sponsored public administration reform programs is poor and the process is widely recognized as intensely political (Heredia and Schneider 2003, Gisselquist and Resnick 2014, Hillman 2013.
This brief review of the literature suggests that in considering systemic reform programs aimed at transforming public sector organizations, there is much scope for elaborating on mainstream change management techniques.
Specifically, we suggest that change management models for public service organizations need to take on board the specific challenges of the relationship between the administrative and the political spheres, and the particularities of institutional culture in various public sector agencies. The argument of this paper is that political economy analysis can offer important insights supporting the development of change management strategies for governance and public sector reform in developing countries.

Application of Political Economy in Development Operations
There are different approaches to political economy analysis in the literature and this has left development practitioners confused or remain in doubt regarding its utility. Here we summarize four key approaches to political economy in the literature, and draw out the potential of political economy analysis for informing approaches for change management.
A "public choice" model of political economy analysis proceeds from the assumption that the key actors are individuals, who operate in a broadly level playing field, maintain complete utility preference sets and make rational choices to maximize utility in any given situation. The emphasis on utility maximization in public choice models tends to promote a conception of civil servants as self-interested, rather than as dedicated to serving the public interest, and to explain civil service failures as a result of a tendency on the part of civil servants to attempt to divert resources to their own ends at the expense of the common good (Mueller 2015). These approaches pay scant attention to the issue of political values or ideas (Hudson and Leftwich 2014). It is challenged by a significant empirical literature based on attitude surveys undertaken with cohorts of public servants in different countries, which produces fairly consistent evidence that public servants are more likely than other citizens to have altruistic attitudes (Houston 2000;Wright 2001;Leisink and Steijn 2009;Kim and Vandenabeele 2010). Furthermore, the pessimistic assumptions underlying public choice theory makes life extraordinarily difficult for proponents of progressive reform, as Grindle and Thomas note: "if policy makers and public managers are conceptualized as "rent seekers" motivated only by the desire to remain in power, little can be expected of them in terms of leadership or the management of change" (Grindle and Thomas 1993, p. 4-5). This would limit the possibility for any reasonable dialogue on specific policy and incremental changes, which actually take place. Such stripped-down models of the motivation of public servants and their leaders have a certain methodological elegance, but are less useful to practitioners engaged in the messy day-to-day processes of dealing with public servants working through difficult and contested processes of reform.
A second approach to political economy analysis draws from a different tradition, taking structural issues rooted in the relationship between politics and the economy much more seriously. This approach proceeds from the assumption that the appropriate focus of attention is not the individual, but key social groups. These groups are differently located in relation to the economy and consequently not only have different interests, but also different resources of power for promoting their interests politically. In these models, society is conceptualized as structurally unequal, with certain groups consistently dominating agendas and decision-making over time and across a range of issues areas.
Analysis of the relative power of different groups to realize their interests thus becomes a significant focus of attention for development practitioners seeking to promote reform programs, particularly those that seek to significantly redistribute resources among social groups, as in, for example, pro-poor programming or anticorruption programs (Hutchison et al 2014;Hudson and Leftwich 2014). Many development programs sought to unpack issues of "vested interests" and "political will" by exploring the links between institutions, power relations and social forces in pressing for or blocking change (Routley and Hulme 2012;Hudson and Leftwich 2014).
Structural issues are useful in understanding the change context and potential opportunities for change, but are less good as a basis for identifying useful interventions in support of change.
More recent approaches (the third approach) to political economy analysis have drawn from the legacy of this first generation, but have shifted focus to the interaction between structure and agency in different development contexts (see Hughes and Hutchison 2012). One model for this is problem-driven political economy analysis. As articulated by Verena Fritz et al (2014), the key to problem-driven political economy analysis is understanding the interactions between structural and institutional constraints and stakeholder interests and constellations in any particular reform contexts. A central tenet of problem-driven political economy analysis is that the activities of "stakeholders" are primarily determined by "incentives". A significant innovation of this approach is the incorporation of a specifically political analysis of the incentives faced by elites, and the centrality to problem-driven political economy analysis of the ways in which elites distribute economic resources in order to mobilize the political support of their most significant constituencies. Problem-driven political economy analysis thus illuminates the reasons that development projects that appear worthy from a technical or efficiency point of view may be neglected if they do not contribute specifically to over-riding political calculations relating to the need to win or maintain political support from key groups.
The fourth approach to political economy analysis, developed through a research program sponsored by AusAID from 2009-2013, takes as a starting point that development itself is a process of contestation and development outcomes are the product of this contestation (Hughes and Hutchison 2012;Hutchison et al 2014;Hout 2015). This view shares with problem-driven political economy analysis the concern with the different capacities of different actors to defend their interests by wielding power. However, it is distinguishable from the problem-driven political economy analysis in three ways. Firstly, this model adopts the concept of "interests" rather than "incentives" and views these as structurally determined, relating directly to the nature of economic production. Secondly, this model focuses less on stakeholders than on broader social groupings or constituencies. Thus development outcomes are regarded as the product of contestation by different constituencies with different structurally determined interests that relate primarily to their social and economic situation (Hughes and Hutchison 2012;Hutchison et al 2014;Hout 2015). Finally, this model also pays attention to the ways in which particular groups articulate their demands.
Drawing upon Gramscian theory, this model analyzes the ways in which different constituencies frame political claims in a manner that can assist them in forming alliances with powerful actors (Hughes and Hutchison 2012, Hutchison et al 2014, Hout 2015, Hughes 2015.

A Political Economy Informed Approach to Change Management
It is the contention of this study that change management techniques need to be informed by a flexible and real-time approach to political economy in order to effectively diagnose and deal with different challenges to change processes as they arise in a dynamic context. In considering systemic reform programs aimed at transforming public sector organizations, there is much scope for elaborating on mainstream change management techniques. Specifically, we suggest that change management models for public service organizations need to take on board the specific challenges of the relationship between the administrative and the political sphere, and the particularities of institutional culture in various public sector agencies, reflecting entrenched distributions of power between different arms of the state. Building these areas of concern into more mainstream models of organizational change, this study proposes consideration of the structural and contextual condition of each country undergoing reform to help inform the design of a change management tool ( Figure 1).
Behavioral change on the part of actors working in particular institutional settings is the intended outcome of any change management effort. A political economy informed approach to change management, we suggest, is cognizant of constraints imposed by a broader country's political economy, which shapes for example, public opinion; the political strategies of governments and leaders; the institutional culture of the bureaucracy, and the distribution of power between the public, leaders and bureaucrats, thereby ultimately determining the possibilities for reform.
However, it also suggests that "smart design" can be built into reform program mechanisms for leveraging opportunities identified by political economy analysis. Any particular reform project has instruments at its disposal that can maximize the scope for change within existing constraints; and can promote change in particular directions.
In the latter case, where powerholders may have an appetite for change but are uncertain which direction to go in, change management tools can promote the likelihood of particular solutions being adopted. Strategies that promote the conceptualization of problems in particular ways that predispose towards certain types of solution; familiarization of key actors with the implications of those solutions; training with respect to new technologies and confidence-building with respect to new relationships; and experimentation with unfamiliar modes of practice can all help to make a particular solution appear more attractive to decision makers. At the same time, efforts to promote the confidence, resources and assertiveness of particular groups of actors can enhance political pressure in favor of particular trajectories of change. Thus even though change management strategies may be limited in forcing change through where there is no appetite for change at all, they may nevertheless be decisive in contexts where a government is looking for solutions to a particular problem.
Where a systemic reform program is envisaged, change management strategies must address a broad set of issues emerging from the institutional culture of the public service unit. In line with the political economy analysis approach outlined, in this study institutional culture is defined as a set of working practices, relationships and values that characterize a public service organization. The particular form of institutional culture in any state institution is partly endogenous, maintained by internal working practices and relationships and the values upon which they are based and which they reinforce. It is also strongly affected by external relationships with political leaders, with international development partners in an emerging country context, and with the broader public constituted as national citizens, as public service clients, and as civil society. Institutional culture is not static, but it is closely connected to dynamics of power and the role of the state generally, and of the particular public service unit in question, in managing the relationship between ruling elites and society more broadly.
External pressure emanating from a changing political situation, both directly on public servants through their engagement with the public through their work and more widely, and via the articulation of new political strategies by political leaders, will affect institutional culture, but the effect of this will be contingent, depending upon whether such pressures can be accommodated by, or require the relinquishment of, key norms of behavior. Public awareness of and concern about reform issues in turn partly determines the articulation of the specific reform program with broader political strategies, thus closing the circle.
Understanding change management strategies as an element in the broader political and administrative context allows both a wider conceptualization of how change might be promoted, and a more sophisticated awareness of the different sources of potential resistance or readiness for change. In this study, consequently, we consider the evolving political economy of the reform context as the key set of structures within which both the political strategies of elites and the institutional culture of the state apparatus emerge; however, this does not imply that these two things are necessarily congruent. Political leaders and bureaucrats have different relationships to the political economy contexts and therefore respond differently to structural change, and this gives rise to the potential for change that is mandated by the government to be resisted by public servants.
Combining political economy analysis with a portfolio of change management techniques can tailor those techniques to an integrated analysis of how the political economy context informs the interaction between a political strategy for change and the institutional culture of the public sector organizations involved in that change. It can unpack the implications for particular reforms inherent in the political economy context and the way that this is reflected in political strategies and in the institutional culture of public sector organizations. Consequently, political economy analysis can offer change managers important information about whether there is a significant window of opportunity for reform -do sufficiently powerful sections of either the political elite, or the public service, or both have enough interest in reform to make significant change possible? Second, it can disaggregate the nature of different constituencies within a particular public sector organization, assisting in designing change management tools that will mobilize and promote the influence of the most likely advocates for change and assisting in understanding the motivation for ambivalence or outright resistance. Change managers can then adopt a bespoke package of techniques that will make forces for change generally, and for specific types of change in particular, more effective.
In order to demonstrate this process, we have disaggregated change management tools into five areas, each of which is acknowledged in the change management literature as essential for promoting effective change. These are leadership; project governance; engagement and communication of stakeholders; workforce enablement; and organizational realignment. Each of these categories describe a particular aspect of the envisioning, directing and embedding of change. Although traditional CM approaches focus on individual "readiness" for change, measured in terms of technical competence, each of these dimensions draws our attention to a particular aspect of the relationship between power and change in specific political economy contexts, and in relation to the political strategies and institutional cultures emerging from these. Consequently, political economy analysis offers insights into the design of change management tools in each of these categories.

Leadership
Change management approaches have long regarded leadership as essential to successful change. Leaders articulate a change vision, and even in relatively small change processes, their commitment to that vision is important in convincing lower level actors that change is not only beneficial but inevitable, thus increasing the likelihood of compliance. In more fundamental or transformational change processes, where institutional structures, relationships and values are likely to be affected, leadership is even more important. "Leadership of successful change requires vision, strategy, the development of a culture of sustainable shared values that support the vision and strategy for change and empowering, motivating and inspiring those who are involved and affected" Gill (2003, p. 307) .
However, formal statements of support for a change program are inadequate of themselves. Leaders must make the change intelligible in terms of normative frames of reference shared between themselves and their subordinatesthe "how we do things around here" of a particular public service organization (Higgs et al 2006) -and this may include shifting such frames of reference in important ways. This pertains to the concept of "political will", which can be understood as encompassing not just the incentives motivating a particular reform at a particular time but also as a way of explaining how the normative collective project of a public service organization is being adjusted to accommodate a reform. Political economy analysis allows the unpacking of both these layers of political will. It suggests the need to employ a conception of power that goes beyond incentivization although our case studies below suggest that material incentives like salary supplements and hiring and firing power are likely to be significant also.

Project governance
Project governance that reflects the local political economy context can help to ease the implementation of the change program. Given the account of leadership offered above, it is clear that decision-making in a change program needs to be located at a level where power is sufficiently concentrated to mandate action. The nature of political economy in fragile contexts often reflects highly hierarchical/centralized decision making and a low level of coordination among different agencies. If decision-making is located at the wrong level, difficult challenges of coordination or resourcing can arise, prompting major disruption during implementation. Equally, it may be possible that front-line service providers who interact daily with the public experience more direct pressure for change than their superiors.
In this instance, project governance that is designed specifically to give lower levels of employees a voice and to facilitate means by which they can exercise power to exert pressure on higher levels of government might be useful.

Engagement and communication with stakeholders
"Engagement and communication with stakeholders" encompasses two aspects: the first is ensuring that employees know about a change, and the second is that they regard it as something they will accept or, preferably, promote.
These two aspects have sometimes been conflated in traditional CM approaches, where communication strategies have been formulated on the assumption that once employees properly understand a change, they will inevitably regard it as beneficial to themselves.
Effective engagement and communication is more usefully seen as a means of devising and circulating a narrative which constructs the desired change as a mutually beneficial common goal, given the particular material and normative context of the public service agency. As in the case of leadership, above, this requires a deeper analysis of the normative frameworks informing institutional culture than is implied by the notion of incentives. It also requires the exercise of power, since it requires encouraging a range of actors at different levels to reorient not only their actions but also their beliefs about the relationship of their work to a broader set of institutional or social goals.
Furthermore, effective change management may require the reorganization of channels of communication within an institution over the long term. This again requires exercise of power, particularly in contexts where personalist relationships are important. Change managers need to implement strategies that can "break the ice" between different groups that would normally interact rarely; it may also imply developing close practices of day to day communication between groups that regard one another as competitors or even enemies. Furthermore, promoting such channels of communication between groups that might share a tactical interest in change, even if for different reasons, can be useful in leveraging their combined influence. Political economy analysis of the functioning of an institution can assist in determining how these kinds of relationships work, enabling design of change management tools that are realistic about both the potential problems and opportunities inherent in a radical reordering of institutional relationships, about the amount of power that will be needed to implement change successfully, and about the types of actors who might be able to exert the necessary level of pressure, whether from above, below, or from outside the organization altogether.

Organizational alignment
Organizational alignment is another important element in change management, and it refers to the process by which organizational practices are reoriented to reflect the new way of working. For example, shifting from manual to automated processes requires a change in work practices, organizational structures and job descriptions, which need to be appropriately realigned. Transformational change in the structure, function, and culture of the ministry can spark major resistance if it has distributive consequences for specific stakeholders. In our case, FMIS implementation is not simply a technology upgrade. It also involves the transformation of working practices, with enormous amounts of basic clerical work rendered redundant, allowing opportunities for a more analytical or forensic approach to accounting and budgeting. However, this implies a change in working practices, in the types of skills and competencies required, in the kinds of interactions and activities each worker will engage in on an average day, and in hierarchies of control.
Because of the level of resistance to wholesale transformations that fundamentally redistribute resources and power, often these can only be achieved in conditions of crisis, when leaders are under enormous pressure to take radical decisions quickly. However, in the absence of crisis, incremental change may be possible. Once again, political economy analysis can provide information on the distribution of pressures for change and for resistance to change, as well as a range of outcomes in between such as manipulation or negotiation of change. This can assist in the development of an organizational alignment strategy, by analyzing the political sensitivity of different aspects of change. For example, it might be expected that changes that involve mass lay-offs of staff will be highly politically sensitive, whereas changes that reduce the need to require staff to work unpaid overtime will be less so.

Workforce enablement
Workforce enablement is used in the change management literature largely to mean training in effective use of and engagement with new systems. However, in line with the analysis here, the term can also be viewed more broadly as representing a level of confidence and familiarity with new processes, systematic learning, and an acceptance of new ways of doing things as appropriate and valid means of achieving common goals. Adequate training, particularly for lower level staff who will have to use the system, combined with provision of monetary or status incentives linked to the acquisition of new skills are important in the narrower sense of the term, while gaining familiarity with new ways of working requires further attention to realigning institutional culture in a manner that can create new relationships and modes of interaction which facilitate new attitudes on the part of staff. This paper uses the combination of political economy analysis with change management techniques in these five key areas to explain the different outcomes of two FMIS projects that were recently implemented in Indonesia and Cambodia. The two countries shared similar governance problems in the early 2000s, but quite different political economy contexts, giving rise to distinct political strategies and institutional cultures. The introduction of similar FMIS programs unfolded very differently in the two cases, with clear implications for the way that political economy and change management can be folded together to provide useful practical guidance to reformers in a fast-changing situation on the ground.

The Political Economy Context of State-building and PFM Reform in Cambodia
Cambodia is a post-conflict state whose current institutions emerged quite recently in a context of high levels of violence and low levels of human development. The current political settlement and state apparatus in Cambodia have their antecedents in the aftermath of the catastrophic levels of death and destruction unleashed upon Cambodia by the Cold War and the Democratic Kampuchea or "Khmer Rouge" regime. Gottesman (2004) explained that the Vietnamese invasion of Cambodia in 1979 ousted the Khmer Rouge, and began the task of rebuilding the state and state services in the context of international isolation and continued internationally backed insurgency throughout the 1980s. The end of the Cold War and the Vietnamese withdrawal from Cambodia in 1989 brought a new era of free market and democratic reforms. A peace deal in 1991 gave rise to a peacekeeping mission which presided over elections in 1993 and the remnants of the insurgency were finally defeated in 1998 (Hughes 2003).
Cambodia's particular history of war and state-building has important implications for the political economy and institutional culture of its state bureaucracy and imposes particular kinds of constraints on the prospects for governance reform. The close link between warfare and state-building in the 1980s entailed that contemporary state institutions emerged in a climate of dislocation, distrust and intense fear. For state managers at the time, shoring up the regime required rapidly developing the ability to exercise power across the territory and get things done in a context of extreme material shortages and an absence of formal institutions. Consequently, the state that emerged was heavily dependent upon informal and personal relations of reciprocity, comradeship, and patronage (Gottesman 2004;Hughes 2003). Public financial management in this era was rudimentary.
During the 1990s, as free market reforms took place, Cambodia regained recognition from Western countries and aid and investment began to flow in, regularizing of public financial management began to occur, but it remained very weak throughout the 1990s as a result of a range of factors: weak or absent processes for collecting revenues such as tax; unmanaged natural resource revenues flowing in an unregulated way to actors at all levels from central to local; the existence of unknown numbers of ghost soldiers and civil servants on the public payroll; reliance on discretionary cash payments and advances to cope with urgent problems; and excessive donor fragmentation and projectization that overloaded Cambodia's ability to manage its heavy aid dependence (Hughes 2003(Hughes , 2009RGC 2007;Ear 2012;Cock 2016).
These problems were compounded by three inter-related factors: low public service wages; politicization of the civil service in the context of post-war democratization; and reliance on personalist networks of patron-client relations to get things done in the public service. The postwar Cambodian state was built by bolting together networks of individuals from different political backgrounds into central and provincial structures. Virtually all administrative decisions were regarded as political, and were made only by the very top levels, and therefore coordinating mechanisms needed only to be vertical and top-down. The impact of this on what change managers call state "business practices" includes: prioritization of loyalty over competence in recruiting, evaluating and rewarding staff; entrenchment of close-knit vertical relations of hierarchy and upward accountability combined with poor coordination across units; the emergence of complex and often informal procedures on values of trust and reciprocity, designed to foster the inclusion of all relevant "insiders" while remaining opaque to "outsiders"; and retention of discretionary power over key issues in the hands of a few individuals combined with high tolerance of small-scale infractions of rules and regulations within clearly understood limits by subordinates (Hughes 2003, Un 2005. Examining public expenditure management (PEM) at subnational level in Cambodia in 2008, Pak and Craig argued that business practices for PFM were highly centralized and powerfully shaped by "patronage networks of personal relationships and loyalty, institutionalized rent seeking and political agendas." They argued that although technically informal, "patronage around provincial PEM is dense and institutionalized, especially within the mainstream system… Concerns for compliance have multiplied opportunities for informal deductions, and personalized yet regularized relationships have distorted processes (Pak and Craig 2008, 7)." A further PFM assessment (based on Public Expenditure and Financial Accountability -PEFA) in 2015 concluded that although the Cambodian government had an orderly process for making budgets and releasing money to different state agencies, there were weaknesses in monitoring expenditure, and frequently large deviations between the allocations of resources in the budget approved by the legislature and the actual allocation of resources between ministries.
A further particular problem identified was payment arrears. The Cambodian government historically set a limit of 90 days from the date an invoice was received by the General Department of the National Treasury (GDNT) for paying suppliers and contractors. This was reduced to 60 days in 2015. Despite the improvement, the new limit still exceeds international norms of 30 days. Furthermore, invoices have already been through a long process of approvals before reaching GDNT, with the result that it commonly takes six months or more for suppliers to be paid for services, creating opportunities for gatekeepers to attract rents by offering fast-track services. The existing KIT system allows the government to generate reports on arrears, but there is no system for routinely tracking this.
The PEFA review suggested problems with both the complexity of vertical hierarchy within the ministry, and also the limited horizontal liaison between different departments. With regard to control mechanisms for expenditure, for example, the review concluded: The government's pre-existing processes were highly centralized and organized around tasks/functions. An analysis of existing practices within the MEF conducted for the purpose of facilitating FMIS was critical of the very time consuming, labor intensive procedures by which approvals were gained. The Change Impact Assessment report described the system as incorporating "many work steps that do not add value… and consume unnecessary time and effort. On average, between 70 and 80% of work steps are checks, initialing, signing, registering, copying and approving… Most of these steps are unnecessary even in a manual environment (Change Impact Assessment 2015, p. 3)." A further issue highlighted in the report was the very high ratio of managers to staff within the MEF, to the extent that some departments in the ministry exist which have several managers, at the level of deputy director, but no staff. The number of transactions processed per staff member in many departments is low (Change Impact Assessment 2015, p26).
The primarily vertical orientation procedures have the effect of limiting horizontal linkages within government, and this is reflected in the formal processes in the central Ministry of Economy and Finance. The 2015 PEFA report commented: "Unconnected manual and semi-manual processes and software systems are still largely used in the budget execution, reporting and accounting processes. Even within MEF itself, each department tends to use its own system with no electronic connection to other systems (PEFA Review 2015)." Lack of horizontal interaction entails that requests and permissions have to travel up through extended vertical hierarchies before a sufficient level of seniority is reached to be able to cross the gaps between different departments. Organizing things this way ensures close central control and strong upward accountabilities between individuals who work closely together, but limited horizontal accountability or delegation of responsibility to agencies at lower levels. This is costly, time-consuming and inefficient, and creates opportunities for mistakes, waste, and rent-seeking.

The Beginnings of Public Financial Management Reform
The reform plan for improving public financial management that emerged took an innovative platform-based approach, comprising four phases or "platforms" each with its own set of reforms and objectives. The first platform was focused on the immediate priority of enhancing budget credibility in order to ensure that budget holders had the resources they needed to deliver expected services and promised development projects. Platform 2 focused on increasing the accountability of managers for expenditure by improving accounting and audit practices and introducing the first stage of an online Financial Management Information System. Platform 3 was aimed at promoting the link between policies and budgets by reforming the budget cycle, improving budget analysis and promoting fiscal decentralization. Platform 4 aimed to strengthen accountability for result. Thus the platform approach represented a stepped approach towards increasingly challenging goals. Platforms 1 and 3 focused on facilitating the links between policy and expenditure, by improving the ability to move money to where it was needed at the right time. Platforms 2 and 4 focused on promoting accountability for money spent, thus reducing corruption and waste and enabling evaluations of government action that could help the government distinguish between effective and ineffective policies by monitoring results.
The significance of the platform approach is that it allowed incremental change towards a series of short-term goals.
Platforms 1 and 3 link closely to the government's effectiveness agenda and might be expected to be politically noncontroversial as a result. Platforms 2 and 4 are more complex since they involve closer scrutiny of the competence and integrity of individual civil servants in an institutional culture characterized by strong and personal links of reciprocity, loyalty and mutual support. According to Taliercio, the government regarded the first platform, promoting the credibility of the budget and improving budget execution, as crucial to its strategy of consolidating support among the electorate (Taliercio 2009). However, Taliercio also points out that there were significant challenges and that certain parts of the PFM program were not able to be achieved due to what he calls "vested interests" (Taliercio 2009).

The Challenge of FMIS
As noted earlier, the significance of FMIS is less in the technological skill required to install and operate the software than in its implications for the procedures by which public finances are managed and by which decisions about expenditure are taken. In Cambodia, implementation of the technology is proceeding broadly according to plan and political leaders express support for its goals of promoting transparency, tackling corruption and increasing efficiency. However, delivering these goals also means fundamentally reorganizing relationships within the MEF, in a manner that directly challenges some of the principles on which Cambodia's state institutions were built and 14 by which they continue to operate. This affects not only their internal functioning and institutional culture, but also the way they relate to the broader political environment, to citizens, and to political leaders.
A Change Impact Assessment of business processes in the MEF conducted in preparation for FMIS noted that many of the unnecessary steps in existing manual approval procedures were redundant even without the introduction of the online system (Change Impact Assessment 2015, p.3). The report further explained: "there is little to be gained in terms of efficiency and effectiveness by implementing FMIS and retaining the current cumbersome, inefficient manual processes, excessive approvals, duplicated practices and unnecessary forms (Change Impact Assessment 2015, p. 4)." The challenge for FMIS is to act as a catalyst for reforms that are not only necessary to realize the potential of the new computerized system but are in any case long overdue. The tackling of systems that perpetuate waste, inefficiency and lack of transparency requires a reorientation of institutional practices which relate not only to a historical legacy of warfare and insurgency but also to a contemporary political settlement. Changing this will require prioritization of the reform within new political strategies of legitimation pursued by the top leaders of the country.
The FMIS is configured with a simple two-stage "Add" and "Confirm" process, which is transparent to all departments while existing lengthy processed operate in parallel with a commitment to gradually streamline the processes for improved efficiency. The streamlined and improved processes will create more time and new opportunities for MEF staff to engage in critical and analytical work using data provided by the system. Furthermore, the accessibility of the system across the Ministry means that work does not need to be manually transferred across departments, but can be accessed by a range of agencies at the same time, providing new opportunities for horizontal interaction and for eliminating existing duplications of function. Making the most of the system then suggests the need for fundamental change in terms of its organizational set-up and its staffing profile. New tasks and job descriptions are required as well as a new structure of decision-making.
From a political economy perspective, PFM Reform in Cambodia generally, and FMIS in particular, represent a complex challenge, with contending pressures facing the government. On the one hand, the government has prioritized PFM reform as a key plank in the strategy to improve the quality of service delivery. On the other hand, the dynamics of existing informal and personal relationship within the civil service means that reform of civil servants' roles and responsibilities has historically proceeded only slowly, cautiously and following protracted negotiations. Even though part of the pay-off from the reform would be improvement in budget execution efficiency, the program raised specters of redundancy, loss of job, status, and other benefits.
In our model, institutional culture is affected not only by the history of the institution and the values that circulate within it, but also the political economy context within which organizations operate. The lack of powerful horizontal accountability institutions makes it harder for the government to tackle such obstacles. Cambodia's anti-corruption institutions were established fairly recently with limited capacity, compared with other countries in the region.
Further, the FMIS has been implemented in a context where broader public discussion outside the government of public financial management issues is limited. Although corruption and quality of services are election issues, there is no strongly and publicly articulated link between PFM reform in general, or FMIS in particular, and moves to promote public sector integrity and service delivery.

Managing the Change to FMIS
Overall the experience in Cambodia can be characterized as a process that got off to a rather sluggish start but picked up momentum later on, due to a combination of circumstances and an implementation and design shift.. The Cambodian case suggests two main conclusions: first, that well-designed change management needed to be given more weight from the outset of the project, given the relatively narrow window of opportunity for reform; second, although technical skills are of course important in any IT upgrade, the key problems here were political and administrative and thus change management techniques needed to be better connected with a broader analysis of the political context and of the specific institutional culture.

Leadership
The FMIS had formal backing from the highest levels of government as well as from leaders within the MEF. It formed part of a key post-2013 political strategy to increase fiscal revenues and to improve the timeliness in payment processes. As such, it represented a political strategy designed to solve what the government considered to be a pressing problem in addressing public service delivery.
However, some actors at middle management levels were less supportive. In part, this was considered by the incoming change management (CM) team to be an issue of "ownership." Initially, the FMIS program was not widely understood to be a transformational change project within the MEF. Rather, it was understood primarily as a software upgrade, to be implemented by the IT Department that could operate alongside existing systems until such time as managers were comfortable to switch over. This attitude fit comfortably with the government's formal "evolution not revolution" approach to the program.
The danger of this approach, however, was that it represented a strategy by which some middle managers could uphold existing practices by minimizing the impact of change. For example, some middle managers began to ask for the software to be customized in order to better fit with existing practices, rather than beginning the process of aligning their own manual procedures and working practices with the new FMIS system. Requests for customizations are generally regarded as lowering the efficiency of the system and increasing the costs of system upgrades, as well as increasing the risk that old, inefficient manual processes will be replicated, rather than taking the opportunity to streamline procedures.
The complexity of the FMIS system, the lingering perception of it as an IT issue rather than as a vehicle for transformational change of the functioning of the Ministry, and the reluctance of some middle management to begin to realign their departments to this change were significant features of Cambodia's experience. When this became clear, more attention was paid by project managers to designing change management tools, and change management approaches focused considerable effort on both senior leaders in the Ministry and on middle managers. When the depth of resistance of middle management became evident, the CM team updated their stakeholder analysis, holding face to face meetings with directors of the different departments in the Ministry to gauge their concerns. These meetings revealed different levels of support among different Department Directors and Deputies, with the most impacted department in particular maintaining a neutral attitude to the FMIS system.
The department that was most strongly associated with FMIS was the IT Department and the Director of this department was the Project Manager for FMIS. For the IT Department, the roll-out of FMIS represented a massive increase in workload, particularly since procurement of required materials and staff recruitment processes took a long time. Members of the IT Department were very dedicated to the program. However, the IT Department is a relatively low-status department within the structure of the MEF, compared to the main budget-handling departments. Furthermore, because the IT Department is located in the organizational structure of the Ministry under intended to bring together heads, deputy heads and officials of different departments of the MEF to resolve policy issues to enable smooth implementation of the system and to ensure that all departments had input into the change process.
The FMWG was vital for disseminating information about, and enthusiasm for, the reform across the Ministry's general departments and departments. Its main role was to facilitate coordination among key general departments and departments that are impacted by the changes. However, FMWG members also had full time reponsibilities in their respective general departments and departments, reducing the time they could spend on coordination. Further, the team represented a team of equals rather than a group operating under the direction of a senior leadership, and this state of affairs has proved unsuccessful in a variety of reform contexts in Cambodia, due to the power of vertical relationships within the Cambodian state (see Hughes and Un 2008). In the MEF context, it proved difficult to maintain interest and attendance from FMWG members. The FMWG had to issue a strong statement to demand greater commitmnet from FMWG members, which had some effect on boosting attendance at meetings, if not creating genuine ownership among members. At least until the system went live, the FMWG was not able to overcome the perception of FMIS as an IT project which their department had been dragged into and develop a shared goal of a transformational change for the entire Ministry which they had a responsibility to direct. One FMWG member interviewed for this report commented that dissemination of information to different departments occurred through the FMWG and without this, the change process would have been much more difficult, but nevertheless for much of the life of the project it represented a somewhat token effort rather than a genuine and proactive embrace of reform.
A further weakness in project governance was the status of change management itself. There was a consultant to support change management work under the project; however, this did not work well even though the consultant was supported by the CM team comprising of selected staff or representatives from General Departments. Subsequently, a local CM team was developed with both government officials and new recruits (11 persons mostly from communication background), supported by the World Bank task team. The Bank task team also includes a team member from Indonesia who had been responsible for the change management strategy in Indonesia's own implementation of FMIS reform. However, this took place only at a relatively late stage in the process, and delays in approvals for recruitment of needed staff under the project entailed that the CM team reached full strength only after the FMIS system had already gone live. Consequently, there was little initial analysis of what kind of support for and resistance to the project might be expected.
In the absence of such an analysis, the governance of the FMIS project was organized in a manner which replicated the institutional structures of the ministry, rather than improving them. The project was located in a relatively low status position in a strongly hierarchical structure. Attempts to formalize a horizontal steering group of "stakeholders" from across the ministry foundered because of inadequate involvement of sufficiently senior leaders who could demand compliance. The steering group thus became an intermittent forum for negotiation between departments that were skeptical of the reform, and the project leaders did not have the power to challenge this perception. Once the CM team was in place, emphasis on leaders' education was increased, and facilitated through the circulation of newsletters, the production of video messages, the conduct of face to face meetings and the plan for power lunches of key department directors.
A series of face to face meetings were conducted with directors and deputy directors of departments in October 2015. The purpose of the meetings was to convey three key points to the directors. The first point was that excessive customization would result in eroding the efficiency of the system and failing to achieve the intended results. The second point was that FMIS was not designed to undermine their authority but to enable them to make better decisions. The third point was that FMIS would enable them to contribute more effectively to the achievement of national development. The meetings also allowed an opportunity to gauge the level of support of the leaderships of the different departments.
This occurred at a late stage, however, with the result that few efforts at organizational alignment had been made before the software went live in January 2016. Partly in response to this, the PFMMP was extended for a further twelve months in order to facilitate embedding of the reform following the roll-out. During this time progress began to be made on producing Standard Operating Procedures (SOPs) that would serve as a starting point for rethinking staffing. The SOPs was a streamlined process both inside and outside the system and it provided the system users the required information for operating with the FMIS.

Engagement and communication with stakeholders
The existence of many formal statements of support for PFM generally and FMIS in particular by the highest political leaders was interpreted as evidence of "political will" to proceed. Although there is a clear political rationale for reform, this has required complex negotiations between political leaders and groups of civil servants who are powerful political "insiders". Stakeholder engagement in this context is therefore a highly political and complex affair involving not only formal attendance at project-related events but intensive behind-the-scenes negotiations between political allies that are likely to remain entirely opaque to outsiders, including development partners.
A Stakeholder analysis was conducted in September 2014 to facilitate the change management strategy and activities. However, the analysis took a quite narrowly technical/administrative rather than specifically political view. Stakeholders were identified in three groups: key influencers, users, and "need to know". These were identified entirely from within the MEF and provincial treasuries. Interviews were conducted with ministry leaders including the Minister and a range of directors, deputy directors and officers across the Ministry, and resulted in a very inward-looking understanding of the purpose of the reform. This makes sense from the point of view of training, but it reinforces a top-down, administrative conception of the motivations for change which over-emphasizes the internal environment of the MEF at the expense of the external political strategy that the government was attempting to pursue through better PFM management.
In the Cambodian context, there are few organized actors working to promote public advocacy on issues to do with budgeting, public spending and corruption (Public Expenditure Review 2015). This implies little direct political pressure, beyond a broad electoral imperative about quality of services, to show publicly that the reform is proceeding according to plan. This is an important contrast with the situation in Indonesia, where action against corruption and to promote the integrity of public finances is significantly higher on the political agenda and has organized constituencies within civil society and the media working to promote day to day action within government on the issue.
A Change Readiness Assessment (CRA) concluded that in this situation, the lack of consensus within the MEF over how far business practices were going to change had produced a great deal of uncertainty among the future users of the system with CRA score clustered above 3 points in which the score of 1 is positive perception and the score of 5 is negative perception (Figures 2 and 3). Three different versions of the business processes were produced which caused confusion, particularly among provincial treasuries whose existing IT competencies were far lower than in the central MEF. This was supplemented with face to face meetings with middle managers, as described in the previous section, and the organization of a major public launch event to mark the system going live. These high profile events would be supplemented with regular newsletters describing success stories and the development of a "Change Agent Network" of trained staff in provincial treasuries who could spread awareness among their co-workers about the forthcoming reforms. The overall focus was on obtaining enthusiastic high-level endorsement of the project as a key plank of political strategy, in the face of increased bureaucratic resistance at the administrative level. Also, a new plan was prepared to focus attention on the development of SOPs to be used in conjunction with the new system, which would assist in working out the details of how future business practices were to function.
Once the system was rolled out to the Provincial Treasuries, the enthusiasm for FMIS at this level increased further (findings of field trip to provincial treasuries, April 2017). Provincial Treasury directors reported that they were frequently subject to sudden demands from the center for reports on particular issues, and that previously, with the manual system locating and analyzing the necessary paperwork to prepare such reports had been difficult. With the FMIS system, the 2-3 days work was replaced by a few minutes taken to print out the relevant information from the electronic system (personal communication, Provincial Treasury Director, April 2017). For Provincial Treasuries, FMIS had evident and immediate benefits in lightening the burden of work, and offered them an assurance of accuracy that freed them of a significant weight of worry. Early apprehension among the PTs thus appeared to have been largely due to their lack of familiarity with the system, suggesting that efforts to allow PT staff to have handson engagement with the software at an earlier stage might have boosted pressure for reform from below. Change management tools to measure "readiness" had already identified PT staff as potentially more committed to FMIS than central ministry staff. Supplementary political economy analysis to identify the potential of PT staff to act as drivers of change could have augmented change management strategies, allowing PTs to contribute to the processes of driving redesign of business processes from a more customer-centered perspective.

20
Organizational alignment As discussed above, the Change Impact Assessment conducted in April 2015 found that FMIS could facilitate enormous efficiencies in the work of the MEF if it was permitted to take over from time consuming manual processes and if work practices, organizational structures and job descriptions were appropriately realigned.
This represented a potentially transformational change in the structure, functioning, and institutional culture of the ministry.
The key challenge for FMIS in Cambodia is that, up to the launch of the system, the full extent of the potential for change had still not been grasped, and few organizational alignments had been put in place. As a result, early indications suggested that FMIS was being used as a reporting system alongside existing manual procedures, an approach which actually increased workload rather than reducing it. However, once the system went live in provincial treasuries, this problem subsided. Some duplication of effort continued, in particular because a number of provincial treasury staff reportedly "distrusted the system" and were concerned about their answerability for any mistakes the system might generate. However, in some provincial treasuries, particularly where the director saw opportunities in the system to reduce workload by easing the onerous task of manually accounting for money, uptake of the system was relatively fast. Arguably, for provincial treasuries, where staff face conflicting pressures of demands from superiors for accurate reports and demands from customers for efficient service, the new system represented a significant improvement (findings of field trip to provincial treasuries, April 2017).
An important lesson from this experience is the need to examine not just the functioning of an institution but the values and relationships that are expressed through its modus operandi. While top political leaders increasingly support the need to improve efficiency as a means to enhance broader legitimacy, the difficulty of negotiating this in the particular environment of government ministries requires enormous expenditure of political capital, and this level of engagement was not evident in this project. Indeed, the mantra of "evolution, not revolution" suggests a concern to accommodate the worries of civil servants. Although organizational alignment remains a highly contested process, it is clear that some efficiency gains are being achieved. At the same time, some practices persist that could have been rendered redundant by the FMIS.

Workforce enablement
Workforce enablement focuses on the attitudes of lower level staff who will have to use the system. In Cambodia, the key issue has been on training and technical issues, particularly in the context of the Provincial Treasuries where experience of using online tools is very low. Change Readiness Assessments indicated the need for further training, information and awareness sessions to be carried out. A key sticking point was the slow pace of approvals for staff, procurement of equipment and access to needed resources. Addressing this requires improved coordination and closer working relationship between concerned agencies within the MEF.
However, as indicated in the previous section, once the system actually went live across Cambodia's provinces, it became evident that there are more interests in the system at this level, and change management plans designed to leverage this finding were established. This experience suggests that conducting political economy analysis at different levels of government could have assisted in identifying differences in receptivity to the project by different types of staff located differently within the institution of the state.

The Challenge of Budget Execution and Treasury Modernization (SPAN)
The SPAN project had several core objectives. Its primary goal was to establish a core financial management system that could underpin budget planning and budget execution for the Government of Indonesia, increasing transparency in all aspects of financial management and improving the efficiency of Treasury functioning. This would translate into faster payments, reduced errors in payments, and ease of operations for those making or pursuing payments.
Furthermore, it was intended to assist in transforming the cultures of the budget planning and budget execution business units through increasing the structure and discipline of daily operations, improving compliance with existing standards and regulations and reducing the time taken to perform activities and tasks. Finally, the project was intended to improve the accuracy and reliability of financial reporting, ensuring that budget execution is well aligned with budget plans. These objectives have broadly been achieved.
From the outset, SPAN was regarded as a complex undertaking incorporating changes in processes, structures, technology, and people management. Because root and branch systemic change was envisaged, this was understood from the outset as having a significant impact on staff as well as decision makers in terms of the knowledge and skill to perform the business processes. The integration of previously isolated systems within the Treasury and budget processes brought new knowledge and skill requirements, which affected the competencies required of the staff.
Furthermore, the change to the business processes and ICT (Information and Communication Technology) created significant changes of roles, responsibilities and workload within the organization structure of DGB (DG Budget) and DGT (DG Treasury). These changes raised the need to equip employees with the knowledge of new roles and responsibilities.
The ongoing BR program provided a window of opportunity since it had already begun the process of reforming the institutional culture of the MOF. The long time-frame for the GFMRAP program entailed significant turnover in leadership -there have been three Indonesian presidents, six finance ministers, and four World Bank task team leaders over the life of GFMRAP. Nevertheless, the firm basis for the project in law and the continued commitment to the reform agenda within the Ministry of Finance, bolstered by continued public concern about the issue, ensured political support.
Nonetheless, this does not mean that the reform process was uncontested. The project had to contend with some conservative attitudes on the part of civil servants. Even though the project had built-in pay incentives for employees to promote their engagement with the change process, this did not entirely offset concern about the impact of SPAN on jobs and positions. Initially, SPAN leaders took the view that the reform could be driven in a top down manner; however, discussions between stakeholders revealed the extent to which different agencies had different attitudes and interests emerging from different ways of measuring performance and organizing staffing. The potential for significant institutional resistance became evident.

Managing the Change to FMIS
The CM Team faced several challenges. The initial de-prioritization of change management entailed that the early management of SPAN within MoF occurred without any change management strategy in place. Even once the CM team was formed, the structure of the SPAN project saw the SPAN-Change Management & Communication (SPAN-CMC) Project Team designated as one of the work-streams under the Technical Coordination Committee. This implied only moderate access to change leaders, presenting challenges in developing the leadership dimension of the Change Management Strategy. Furthermore, the delays in making progress on mapping the business processes meant that the CM Team was delayed in formulating a strategy. Ultimately, the team formulated a strategy that covered not only the change management plan itself, but a range of training, monitoring and risk management activities also.
Preparing the change management plan required assessing a sense of change readiness and the extent of required change, and then designing and implementing appropriate change management strategies.

Leadership
Change managers supporting the SPAN project took the view that one of the main principles for managing change in an organization is "Change Flows Downhill". This principle emphasizes the role of leaders of the organization as the sponsor, driver and sustainer of the change initiative articulating the "change vision" that provides the context for every individual's effort to promote change.
In the first, "Assessment", phase of the change management strategy, the results of a Stakeholder Mapping and Analysis and Change Readiness Assessment (CRA) uncovered the perception of the limited "visible" involvement and support of the leaders in key MOF departments. Based on that discovery, a set of activities for leaders were designed to increase the effectiveness of leaders in promoting reform. A SPAN Implementation Committee of top leaders was established to ensure the engagement of all relevant department heads in planning and monitoring of the program. This allowed a mechanism for ensuring that SPAN leaders were always updated with the latest essential information on implementation, provided with key messages that needed to be communicated throughout the Ministry, and providing them with a schedule of activities that could make their commitment "visible" to lower levels of employees. Leaders were also provided with training on the role of leaders in organizational change.
Specific activities undertaken included the creation of a SPAN Message Video from the Minister of Finance, articulating the case for change and showing clearly the Minister's personal commitment to the program. This could be shown to lower level employees of the Ministry of Finance (MoF), boosting their perceptions of the political support for the program. The change managers also organized regular "power workshops" in which key leaders, primarily the directors of DGB, DGT and PUSINTEK 2 , met, discussed progress and shared the latest information.
The same leaders were also engaged in "discussion roadshows" with the leaders of the technical coordination teams.
This focus on leadership was successful because of the Indonesian government's strong commitment to improving efficiency following the change in regime. However, despite this broad support, one issue that hampered the success of these strategies was the regular changes of personnel at the top management of the ministry.
Finally, a committee called the SPAN and SAKTI Implementation Committee (KISS) was established by decree of the Director General of the Ministry of Finance in each affected work unit. In each unit, the KISS Committee comprised the Head (Director/Head of Work Unit), the Deputy Head (Head of a Department), and the Secretary (identified SPAN change agent of the work unit). These committees operated to connect leaders in regional work units to the central Ministry. The committees were regularly provided with up to date information from the Centre, were involved periodically in SPAN related activities, and offered an opportunity to build relationships between the central team directing the change and regional implementers. The committees were expected to take responsibility for mobilizing support within their own work units for the change, based upon this information. The committees also provided a context for a network of "change agents" to work to mobilize support for the initiative.

Project Design and Governance
To ensure the availability of dedicated full-time staff to manage and implement the SPAN, in 2008, a new Echelon II unit called the Directorate for Treasury Transformation was established within the Directorate General for Treasury, primarily to manage the reforms and oversee all activities related to the development and implementation of SPAN. Furthermore, the governance structure for SPAN was developed with a number of specially defined roles.
The role of Project Sponsor was defined for the Minister of Finance, while the Director General of Treasury, the Director General of Budget and the Secretary General of the Ministry were each assigned the role of Project Owners.
The Director of Treasury Transformation was assigned the role of Project Director. A Steering Committee was created along with a number of Technical Coordination teams, a Change Control Board that evaluated all change order proposals before they were submitted to the Steering Committee for approval, and a PIU.
The emphasis on strong leadership in SPAN was embedded into the SPAN project structure, which comprised two

Engagement and Communication with Stakeholders
The key stakeholders identified with respect to the implementation of SPAN were MoF employees within DGT, DGB, PUSINTEK and other related units (approx. 10,000 people) and other government ministries and agencies (approx. 96,000 people). The main challenge during the project was the acceptance or buy-in level from affected employees during the implementation stage. Although the promised benefits of the project for many employees were clear, communicating the benefits of change was not as easy as expected. Reaching out to all layers of affected employees effectively with comprehensive information concerning the project, its benefits, and the progress of reform while keeping users and other affected employees was not easy. Also, the introduction of SPAN would render obsolete a range of pre-existing Treasury systems, and consequently it invited enormous resistance from groups who had created and maintained these many island systems. These users feared a loss of their role or their power within their agency.
A number of responses to this problem were required. Initially, a stakeholder mapping exercise was required, to determine different categories of staff with different kinds of concerns about the change process. Furthermore, a CRA was conducted to gain understanding of different stakeholder groups to ascertain their level of awareness of the vision, objectives, features, and benefits of SPAN implementation, and specific benefits for individual stakeholder groups. The CRA assigned three categories of readiness across the five areas measured, and found that there was progress in all areas between the first and second assessments (Figure 4). The assessment also investigated the current level of commitment of key stakeholders, their perceptions of critical success factors and impediments to implementation, their willingness to adapt to the new environment and potential constraints on this, and their current level of involvement in project implementation. Data about these issues were collected through a combination of detailed interviews, surveys, and focus group discussions and used to generate a set of action plans. Together the stakeholder mapping assessment and the CRA yielded insights into the political economy factors underlying attitudes towards change. The CM Team worked on the basis that tailored messages were required for different audiences and that the effect of messages disseminated had to be continuously monitored to check for understanding by recipients. The case for change and benefits of SPAN were also re-emphasized periodically through SPAN Key Messages and SPAN communication programs. However, this was not effective for all staff: some pockets of staff showed consistent resistance to the end of the project. The CM Team tried to come up with specific communication activities for these staff, but stretched resources entailed that these were not consistently performed. In this area, performance of a more detailed political economy analysis of the groups of staff involved and possible ways in which their concerns might be addressed through forms of compensation or incentivization, rather than simply reiteration of the core message, might have been useful.

Organizational Alignment
This vision and the strategy for rolling out the program also had to be embedded in a human resources transition plan which could address the problem of staff concern about potential job losses. This was not done at an early enough stage in the process. MoF should have been able to communicate a clear plan for managing the transition in job responsibilities after SPAN and to counsel staff effectively to reduce fear, uncertainty, and resistance. The transition plan should have been decided and agreed upon early in the project at the political level. The initial proposal to offer a golden handshake to some staff and to transfer others was not approved by the learning among co-workers, blended formats combining classroom and hands-on approaches, "Train the Trainers" programs, focus on core processes, and e-learning. The results of the CRA suggested that this highly elaborated approach to training was significant in promoting engagement with change among MoF staff.
A further issue in workforce enablement is the incentivization of employees. Monetary incentives were considered to be the most significant; these needed to be continued after the change was implemented to ensure continued adherence to new business operating procedures. There was a risk that the government might not have the budget or even the willingness to provide this monetary incentive due to budget constraints or in the context of any deterioration in the economic situation. However, following the modernization of KPPN (payment) offices in support of PFM reform in 2007, incentives in the form of additional allowances resulting in higher monthly take-home pay were rolled out to all staff working in DG Treasury and DG Budget, starting with those who worked in payment offices deemed "excellent" (or "KPPN Percontohan").

Conclusion: What Does the Study of FMIS Suggest about the Contribution of Political Economy Analysis to Change Management?
This study has investigated the claim that combining change management tools with political economy analysis can provide an improved approach to promoting governance reform, with specific reference to the implementation of FMIS in Cambodia and Indonesia. The discussion proceeded from the following premises: 1. Political economy analysis can offer an assessment of whether there is a window of opportunity for reform to take place in a particular political, social and economic context.
2. Political economy analysis can also offer a detailed assessment of the various social forces that will tend to support or resist change, and the relative resources of power that different social forces have at their disposal to promote, block or somehow influence the course of reform.
3. Given that any reform context is likely to include forces for and against, the field of reform is a contested site, where outcomes are at least to some extent contingent. In this context, change management can make a difference, but is more likely to be effective in promoting particular outcomes if strategies and activities are tailored to transforming or supporting key groups.
In both the Cambodia and Indonesia cases, change management strategies were employed to facilitate the implementation of FMIS in support of public financial management reform, but late in the day and with insufficient resources, particularly in Cambodia. At issue in both cases were the questions of whether FMIS would be implemented and used, and whether this would set in motion wider programs of public financial management reform.
The empirical record suggests that in both cases, FMIS has been implemented and used, and the Task Team estimates that about 20% or so efficiency gain has been achieved in both cases. In Indonesia, the adoption of IFMIS SPAN has certainly been a key part of a broader and highly complex story of improvements in public financial management.
In Cambodia, the adoption of the FMIS also represents a shift in the way public finance is managed and more efficiencies are expected to be realized as operating processes continue to adapt. The uptake of the system by the central ministry and provincial treasuries has been more enthusiastic than anticipated and that momentum is building for a reorganization of operating procedures as a result. Consequently, both cases represent qualified success.
However, in both cases, also, there were significant delays and problems, offering opportunities for lessons learned.
In this concluding section, we discuss the question of what the integration of political economy analysis into change management strategy might suggest, in hindsight, in terms of new strategies for tackling the problems that arose.
This can form the basis for developing a model of integrated practice with respect to political economy analysis and change management which can then be tested in the implementation of future programs.
In section 2 we discussed the ways in which the change management literature and the various approaches to political economy conceptualize change. The change management literature has evolved from a fairly optimistic approach which assumed all change to be rational and resistance misguided, to be overcome by better communication, to more sophisticated models featuring not only the interests and competencies of different groups in an organization but also their values, norms and cultural practices. A basic tenet of change management is that all of these are changeable given appropriate tactics on the part of suitably empowered actors.
Political economy analysis, of the type employed in this study, focuses on the distributions of power and interest between key actors advocating for or against change. These actors are broadly defined and include groups and individuals outside the institution concerned. They may include donors, businesses, service users, civil society organizations, the media and state agencies themselves. Depending on the context, different groups within these categories will have more or less power and interest to press for change. The relative power of different groups can also change as the political economy context changes.
Political economy analysis tracks the iterative interaction between structural constraints and the struggles of different actors to pursue key goals. From the perspective of political economy, socio-economic contexts are dynamic, but relatively immune to manipulation by change managers. However, the responses and strategies of actors to these contexts are highly contingent and can be influenced. Even though political economy analysis draws attention to the fact that different actors have different distributions of power and interest, actors have a choice over the strategies they use to maximize their influence and whom they choose to ally with to protect their interests. Those who engage in managing change may be able to significantly influence the prospects for change by offering actors more choice, better strategies, and new allies. To analyze the case of FMIS reform we developed a model in which we see the political economy context as giving rise to two potentially conflicting areas of activity within the state: political strategy and institutional culture.
Governments respond to demographic change, political crisis, technological innovation or economic shocks by developing new political strategies that keep up with the times. These new political strategies are intended to shore up important constituencies of support and marginalize opponents. Part of the strategy may be articulated in the form of election manifestos and policy agendas, and other parts may be deliberately hidden. It is important to remember that political strategies may not be internally coherent and they may actually be contradictory -the strategy in one area may directly undermine the strategy in another area, and which strategy prevails will depend upon internal power relations within government.
In both Indonesia and Cambodia, high political leaders supported the adoption of FMIS because it advanced key political strategies. In Indonesia, the goal was to boost domestic popularity with the public and international credibility with development partners and investors by tackling corruption and exhibiting a responsible approach to the management of public finances following the Asian Financial Crisis and the fall of the Suharto regime. The situation was one of regime change in a context of intense political and economic crisis. In Cambodia, the goal was to support improved budget execution and service delivery in response to higher expectations from citizens, as reflected in the government's narrow election victory in 2013 and its much reduced majority in local government elections in 2017.
In both cases, new political strategies had to contend with embedded institutional cultures of hierarchy, informal networks, and patronage within the civil service. We use the term "institutional culture" to include ways of working, perceptions of interest and authority and habits of control within a particular state institution. Institutional cultures emerge over time, but are intimately linked to the role of the institution in mediating the relationship between political elites and the broader society. They are thus closely related to shared beliefs among civil servants about the nature of public service and the national interest.
In both Cambodia and Indonesia, FMIS represented a challenge to the status and power of civil servants in influential and prestigious government departments. Consequently, resistance could be expected creating a space of contestation around the reform in which various actors maneuvered to attempt to maximize their influence and promote their interests. Resistance was, in fact, encountered in both cases, but at the time of writing, appears to have been overcome in both cases sufficiently to claim success in the reform. Political economy analysis can shed light on the political dynamics of this zone of contestation, offering insights to change managers attempting to design effective change management tools and strategies appropriate to the reform and the reform context.
The argument of this working paper is that many lessons can usefully be learned by reviewing already implemented processes of change management from a political economy perspective. Furthermore, the transfer of change management techniques from one context to another can be helpful in improving processes of reform and mobilizing support from affected staff and other groups. Comparison of the cases of Cambodia and Indonesia, however, also suggests that change management strategies are limited by local contexts even as they help to transform that context.
Change experts working on IFMIS SPAN in Indonesia and then moving to assist with FMIS in Cambodia found that identical strategies could not be rolled out in the Cambodian context, and where similar approaches were used, they did not have the same effect. This was due to some significant differences between the Indonesian and Cambodian contexts.
In this working paper, we have disaggregated change management techniques into five sets of issues: leadership; project design and governance; effective communication; workforce enablement and organizational alignment. In each of these areas, we used political economy analysis to show how the problems that arose with implementation of the project arose from distributions of power and interest among actors in this zone of contestation. Change management tools can intervene in this zone of contestation. Although change managers cannot reconstruct distributions of power and interest, they can affect them at the margins, making some actors relatively more powerful and influencing perceptions of interest of others. Here we draw together some key similarities and differences between the two cases to demonstrate the marginal utility of change management in promoting reform.

Leadership
The attitude of leaders to reform is not likely to be strongly affected by change managers, since leaders are oriented to broader political imperatives. In these cases, both Ministries had leaders who espoused the reform agenda. In Indonesia, FMIS represented part of a raft of reforms championed by a high-profile minister with a very hands-on approach and strong backing from the president of the republic. In Cambodia, leaders adopted the slogan "evolution not revolution," in line with a broader prioritization of stability, and day-to-day management of the project was delegated to a working group of middle managers with varying attitudes to the reform and little experience of horizontal coordination. At the provincial level, managers were initially highly uncertain but later saw the most benefit to themselves from the reforms.
The CM team in Cambodia was able to promote FMIS by making video clips of ministry leaders talking positively about reform, which were widely distributed on social media. This was important in convincing provincial employees that the reforms were central to the government's political strategy and vision for the nation, even though middle managers were more resistant. In so doing, the support of the top leadership for reform was translated more effectively into an authoritative drive for change throughout the ministry, linking with greater enthusiasm for the change at the provincial level. Change managers were also able to organize activities such as an overseas study visit for managers from different General Directorates who were not accustomed to coordinating horizontally, promoting relationships between them, making decision-making faster and better, and thus promoting the reform.

Project Governance
In the Indonesian case, the project was governed by a raft of legislation produced in the aftermath of the political and economic crisis, which mandated reform as a matter of law. This created pressure for change, and senior leaders in the ministry were assigned specific roles as project sponsors and leaders, with accountability for progress of the reform. Perhaps as a consequence, dedicated units were established in key directorates with staff who worked full time on reform programs, answering directly to the top leaders. This created momentum behind the reform and provided considerable resources to reformists, despite the fact of resistance by some civil service units.
In Cambodia, by contrast, the project was driven from the IT Department in the MEF. This was a relatively low status department within the Ministry, compared to the powerful General Department of National Treasury and General Department of Budget. Furthermore, the IT Department answered to a separate Secretary General, and therefore had few channels of communication with the General Departments that were supposed to be reforming.
The establishment of the Working Group, which included all the General Directors, was supposed to address this, but it was difficult for the IT Department to manage the Working Group because of their lack of authority within the ministry. Initially skeptical attitudes reflected preference for upward reporting rather than cooperating among themselves. This entailed that meetings were difficult to arrange and decisions were slow to be taken. The CM team in Cambodia was able to strengthen the working group by bringing these issues to the attention of senior leaders.
Senior leaders responded by requiring the General Departments to make the working group a higher priority.
However, the CM team itself was situated in the IT Department, and therefore suffered from the same constraints as the project managers.

Communication and Engagement of Stakeholders
CM teams in Indonesia and Cambodia used diagnostic tools to assess "readiness" for reform on the part of a range of employees at all levels of the respective ministries. This was then used to create effective channels of communication with messages tailored to the particular constituencies targeted. This was in many respects successful, but some pockets of resistance remained in both countries. Change managers attempted to identify different reasons for resistance, and to tackle these with bespoke messages intended to reassure and persuade. At the same time, the morale of pro-reform constituencies was boosted by good news stories about progress and achievements. The contribution of political economy analysis here is potentially twofold. First, translating an assessment of "readiness" for change into an assessment of the intersection between interest in change and the power to drive it might allow better targeting of the most important constituencies, who might be most open to persuasion from whom improved support might be decisive. It can thus allow targeting of effort and conservation of resources, and in both countries, communication strategies were eventually limited by time and money. Second, such an analysis can highlight constituencies that might productively be brought together, so that their combined influence can be leveraged. For example, once FMIS in Cambodia went live, and Provincial Treasuries could experience using it, some Provincial Treasurers became enthusiastic advocates. Organizing interactions between Provincial Treasurers to share their experience could have leveraged this effect, as could organizing opportunities for senior managers to engage directly with the Provincial Treasurers, thus placing pressure on middle managers in the central ministry to come into line.

Organizational Alignment
This was the most controversial and difficult aspect of the reform in both cases. The transformative potential of FMIS can only be realized if standard operating procedures within the Ministries are realigned to make best use of the system. In Cambodia, this was a sticking point in the reform for a long time, because of the nature of relationships within the civil service and between civil servants and political leaders. In Indonesia, the civil service was undergoing radical reorientation in any case. In Cambodia, the civil service has not seen much upheaval since the end of the war twenty years ago, and is regarded by the government as a key force for stability in the country. Reforms that threaten jobs, hierarchies of status and longstanding habits of control are highly contentious in any context. In Cambodia, the political strategy of which FMIS was a part became more pressing and this put pressure on middle managers to accept the reform. However, the work done by change managers to inform middle managers and provincial treasurers about the kind of changes that could be realized was important in encouraging action to take place quite quickly once the opportunity provided by an impending election arose.

Workforce enablement
This has traditionally been understood quite narrowly in the change management literature as provision of training and technical support. For a reform like FMIS, adequate training is essential in enhancing the confidence of users of the system and thus gaining their support for change. In Cambodia, some of the most enthusiastic proponents of reform were the Provincial Treasuries, who are front-line service providers and are thus faced with pressing demands from both their superiors in the ministry for accurate reporting and from their customers for rapid and effective response. FMIS offered significant benefits to these staff; however, their lack of confidence over their ability to use the system inhibited the extent to which they could be seen as drivers of change in the early stages of the program.
Once the system was rolled out and they were able to use it for themselves, they become much more confident advocates for change. Political economy analysis would suggest prioritizing early training, possibly with sample software, for such constituencies who are potentially important constituencies for reform but are held back by lack of confidence in their technical skills. Similarly, a study tour to Indonesia, during which Provincial Treasurers could see their counterparts using the system, was an important turning point for the reform process, according to the Cambodian CM team.
In both cases, at time of writing, initial resistance appears to have been overcome and FMIS has become part of the landscape of public financial management in both countries. In Cambodia, in particular, where adoption is more recent, significant decisions are still to be made regarding further streamlining and organizational alignment, particularly once the system is rolled out to a range of line ministries, all of whom have different working practices and institutional cultures. A lot of work remains to be done to see how the line ministries will engage with the system and further analysis in support of change management will be required.
In Indonesia, active change management has continued to track levels of satisfaction and efficiency gains to maintain momentum and support behind the reform. At present, it appears that the reform is increasingly well-embedded in the day to day operations of the ministry, and momentum for bureaucratic reform has continued with the election of a reformist president, Joko Widodo. However, continued political economy analysis of reform momentum in both countries would represent a valuable resource for change managers given inevitable risks of political and economic uncertainty that accompany dynamic and fast-growing economies such as those of Indonesia and Cambodia.