PDIA to Inform Budget Reform in Mozambique

This note covers our experience working with officials in Mozambique’s public financial management sector, between September and December 2009. This was right at the start of the journey in learning how to do PDIA. We were exploring the basic idea that a conversation about problems could spark new ways of doing reforms. Mozambique’s Ministry of Economy and Finance (MEF) had come to the end of a ten year public financial management (PFM) reform initiative. The reform was widely hailed as a success, having established a new IT-based system for financial management. However, some observers (including experts at the World Bank) were concerned about compliance gaps and weaknesses in the system, which no one could quantify but which were thought to pervade the system. A new Public Finance Vision document had been developed (in June 2009) to design the next decade’s PFM reform, focused on continuing the reforms as they had been done (with more IT roll-out and expansion).

World Bank experts wanted to try and shift the design of future PFM reforms to more actively address compliance concerns (to minimize gaps and weaknesses). These experts asked Matt Andrews (from BSC) to advise on this, in consultation with managers and technical staff of the budget and treasury departments in the Mozambique MEF.

Matt experimented with the first principle of PDIA—helping government officials focus on problems instead of solutions when designing reforms. Rather than designing an improvement strategy for PFM reforms, and proposing a new vision document, he convened and facilitated a discussion about the gaps and weaknesses in the PFM system and in past PFM reforms.

In the first stage of this work, Matt worked with a small World Bank team to develop an external, evidence based analysis of gaps and weaknesses in the PFM process. This was then presented to a group of authorizers in the MEF and a group of MEF technicians. The aim was to make these key agents aware of the problems (expressed as compliance gaps and weaknesses), and to get these agents to adapt their future ‘vision’ strategy to better address these problems. A second stage of the work emerged after the authorizers, struck by the evidence of gaps and weaknesses, asked for a broader engagement around the issue—including budget users from ministries, provinces and districts. These agents were included in a 30-person workshop where Matt solicited the views of all participants about strengths, weaknesses, and gaps in the PFM process (‘opinion evidence’ of the problem) and facilitated a broader discussion on why these gaps existed (and persisted) and in how gaps might be addressed in a different reform process.

The first stage of the convening process was successful in building awareness of authorizers about problems in their process, and having these authorizers nominate a clear ‘next step’ for the process—the broader workshop. The broader workshop was then also successful, in fostering a broad and localized discussion about the problems —what gaps and weaknesses existed—and why the problems persisted. This workshop ended with the participants recommending a new approach to developing the Public Finance Vision; the Vision would focus (at least in part) on closing gaps (which had been clearly identified) in PFM system implementation, especially at the point(s) of service delivery; broad-based teams (including line ministries and service delivery implementers) would identify potential strategies in each gap area, to inform content in the Vision document; the teams would then take responsibility for implementing the strategies. Unfortunately, the MEF did not follow this approach and continued with the Public Finance Vision it had already commissioned. The mainstream PFM reform thus did not directly address gaps and weaknesses identified in these workshops, and instead continued rolling out new IT systems and processes that had up to that time undermined by gaps and weaknesses. Unfortunately, these compliance gaps persisted and contributed to a major PFM-related government corruption crisis in 2016 (where significant finances were found to have been spent outside of the PFM system).¹

There was a positive result from the workshop, however. An entity responsible for internal audit (the IGF) built upon the observations made at the workshops and initiated an audit of compliance with the IFMIS. It was the first time the IGF had conducted such an audit, and it found significant new evidence of the compliance gap. This ‘practice’ enabled the IGF to build its capability to do such work, and motivated the IGF to do the audit again and to expand its mandate to allow more audits of the PFM system. A 2014 law institutionalized these changes, and showed that a more engaged internal audit function emerged from this work (at least partially). A second positive result emerged in 2010 and 2011. Select line ministries involved in the second (broader) workshop engaged the World Bank and MEF to propose a new project focused on addressing PFM weaknesses in the sector delivery processes (similar to the idea proposed at the end of the workshop). The idea was accepted, and a new project was designed over the next two years (the Program for Results, supported by the World Bank).² It took a while to develop³, and was difficult to deliver, but (even though it is called a ‘program for results’) the project managed to introduce a new, broad-based, problem driven approach to doing PFM reform, where teams of government officials at different levels (central, provincial, and local) were mobilized to solve practical PFM process problems and foster better service delivery.

The project ultimately helped ensure that over 1,000 schools complied better with policy requirements related to financing, facilitated process improvements that led to budgetary funds flowing more effectively to over 1,000 schools, facilitated the identification and implementation (through iterative experimentation by public officials in their day-to-day work) of new ways to procure, store, and deliver medicines to health centers; and more.⁴

¹ See Andrews, M., McNaught, T., and Samji, S. 2018. “Opening Adaptation Windows onto Public Financial Management Reform Gaps in Mozambique”. CID Working Paper.

² While the project idea arose from the 2009 problem discussion, the design emerged from a process between 2010 and 2013 (discussed in a PDIA in Practice note ‘Contributing to a Problem Driven Project in Mozambique’).

³ The long process of World Bank project preparation allowed for the emergence of a second externally financed operation, USAID’s Results-Based financing in Mozambique’s Central Medical Store, which focused explicitly on the problems identified through this process and employed an iterative, adaptive approach to facilitating reform. This project has been evaluated as a success. See Spisak et al. (2016). Results-Based Financing in Mozambique’s Central Medical Store: A Review After 1 Year. Global Health: Science and Practice March 2016, 4(1):165-177.

⁴ World Bank (2018). Placing Results Front and Center in Health and Education in Mozambique.

Lessons about the potential and value of PDIA
We learned that the fundamental starting point of the PDIA process—facilitating a focus on problems by a broad group of agents—can change the conversation about policy and reform; broadening this conversation and focusing it on contextual realities of reform.

We also learned that the immediate results of a direct, externally led PDIA intervention (in this case, workshops) can be disappointing. The MEF did not adapt its main PFM reform approach after the workshops, suggesting the broader, problem-driven conversation had limited influence.

However, we learned that the PDIA process really has its impact via indirect or secondorder influences that emerge through local agents—where the externally led intervention nudges internal agents to think and act and interact differently, empowering new engagement and learning-by-doing, building internal capabilities to act, and fostering emergent and potentially surprising responses. (The emergent IGF activities and problem-driven sectoral PFM projects emerged from agents at the workshop who were enthused and empowered to try new things.)

We learned, over time, that the PDIA process can foster very different results to the ‘historical counterfactual’ (what would happen if reformers or policymakers continued creating their reforms or policies according to prior patterns). The main PFM reform pursued under the Vision looked very much like prior reforms and yielded the same kinds of gaps. The PDIA-inspired Program for Results reform was more focused on gaps, broader in engagement, and successful in yielding better compliance with PFM systems and better functionality of these systems.

Lessons about doing PDIA
We were encouraged that our ‘idea in principle’ held firm: an awareness of problems can shift ways of thinking and doing reform. We learned that the process of becoming aware of problems involved at least three parts: (i) using evidence to communicate to inside agents what the problem is and why it matters; (ii) facilitating a discussion across different groups of these inside agents about the problem (to validate evidence and agree on interpretations of the problem); and (iii) to propose next steps that mobilize a practical response to the new problem awareness.

We learned, also, that evidence and ideas developed by insiders themselves holds much more value (and has more impact) than evidence and ideas developed by outsiders. Sometimes insider evidence and ideas is more opinion-based, subjective or slow-in-coming, but it is very important.

We developed our first ideas on problem construction from this experience; where we use simple questions to get insider groups to identify (with evidence and narrative) ‘what the problem is’ and ‘why it matters’. The problem construction activity is the first step in any PDIA process. It builds initial authorization for the work, helps teams develop a common understanding of the work, and leads to discussions about entry points for action.

We also developed our first ideas and designs on adaptation windows from this experience; where the adaptation window is a facilitated engagement (of one or multiple interactions) where a group of insiders are convened to recognize, evaluate, and begin addressing problems in their reform and policy processes. These adaptation windows serve as stand-alone interventions that governments can engage in before deciding to move into the longer-term PDIA process.

We learned about different levels and stages of problem driven convening in government, which are almost always characterized by influential top-down authorizing structures. Given this, one needs to start the convening by building the awareness of high-level authorizers, and have these agents authorize a broadening of the convening process.

Understanding the political and bureaucratic authorizing environment
The process of convening meetings and workshops was not difficult, and if we were just focused on this the exercise would have been easy. However, the goal of shifting behavior and decisions made this much more complex. Achieving this goal meant that the work needed to influence decision-makers, and the decision-making process, and we struggled with understanding who these agents were and how they could be influenced. We worked with political economy analyses and other externally written materials to try and navigate this terrain, but found this ‘homework’ was only of limited assistance. This was most apparent when, even after the message of reform gaps was well received by both the official authorizers and technocrats, we faced the reality that these agents would not change their PFM reform approach (and opted rather to continue with their previous Vision). This suggested we did not know enough about the decision-making process, incentives of agents in such process, or how to influence the incentives of agents.

In retrospect, having seen what emerged after the workshops (and with many other PDIA experiments that we have undertaken), we have learned that our initial work overemphasized the importance of ‘us’ (the external agents) in this work—and thus overemphasized the importance of ‘us’ navigating the political and bureaucratic authorizing environment. What matters more is that the PDIA intervention sparks new ideas and engagements by local agents, which empower and inspire and enable these agents to manage their political and bureaucratic authorizing environment differently.

Dealing with ‘the messiness’ of problem driven conversations
Broad-based problem-driven conversations are not common in the development community, especially when they are focused on ‘the problems’ of past reforms and policies (and gaps and weaknesses of such). These conversations are also quite difficult, given that they bring to the fore realities that many agents would rather keep quiet or leave unspoken: realities about missed opportunities, perhaps, or failures in design or implementation, or weak leadership, and more. Dominant agents can become defensive in these conversations, or undermine new voices, or try to impede the sensemaking process that a problem-driven process requires. Less dominant agents can hold back in offering views or contribute to this sense-making process.

Given such realities, we struggled to facilitate a process that afforded a balanced narrative on the past reforms, that encouraged those who had been involved (‘you have done a lot that is working, but there are gaps’) but also promoted an urgency to shift approach (‘there are significant gaps that won’t be changed through a strategy of ‘more of the same’’). We also struggled to facilitate a conversation of many voices, offering different perspectives that generally lacked objective evidence. The result of this, at many junctures, was a ‘messy’ discourse that emphasized the ambiguity of reality and the tensions of change and power dynamics one finds in most governments. This was a struggle to manage, especially in a context where control and stability and ‘endorsement of the status quo’ was quite important. At times, donors invited to the workshops expressed real concern about the messiness of these conversations and even suggested that we as facilitators were ‘losing control’ and ‘wasting people’s time’.

Over time, we at the Building State Capability program have learned that there is no way to ‘miss the messiness’ in a problem-driven process, and that there is more room for such messiness than many in the development community would imagine. Ambiguity and disagreement can be powerful instigators of questioning and change, and even empower local agents to pursue different ideas and relationships in the pursuit of change. The key, we have found, is to manage the ‘heat’ of the conversation and to ensure inclusivity (where no one is identified as a villain and no one is identified as a saint, but everyone is treated as a resource and potential change agent).

Reflecting on our role, as external agents
We at the Building State Capability (BSC) program are not insiders in the Mozambican PFM reform process. We are located far away and do not have the same interests, incentives or personal and political investments as local insiders. However, we find ourselves working in quite deep and personal ways with insiders, trying to promote a reflection on the problems they face and to mobilize new ways of engaging with those problems in the future. We constantly struggle with our role, as external agents engaging so closely with internal agents. In this Mozambican engagement, we were working closely with a large international donor (the World Bank) that also has a complex role as an external player—informing and funding government reforms and working alongside other international bilateral and multilateral agents—but with a strong local presence. In this context, Matt Andrews struggled to work out if our appropriate ‘seat’ was on the government side of the negotiation table, acting as an external facilitator (or therapist) helping insiders ‘see’ their problems, or across the table from the government, as an external expert advising or educating the government insiders about their problems (from the global or even donor perspective).

The intervention actually saw us playing both roles (in the two stages), with different kinds of strengths and weaknesses. The first stage of work (the small, authorizers presentation) saw us acting as more of an ‘advising or educating’ outsider (an international academic entity working with a global donor, and using internationally endorsed data to show ‘the problem’ in the PFM reforms). The work we did in playing this role was important in gaining some authorization to engage broadly, but in hindsight it did not convince local agents to change. The second stage of work (the broader workshop) saw us playing more of an external facilitator role, drawing opinionbased evidence from insiders about their problems, and fostering a conversation amongst these outsiders about their own evidence. The work conducted under this approach was more messy at the time but led—ultimately—to more influence amongst local agents (with the emergent internal audit activities and PFM project coming from such engagement). In reflecting on this experience, and other PDIA work, we have found that external agents are always moving between roles, and should most importantly be aware of these roles and not presume how they are engaging or engage passively, which usually manifests in taking an ‘expert’ or ‘advisor’ role. We have also found that the PDIA work is most effective when we are playing the role of facilitator, and not expert, and ‘give the work back’ to the insiders.

Following this work, we continued engaging in Mozambique—taking the next step of helping to design the problem-driven PFM with government counterparts in the MEF and sector ministries, and with the World Bank team. This work caused us to further hone the approach we took in facilitating problem construction (identifying problems that matter) and also working on problem deconstruction (breaking a large and intractable problem down into manageable pieces and finding entry points for action) and then moving into rapid action and learning.