Author Topic: Conversation with Matt Andrews on limits to externally influenced PFM reform  (Read 43252 times)


  • *
  • Posts: 18
Responding to question 12

I am not sure of the situation but I would suggest that the positive progress in Albania is a blend of patient engagement, discussion and facilitation by external consultants, growing awareness over time by local folks, and serious steps to break down the incumbent mechanisms that held PFM back and look to new mechanisms. Note that I have not spoken about either coercive pressure by donors or pressure to copy others. Knowing some of the folks who have been involved in Albania, it seems to me that the big donor pressure has not been a huge part of the success here. So: I would not say that the key here is not about external pushing... rather eternal facilitation of internal change.

But from your explanation it seems to me that the external change may still be quite narrow. What i expect is that a small group of decision makers in the MoF and some central budget departments have bought into the new approach, and some in the parliament are along for the ride as well. But most agents in what I would call 'deconcentrated' or 'distributed' areas of government--the majority of parliamentarians, line ministry budget departments, line minsitry policy offices, and districts, are not as embedded in the new system. So the reform needs to diffuse to these folks if it is to stand any chance of really influencing how budgets are executed and what the 'ruls of the game' turn out to be in ten years time.

I derive this story fom my gneral understanding of PFM reform as something that is focused on narrow groups of specialists but seldom reaching distributed agents. This problem can only be rectified by increasing enegagement internally--not with more external push and pressure, as the folks who now need to buy into reform are not subjected to these pressures . These folks must be bought into the reform by local people, through connectors and conveners.

I am including the conclusion of my discussion about this in the book I am now reproducing, piece by piece, on this blog!

Getting inclusive about agency

The basic message in this chapter is that institutional reform requires more than the narrow engagement of institutional entrepreneurs, no matter how impressive these may be. It also calls for the involvement of distributed agents who ultimately have to implement the reforms. Unfortunately, externally influenced reforms are seldom this inclusive about agency. These reforms emphasize the roles of concentrated champions and heroes, and ignore those who are more de-concentrated or distributed. Andrea Whittle and colleagues call this latter group the “others” responsible for “more mundane and less prominent, but nevertheless essential, activities … associated with emergent institution-building.”  Their exclusion is a major reason why many externally influenced institutional reforms are poorly implemented and ultimately fail to change behavior or improve government.
A core part of the theoretical argument proposed in this chapter shows that distributed agents are required in most stages of institutional change. These stages are conceptualized in light of Greenwood, Suddaby and Hinings’ model, which posits that institutions do not change just because they are threatened by jolts or crises. Change only comes when incumbent institutional mechanisms are de-institutionalized and agents become engaged in a process of finding, introducing and re-institutionalizing a new set of rules. Table 5.1 shows the stages in this process and the roles institutional entrepreneurs and distributed agents play in each stage. It is obvious that distributed agents are key players in the final roles—diffusion and re-institutionalization—where others may prefer to use terms like implementation or roll-out. It is important to note, however, that distributed agents must be involved in earlier stages as well, or else these latter stages could be compromised. If distributed agents responsible for implementation are not engaged in finding the new solutions (during pre-institutionalization), for instance, there is a risk that such solutions will not be technically viable.   

The Mozambican study shows that distributed agents are seldom involved in early stages of institutional change, however, especially when reforms are influenced externally. Such agents were excluded in Mozambique on the apparent assumption that high-level champions and heroes would be able to force implementation by edict once designs were completed. This assumption is often made when reforms are adopted as signals and the questions about how they actually diffuse and drive behavior wait until after loans are disbursed. The assumption has been found wanting in places like Mozambique, however, and one can expect this to be the case in other cases of externally influenced reforms in development as well. Heroes often lack the influence they are assumed to have, and often they end up pursuing private gain instead of public interest as well. Both problems emphasize the risk external organizations take when they depend too much on champions to make reform happen. Alluding to such issues, Henry Mintzberg concludes his discussion on Ghana’s leadership problems by quoting from Bertolt Brecht’s play Life in Galileo.  A character in the play laments, “Unhappy is the land that has no heroes” to which another replies, “No. Unhappy is the land that need heroes.”

The next chapter discusses limits one can expect from institutional change. One of these relates to the idea of restricted reach in reforms. Restricted reach is a problem that results from an exclusive focus on heroes and champions. Not all externally influenced institutional reforms are so exclusive, however. Chapter nine discusses examples of reforms—in communities and governments across some of the more difficult countries of the world—where international organizations have facilitated inclusive processes of change. The engagement of multiple players in these processes has led to the shaping of endogenous solutions to contextual problems. Tying back to the book’s leading metaphor, this engagement fosters group-based craftsmanship that seems vital to ensuring the fit between reform pegs and the holes in which these need to fit.
« Last Edit: January 25, 2012, 08:26:15 GMT by Napodano »


  • *
  • Posts: 18
Answer to Question 13...and Glen Wright's comment

Nice comment. I agree completely with teh description you give about how external reform support can provoke change, incrementally, over time. The work in Albania has been ongoing for a good while now, and has been pretty consistent--a number of the same internal and external faces, building on small steps, some that have worked and some that have not worked, patiently, with trust-building, growing capacity and political appetite, etc. In this example one sees that reform does take time. But one also sees that reform happens over time if (and this is a bog if, given Glen's comment) the process of reform is consistent, truly incremental (building steop by step on past interventions), and external parties don't push governments to do stupid things in the short run. I have added some recommendations for improving external influence at the end of this email, as ways in which I see donors need to adjust their processes of reform to facilitate real reform.

You ask if I think the ground swell for change within the development monoliths will work. I think we all have reason to expect they won't. In my research I found serious people within narrow areas like PFM and in broader areas noting the need for less mimicry, less best practice focus, more localized engagement, etc. since 1992 (almost as soon as this work began). The voices of critique are far more influential and impressive than mine and others (think of Dani Rodrik, Amartya Sen, Evans, Chan, Amsden, Grindle, and many more). I even find World Bank commitments to change, almost every ten years. And little change.

There are obviously multiple reasons for this, and the link between aid and reform is one of them (disbursements need to be triggerd by visible, short term improvements, fostering what I call 'reform as signals'). I am attaching a working paper by Lant Pritchett, Michaal Woolcock and I on this topic. We believe taht the way donors do development is creating a 'Big Stuck' situation where governments are not being encouraged to do things taht make them more functional. In my experince, many folks in the development comunity read this kind of work and say 'yes you have a point but I am not part of that problem' or 'we can't really change the way things work'. I discuss the problem in the last chapter of my forthcoming book (you must know by now that I am spending lots of time on this book at the moment!):

The book’s last chapter acknowledges that the problems of poorly fitted institutional reform in development cannot be solved without change in the way organizations like the World Bank think about and do such reforms. It also recognizes that these organizations face constraints in adopting new ways of doing reform. Such entities have noted their poor records with reform in the past, even in high profile evaluations on the topic. They have introduced strategies to do things differently. These seldom alter the way business is done, however.  It is much like a carpenter acknowledging his pegs do not fit the holes but insisting that the process he follows is correct and does not require adjustment. “I don’t need to change,” he might say; “It would be great if all the pegs fit into all the holes, but give me some credit for what I have done. Look at all the nice pegs I have created, and remember that some of the holes were filled by the pegs. I am sure more pegs will fit more holes in time.” The chapter asks why organizations like the World Bank would take such a position on failure and routinely resist change. It identifies why development organizations and developing country governments keep producing institutional reforms with limits and ultimately presents some practical ideas to address these issues—changing the rules of development itself.

An excerpt on how I think external reform recommendations could be improved:

The process of finding and fitting institutional reform is similar to the iterative route carpenters might follow when crafting pegs for difficult-to-fill holes. They may look at different types of pegs for inspiration, considering some square oak pegs, round pine pegs, small pegs, rough pegs, smooth pegs, and so forth. Ultimately, their selection would draw on attributes of many examples—perhaps leading to a hybrid round, rough oak peg that fits the hole needing filling. This kind of hybrid is evident now in Rwanda and Indonesia, where decentralization and anti-corruption reforms have yielded new institutions that incorporate some aspects of external best practice with other aspects of internal tradition. Neither of the two examples is a pure form of any pre-existing best practice, and neither emerged by pursuing best practice. Rather, as has been shown, they came about as a result of incremental processes supported by external partners.

Experience in these cases yields recommendations for ways to improve the process and product of externally influenced institutional reforms. First, the content of these reforms should emphasize short term actions that facilitate long term change. Best practice reforms described in chapter four contrast markedly with this. They introduce long term solutions—like best practice international accounting standards—in short time periods, promising final products from four to six year projects. The idea behind incrementalism is different. Reforms should start with small first steps that lead to learning and next steps, which cumulatively provide a path to long term change. External short-term projects should focus on the short-term actions that facilitate this kind of gradual change, and emphasize as results the degree to which governments ‘find’ new ideas and build new capacities and political appetite for bigger next steps—fostering longer term, endogenous change.

This leads to a second recommendation, centered on the importance of external parties allowing and even cultivating ‘muddling through’. This can be done by fostering experimental interventions, crafting ongoing evaluation and learning opportunities in these interventions, and ensuring that lessons feed future experiments.  Again, such approach contrasts with dominant ways of doing institutional reform in development. The tendency is to focus projects on specific pre-designed interventions, with a linear route to implementation—given indicators used to determine whether project funds disburse, for instance. This approach focuses reformers on routine, mimetic, reproduction of best practice content and ultimately leads to the kind of limits discussed in chapter six; reforms are adopted in form but lack functionality.

Building on this idea, those designing external reforms should be careful not to focus interventions on reproducing what they see ex-ante as the ‘right rules’. Rather, they should help construct a problem focus for internal reformers, and provoke an awareness of contextual realities that will shape solutions. This means constructing issues as problems (discussed in chapter seven) and bringing local actors into processes of discussing and addressing these problems (discussed in the next chapter). This is the opposite of a ‘Cargo Cult’ approach to doing external reforms (discussed in chapter three) in that it focuses the process and product of change on local issues. As discussed, reforms need to be defined in this way if they are to be relevant—politically accepted and practically possible.

Another set of recommendations relates to the way in which external agents can help facilitate broad-based scanning and the translation of reform ideas. First, external agents should realize the important role they can play in providing external ideas into a menu of institutional reform options. They should note, however, that this role requires more than representing one-best-way models and narrow sets of ideas. The role requires providing broad sets of ideas, with details about where they emerged, when, under what conditions, and with what pre-requisites. The goal, as discussed, is to construct a host of menu items with the attached recipes, facilitating choice. Simultaneously, external agents should develop ways in which internal ideas can also enter the menu. This requires seeking out internal agents with ideas, working with these agents to flesh such ideas out, and making it clear that these ideas have equal standing (and even more standing) with external ideas.

Finally, external agents should facilitate processes where internal reformers can experiment with combinations of the ideas they have in front of them. This is a core part of the ‘muddling through’ process and involves creating opportunities, through externally supported grants, projects and the like, for action-based interventions, with organized evaluation, learning and feedback mechanisms. External agencies pursuing such approach should be open to the creative, locally relevant but potentially strange looking hybrids that emerge from such processes. The goal should always be to facilitate the process itself, more functional institutions, and solved problems. 
« Last Edit: January 25, 2012, 08:39:00 GMT by Napodano »


  • *
  • Posts: 154

An important set of factors that are worth considering in this context is the decision-making framework that applies to an issue in the particular country. An illustration is from my experience in Indonesia. There was an issue that I feel I can discuss now as the Indonesian players are no longer in this life: I was working with the bright young Indonesian on improving the supply of housing. A/the major barrier was the decision that the President had made to keep interest rates for mortgages very low for low income borrowers (I won't give the actual figure), effectively keeping the private sector out of the low income market. However, there existed a substantial market and supply mechanism if the government would make land available (which they were willing to do) if private developers would build low income housing, and the private developers were more than willing to build the housing if they could make the loans at an interest rate that was higher. My young Indonesian counterpart/co-conspirator and I developed a model that used different interest rates to project different levels of housing expansion and affordability. This model indicated that a 50% increase in interest rates would generate a more than doubling of annual housing supply to relatively low-income families. My counterpart used his connections to have this analysis presented to the President, and in a week the new interest rate ceiling was announced, and the resulting program exceeded even our highest projections. On the issues of decentralization, on which I worked for years (almost a decade) in the same collaborative way, never made significant progress during my time and years after, despite that almost all the Indonesian technocrats and external agencies came to support it. Years later, the crises in certain (3 key) provinces pushed the resolution over the hill. All the work by many people over many years that had not seemed to have been effective, made it possible for crisis solutions to be rapidly deployed.



  • *****
  • Posts: 678
PFM Boarders,

the conversation is now closed. I leave to Matt and Martin the closing remarks.

I take this opportunity to personally thank Matt for his time and kindness in sharing excerpts of his upcoming book, exclusively for .


Martin Johnson

  • *
  • Posts: 76
Re: Conversation with Matt Andrews on limits to externally influenced PFM reform
« Reply #34 on: February 08, 2012, 18:09:28 GMT »

When you agreed to come to the fireside for this chat, we were confident that we would be in for a fascinating discussion. We had not quite anticipated being treated to such depth and to such effort on your part.  I think it is safe to say that we are all looking forward to publication of your forthcoming book. In the meantime, the extracts you have shared and the arguments you have presented provide the many external agents among us with much food for thought as to how we can become more effective in our work in facilitating change (and perhaps even in choosing the right trees to bark up!).  On behalf of all PFMBoarders and all visitors to this debate, thank you very much indeed.

I have attached a transcript of the discussion for download.



RSS | Mobile

© 2002-2023
Powered by SMF