When one examines these reforms more closely, however, the picture is less impressive. While Afghanistan's public financial management system may look like one in a middle-income country, donors still express reservations about whether funds are actually spent where the budget says they are. Allegations of corruption are rife, and financial controls that work in the central ministry of finance buildings in Kabul are still not operational in many sectoral ministries (especially those located outside the capital) where most spending remains outside of the national budget. The gap between appearance and reality becomes evident upon examination of the PEFA data used to assess the system, which shows that laws and budgets on the books bear little relationship to their implementation. Similarly, while Uganda's anticorruption laws may look perfect, that's hardly how they function in practice. Corruption is commonplace, and a string of high-profile corruption cases have caused donors to withdraw funding from the country as recently as November 2012. This is reflected in the data from Global Integrity, which scores the country's anti-corruption laws at a perfect 100 but rates implementation of laws at only 51, leaving a "huge" implementation gap of 49.
Such gaps between what institutions look like and how they're put into practice are commonplace in developing countries. They reveal the limits of institutional reforms in development and suggest the need to correct the way these reforms are currently done.
My research shows, for instance, that over 80 percent of developing countries experience major implementation constraints after years of public financial management reform. Countries make budgets better after reforms but commonly fail to execute the budgets as they are written. They draw up best-in-class procurement regulations that are seldom implemented, and adopt internationally accepted accounting standards required by donors -- but with little hope of actually doing their accounting in accordance. What's more, studies suggest that there is little reason to believe this gap between form and function closes over time. In fact, the limited existing evidence shows that the gap widens with the years, as reforms help countries craft ever-better laws and formal procedures instead of focusing on better implementing the laws and procedures that already exist.
When looking at Global Integrity data, for instance, it appears that the average gap between the quality of anti-corruption laws and their actual implementation widened for developing countries between 2008 and 2011, from 32 to 35, regardless of institutional reforms. (I'd argue, in fact, that the reforms may have exacerbated the problem.) This gap is more than twice that in richer countries, where corruption is less of a problem. My research shows that blind confidence in the "best practice" institutional approach often leads reformers to devise plans without any real attention to the context. External reform planners promote internationally "appropriate" institutions but neglect to ask if these new rules of the game can actually be implemented. Developing countries are being pressured to introduce international accounting standards, for instance, even when there are no internationally accredited accountants in the country, the idea of an accounting "profession" is not established, and businesses do not accept the high levels of disclosure demanded by these new standards. Officials in the developing countries have every incentive to act as champions of reforms focused on introducing such poorly fitted institutions to garner external support (like funding from a donor). But they either never intend to fully implement the new institutions (knowing from the outset that they won't work), or soon run into problems when they attempt to carry them out. In this kind of game -- where incentives are tied to measuring what the funding institution can tangibly evaluate -- it should not be surprising that the real and often hidden rules of the game remain unchanged and dysfunctional even after reforms are completed. What you see is not what you get after reforms.
Not all efforts at institutional development have failed, however. Consider, for instance, Rwanda's tradition-based Imihigo initiative to increase the delivery of government services. This reform emerged over the past fifteen years in order to dramatically improve provision of basic services in districts where deficient services had led to dissatisfaction and violence in the 1990s. Initiative to take action began after 1995 through the engagement of many parties, including (though not dominated by) international organizations. In contrast to many institutional reform experiences, these parties did not start with a script for "institutional reform." Instead, they tackled a specific problem: The lack of service delivery in local governments, which had fed tensions that led to the country's genocide.
Spurred by common concern over this problem, groups of officials began meeting in the late 1990s. Those involved included the president, the minister of local government, other civil servants, and international development specialists. They built on lessons from prior (failed) efforts to establish local governments and, in 1998, started experimenting. Using a small World Bank loan of $5 million, they financed different service delivery initiatives in a variety of districts. The results and lessons of these experiments inspired a second wave of experiments in the early 2000s, which led to a rolling process of policy formulation and decentralization (supported by a range of external organizations and donors). Every few years the government assessed the lessons learned in this process, adjusted what it was doing, and took another step forward.
Progress made from Rwanda's problem-driven and history-minded approach became evident by the mid-2000s. Reforms matured to a point where district administrations were well established, financial flows had been structured to fit Rwandan realities, and districts were starting to lead service provision. But there was no way to monitor performance or create incentives for the districts to pursue high levels of performance. The reform group started looking for ways to address this specific problem. They eschewed many "best practice" modern models of performance management as poorly fitted to their context, and looked instead to a forgotten Rwandan tradition called Imihigo. According to this tradition, monarchs used a simple reporting and rewards process to ensure that regional appointees were managing their regions well. These appointees gathered together at various points of the year and had to report, in a group, on what they had done. If the monarch deemed their performance acceptable, they could drink from a common calabash and were publicly recognized. If not, they were publicly passed over.
A version of this was adopted on an experimental basis in 2006. It involved mayors reporting in a public space to the president or prime minister; and providing photographs, testimonies, and other visible evidence of performance. Early lessons were captured, and the Imihigo process was improved and diffused throughout the country, at first on the district level, and more recently on the national government level. It is a peculiar-looking Rwandan performance management hybrid that does not fit any best practice script, but that addresses specific problems and enhances functionality.
The Imihigo example points to various principles common to many institutionally-focused reforms that do yield more functional governments. First, they're problem-driven, meaning that they're not informed by pressures to adopt a script of the "right" rules. Second, they emerge through iterative, trial-and-error processes of experimentation and learning (with tight feedback loops) that allow adjustment -- rather than rigidly programmed projects that lock a "best practice" reform design into place for five to ten years. Third, they involve groups of leaders distributed across the government, not singular champions sitting atop a closed hierarchy.
My colleagues Lant Pritchett, Michael Woolcock, and I merge these principles into a new way of thinking about doing institutional reform in development, called "problem-driven iterative adaption" (PDIA). PDIA embodies our belief that context has to be taken into account if we wish to promote institutional reform and good governance in developing countries. It eschews the idea of generic "right" rules, and instead promotes flexible processes that allow countries to find the kind of hybrids that are accepted and implementable in their particular contexts. These kinds of institutions actually work (by influencing and shaping behavior), and are vital for development.
Note: The posts on the IMF PFM Blog should not be reported as representing the views of the IMF. The views expressed are those of the authors and do not necessarily represent those of the IMF or IMF policy.
