Kunduz, Afghanistan

Special Report: Changing the rules of the game

Rebuilding fragile states

Frauke de Weijer | October 10, 2011

Development organizations have yet to come to terms with the inherent complexity of institutional change. Institutional change takes time, and the kind of institution best suited to a given situation depends on the context. In other words, a successful institution in the West is not necessarily going to work in Afghanistan or Sudan. Institutions understandably tend to mimic other successful organizational structures, but this often only creates the illusion of capability and legitimacy. Development organizations therefore need to build a deep understanding of the rules systems at work in the society in question and acknowledge the unpredictability of change in the complex social systems of fragile states. Only then can they adapt their practices accordingly and help build institutions that work.

Foreign interventions in fragile states – whether aid donors or occupying armies – often try to create new and ‘modern’ state structures, mainly copies of Western institutions. But the assumption that good governance, democracy or indeed development can be imposed from the outside is increasingly being challenged. They can only succeed if rooted in local political, social and economic processes.

Building modern state institutions means having the right policies implemented by effective economic, political, judiciary and governance organizations. However, recent research by Lant Pritchett and Frauke de Weijer shows that creating effective institutions takes much longer than expected. 1 Their 2010 background paper to the World Development Report 2011, entitled Fragile States: A Capability Trap? argues that at the average rate of improvement in bureaucratic quality, a typical fragile state would take 116 years just to get to the level of a country like Kenya. 

Governments of countries such as Afghanistan, Haiti and South Sudan are not only overly optimistic about the possible pace of change, but they are also increasingly expected to perform an unrealistic range of functions. All in all, this places very high demands on the degree of institutional change that has to occur in a short time. Development programmes have therefore spent tremendous energy and resources on policy reform and capacity building, with the aim of facilitating institutional change. Unfortunately the results have been disappointing. 

Recent insights into the dynamics of change in complex social systems show the inherent unpredictability and uncertainty surrounding change. The international community has yet to come to terms with this and adjust their approaches accordingly. International development organizations still operate on the premise that once an institution has been successfully established in one place, it will perform equally successfully elsewhere. They also believe that change is manageable from the outside. This leads to perverse incentives that may even reduce institutional effectiveness by creating fragile institutions that lack the necessary robustness.

There are other reasons why attempts to introduce institutional change from the outside have been so disappointing. Institutions are deeply rooted in social contexts, and new institutions have to operate in social contexts where different rules apply. These rules affect how these new institutions function. Patronage, for instance, does not immediately disappear when new recruitment procedures are adopted. Rather, people find ways of working around these rules, and in the end their actual behaviour barely changes.

New approaches need to take into consideration the inherent unpredictability and the underlying values, norms and behaviours that shape people’s responses. This requires a different mindset that creates the conditions for contextualized solutions. This new mindset has to recognize the value of local knowledge and problem-solving capabilities and use them as a starting point. It has to stress variety and a willingness to explore. But it also has to acknowledge that local knowledge and capabilities must be placed in a wider strategic framework that will give it direction.

When rules clash

Development can be defined as a historical process of economic, political, administrative and organizational, and social transformations. The essence of each transformation is a shift in the overall ‘rules systems’ – the established patterns, norms of behaviour and expectations – in which individuals operate. Development therefore necessarily implies a change in social norms and behaviours, and the transition from one system of rules to another (see Figure 1). 2

Figure 1. Development as a fourfold process

Formal rules systems are often only partly developed in many developing countries. There is a profusion of different kinds of systems, formal and informal, sometimes competing for resources, power and legitimacy. 3 During times of transition, citizens are faced with multiple, potentially conflicting rules systems. This creates stresses in society, which can easily lead to conflict and frustrate development efforts. 4 

Dani Rodrik, professor of international political economy at Harvard University, starts his executive education courses with three traffic videos that illustrate this point. One shows highly organized streets with traffic lights and clear lanes and a free flow of traffic. The second, in Hanoi, is highly informal with pedestrians, donkeys, cars and tracks miraculously and efficiently meandering their way across a roundabout. The third video shows the most dangerous intersection in St. Petersburg, where only half the people abide by the traffic lights. This mixture of formal and informal rules has disastrous effects, with accidents occurring regularly.  5

An example of a rules system often associated with development interventions is the value of individual rights and self-determination. Promoting this rules system can have a St. Petersburg effect, easily leading to tensions in a society where social structures are primarily meant to support the collective well-being of a group as a whole. Educated girls in conservative Muslim countries often find themselves struggling with their new-found sense of self and the role their families expect them to fulfil. Two value systems compete for legitimacy, and the long-term outcome is difficult to predict.

An example of a rapid transition from one rules system to another that clearly failed is the ‘shock therapy’ applied in former Soviet republics. The policy of privatization that was meant to increase competition led to more cronyism and monopoly. Because the rules systems among the powerful had not changed, the new institutions became ‘contaminated’ by the old way of doing things.

Though important lessons have been learned from failures such as these, the mindset behind most policy decisions has yet to change. The presumption that institutional change can be planned, directed and managed according to a predetermined plan needs to be re-evaluated. International development and foreign policy strategists have to realize that institutional change in complex social and political settings needs to be managed more dynamically. 6

Support for this recognition comes from unexpected quarters. Scientists from fields as diverse as physics, evolutionary biology, meteorology, organizational behaviour and many others have started to recognize this unpredictability and the non-linear attributes of change in complex systems. 7

Resistance to change

vario images GmbH & Co.KG / Alamy

Training national police, Kunduz, Afghanistan

Implementing institutional change without a deeper understanding of the rules systems at its core is likely to fail. Indeed, resistance can be expected as most reforms aim to change the rules of the game. Think of empowering the poor, protecting minorities, stressing individual rights or adopting meritocratic principles. Political economy analyses are increasingly taking into account active resistance to change by current power holders based on their self-interest. However, these analyses overlook opposition that stems from alternative rules systems. Self-righteousness is more inert to change than self-interest.

Take patronage, for instance. Loyalty systems based on patronage are strong forces that are deeply embedded in social fabric and are not easily severed. Indeed, they are likely to undermine attempts at applying meritocratic principles in organizations. These loyalty systems may stand in the way of building transparent, accountable organizations – or governments, for that matter – but at the same time they may serve, or may at a certain point in time have served, an important function in society.

Thomas Barfield, professor of anthropology at Boston University, elegantly shows in his 2010 book, Afghanistan: A Cultural and Political History, that it is exactly these social structures and kinship systems that have made Afghan society highly resilient after 30 years of war and the absence of a functioning state. Regardless of our moral attitude to patronage, it is a strong force that cannot simply be wished away. Unfortunately, this wishing away is the attitude that prevails in international development.

In light of these realities of institutional change, it would appear naive to assume that the mere introduction of a new set of rules and procedures in an organization would immediately replace the old ways and make it start acting according to the new rules. However, much development discourse is still based on the notion that the adoption of a certain institutional form, one that works well in industrialized countries, will automatically lead to the same outcome elsewhere. This reasoning ignores the fact that the development of these institutions was the output of a complex and long struggle within a particular society. 8 Institutions are an output, not an input.

Unfortunately, this reasoning may ultimately lead to a reduction in state capacity – the opposite of what was supposed to be achieved. Such a dynamic may be explained through the notions of premature overload, isomorphic mimicry and capability traps. Current development discourse, with its emphasis on good governance and effective institutions, overloads governments with demands and expectations that are simply not realistic; a case of premature overload. This may cause a reaction of isomorphic mimicry (see box) that can lead to capability traps (see Figure 2).

Made to mimic

Organizations often purposely adopt a strategy of isomorphic mimicry. They need legitimacy (and the associated financial resources) to survive and therefore need to ensure that they perform according to the standards placed on them. 9 If this legitimacy is based on form and not function, then they may respond by creating the illusion of being capable organizations. In such cases, they adopt the outward form of a capable organization, with little regard for how well their organization actually functions.

Isomorphic mimicry

Understanding how isomorphic mimicry works in relation to organizations can be explained by way of analogy. Lant Pritchett, professor of the practice of international development at the Harvard Kennedy School, uses bugs as an example in his 2010 presentation Isomorphic Mimicry: What and How? 10 'Bugs avoid being eaten by birds by developing glands that secrete poison and signal that they are poisonous. Once established, other bugs may simply develop the signal, [but] not the poison.'

In a background paper to the World Development Report 2001 published in late 2010, Fragile States: Stuck in a Capability Trap, Pritchett and De Weijer explain how this bug analogy can be used with organizations. 'It is much easier to create an organization that looks like a police force—with all the de jure forms organizational charts, ranks, uniforms, buildings, weapons—than it is to create an organization with the de facto function of enforcing the law. The danger of isomorphic mimicry is that it creates a powerful dynamic in which what survive are not functional organizations and institutions but mimics, which can adopt the camouflage of capable organizations without any of the associated drive for performance.' 11

A schooling system whose graduating students barely meet basic levels of learning – unfortunately a common situation in many developing countries – is one example. The education system seems intact on the outside. There is a school building, books are present and there is teaching. But the school’s children leave barely knowing how to read and write. Functionality is not measured, nor are there consequences for low performance.

Similarly, many developing countries have adopted wholesale the form of many bureaucratic institutions that are expected to perform a broad range of state functions, with little regard for their actual performance. The de facto functionality of an organization or system (how it functions in practice) is often not measured, and the de jure, or legal, form of the institution is deemed sufficient. It is good to remember in this context that it is only in very recent years that results-based performance measurement has gained ground in the West.

One would perhaps assume that development assistance organizations have the necessary performance indicators in place to evaluate their de facto functioning. Unfortunately, reality tells a different story. Real outcomes are not always easy to measure, so donor agencies tend to limit themselves to measuring outputs such as policy recommendations, strategic planning workshops and public consultations.

Whether any of these activities really affects how the system functions remains largely unknown. As a consequence, organizations often lack the drive to improve their performance in real terms, as improved functionality is not explicitly recognized. Accountability flows upward, not downward, in a situation with donor funding. The external legitimacy of the organization is based on whether the donor boxes have been ticked, not whether genuine performance needs are met. 

Isomorphic mimicry need not always be a bad thing. It may well perform a function, such as setting an example and creating exposure to a different way of doing things. It may create an institutional island that presents a window to an alternate possible reality. Moreover, a certain degree of divergence between what an organization looks like from the outside and how it functions on the inside is normal. Even in the most effective organization, there is a difference between what it claims to do and what it actually does (the espoused values and the values-in-use). There is a difference between official procedures and what actually happens on the ground. This acts like the grease between the wheels that makes a machine run smoothly. 12 

Perfect storm

In many developing countries, however, and fragile states in particular, this phenomenon is taken to the extreme. The outside appearance of the organization and the internal operations become highly divergent, and two parallel, detached universes begin to co-exist. Policy changes made on paper no longer have any bearing on the lower tiers of the organization, which continue to play by their own rules.

This may lead to a loss of institutional integrity, 13 whereby the outward de jure performance is no longer connected with the de facto performance. This leads to consistent implementation failure, whereby policy reforms made on paper lead to little change in outcomes on the ground. Take public sector reform, for instance. Advisors recommend changes to the organizational structure, retrenchment of a proportion of the civil service and the adoption of new meritocratic recruitment processes. Most of these recommendations stay on the shelf and never get implemented. Or they get implemented strategically, serving the purpose of the human resources officer in charge who uses the new procedures to keep unwanted candidates out, but recruits his own through the back door.

The perfect storm of premature overload, isomorphic mimicry and loss of institutional integrity can lead to a capability trap. 14 The de jure and the de facto become disconnected from each other, and the organization gets stuck in a low-performance equilibrium with external and internal pressures conspiring to maintain the status quo. This is what may happen when pushing too fast for reforms that are incongruent with prevailing rules systems. Exactly the opposite of what is trying to be achieved.

Figure 2. Excessively rapid pace of de jure reform creates a widening gap of de jure and de facto, which can further worsen de facto capability.

Building effective states

These dynamics of institutional change, and the potentially perverse effects of outside intervention compel us to reflect on our strategies for building state capacity. It has implications for good governance, capacity building and the way development programmes are designed.

A scarce resource

Modern-day expectations of what the state must be able to deliver are high and growing, and these expectations are extended to fragile states via the ‘good governance’ agenda of donor agencies. Viewed through the lens of rules systems, a state that is performing well needs to be more than just a well-functioning state apparatus, as the ability of the state to perform is closely related to how well it fits with the social context in which it is embedded.

State capacity should therefore be defined as the ability of the state system to adapt its rules systems to changing circumstances and opportunities. And it should do so in a way that leads to positive economic, political, social and administrative transformations. Pushing too hard and too fast can actually reduce state effectiveness. The conclusion is that although one cannot overstate the importance of strong institutions, the road map for how to get there should be subject to a lot more debate.

A road map for building effective institutions needs to be guided by the likelihood that reforms will succeed. This needs to go far beyond the current frameworks of political economy analysis that focus mostly on identifying and overcoming resistance to reforms that are considered high priority by the mainstream development community. It needs to treat state capacity as a scarce resource – one that is fragile and needs to be nurtured.

Most governments in fragile states lack enforcement capacity. This could be because the state is inherently weak or because there are international conventions that restrict the use of force. In practice this means that for a government to successfully implement reforms, it needs to be supported by a sufficiently strong coalition.

Such a coalition can only be built if the proposed reforms have a certain degree of buy-in in the society, which means they must have a firm base in an existing rules system. They must be perceived as legitimate by the citizens for them to allow the state to impose its norms and rules when these do not overlap with existing rules systems. In addition, the government apparatus itself must embody these norms, which is often quite a challenge.

Take Afghanistan as an example. Does the state have the perceived legitimacy to station a police force in rural villages? Does the police have the legitimacy to conduct house raids? And what if the police force is known to be corrupt? Is advocating a strong national police force still such a good idea if the answer to this question is no, at least in large swathes of the country? It is important to have clearer expectations of what a government can realistically be expected to do.

It is time to move past ideological debates on the size and scope of government, and think pragmatically about what reforms are most pressing, what can be achieved realistically, in which order, time frame and at what pace. 15 Small successes may need to be used more strategically to gain the credibility, legitimacy and trust necessary to move on to the next, perhaps more contentious, step on the reform ladder. Perhaps police should only be given broader powers once they actually start to improve people's security.

Paper tigers

The dangers of premature overload, isomorphic mimicry and the loss of institutional integrity need to be considered when designing a strategy for capacity building. 16 In practice, however, capacity building programmes in government ministries in developing countries and fragile states tend to ignore these factors.

Fabrizio Bensch / Reuters

An Afghan national policeman during a drill, Kunduz, Afghanistan

A major risk is the disconnect between policy makers and the reality on the ground. Local knowledge and capabilities are the building blocks of genuine capacity building. But more often than not, the international consultants and national staff leading capacity-building programmes make no connection with people’s realities. Their formal policies barely reach the ground, and their development jargon falls flat with people used to local rules and value systems.

The opposite is also true. Local problem-solving capabilities and endogenous solutions rarely reach the top levels of policy making. And when they do, these solutions are generally discarded as non-strategic, too idiosyncratic, too steeped in local culture and not in keeping with the high standards of the donor agencies. These may be understandable responses, but they make home-grown, contextualized policy solutions that are an effective blend of international experience and local knowledge less likely to occur.

Another danger of the loss of institutional integrity is the weak organization it creates, a ‘paper tiger’. 17 Capacity is only present in a few individuals and does not filter through the organization. It is a thin layer of shiny varnish that can be easily scratched off. These organizations collapse quickly when the leadership is removed or they succumb to political pressures.

The case of the Ministry of Rural Rehabilitation and Development (MRRD) in Afghanistan is a good example. This ministry rose to prominence quickly after 2002, under the visionary guidance of its minister, Haneef Atmar. It became the poster child for rapid organizational change and state capacity building. Afghanistan’s flagship programme, the National Solidarity Programme, was executed under the ministry's auspices. It began to make real progress in strengthening the links between communities, provincial departments and the central government.

The ministry consolidated its position under the succeeding minister and started to show signs of robustness. Capacity had started to trickle down through the entire organization, and the different tiers had started to work together towards the same goals. Nonetheless, this way of operating had not yet acquired sufficient momentum to be able to operate without a visionary leader. This became clear with the appointment of the current minister, who does not share the same working method that MRRD had now acquired. Meanwhile, MRRD is starting to crumble. After nine years of solid effort, MRRD is still a paper tiger.

There are no easy answers here, let that be clear. The pressure to produce immediate results is high, while managerial and technical capacities are dangerously low in many governmental organizations. These are genuine constraints that may push organizations towards isomorphism. However, international development organizations need to accept partial responsibility for this reality, and limit the extent to which they are feeding into these patterns.

Programmes funded by large bilateral or multilateral donors are most guilty, but even small participatory-minded NGOs fall into the same trap. As much as they try to use evaluation criteria that are meaningful to the people they are trying to serve – which is a major step in the right direction – they are also unable to escape the rigid monitoring and evaluation frameworks that they are subject to.

Home-grown solutions

Assumptions about the degree of institutional change a society can absorb at a given time tend to be overly optimistic, so the effects of policy reform are therefore less predictable than expected. Importing institutional models may cause resistance, and the development community's working method may undermine positive change rather than support it. However, this does not mean that external knowledge and imported institutions cannot play a role in development.

In this context it is useful to make a distinction between adaptive problems and technical problems (see box), a distinction made by Ronald Heifetz, senior lecturer in public leadership at the Harvard Kennedy School.

Technical and adaptive problems

Technical problems are technical in the sense that the necessary knowledge about them is known, has been digested and institutionalized in a set of organizational procedures. For example, influencing interest rates through monetary policy is a technical problem, whereas ways of increasing risk-taking and entrepreneurial activity are adaptive ones. Solving technical problems does not lead to serious losses for those involved, and values and behavioural patterns do not need to change in any significant way.

Adaptive problems, by contrast, are likely to cause losses for at least some of the stakeholders. Adaptive problems are deeply embedded in the complexities of the social system, where change is unpredictable and outcomes uncertain. Solutions are not clear-cut, and people’s opinions on adaptive policy issues tend to strongly differ. The answers cannot be known beforehand, and the effects of each intervention cannot be adequately forecast.

Many problems in the developing world have a highly adaptive character. Most institutional change initiatives in fragile states address adaptive problems, as they are by definition aiming to change the rules of the game. They are therefore likely to cause serious friction between different rules systems, making change much more unpredictable. Imported solutions may work, but they may also cause serious resistance in the social system.

Based on Heifetz, R. (1994) Leadership without Easy Answers. The Belknap Press of Harvard University Press

The current development models of centralized planning and imposing reforms from the top down work well for technical problems, where change is likely to be linear and predictable. But these models are not particularly well equipped to deal with adaptive problems, as they require more humility and an acknowledgement that the answers cannot be known beforehand. One of the biggest mistakes made in development is to treat adaptive problems as technical problems.

Change management and organizational learning literature 18 identify three approaches for tackling adaptive problems:

  • Collaborative problem solving
  • Cross-boundary cooperation
  • Encouraging local experimentation and innovation

What these three approaches have in common is an emphasis on learning and the constant re-evaluation of the strategies and the assumptions underlying them. People have to work together to understand the norms, values and behaviour – the rules systems - impacting the problem at hand. They jointly devise strategies that have local support, build on local capacities and solve local problems. This process often requires changes in values and belief systems, as bridges between these rules systems will need to be built.

Many non-governmental organizations (NGOs) have realized the power of home-grown solutions for decades, which has given rise to a wealth of participatory and community-based methods to identify local priorities and solutions. Unfortunately few of these local innovative practices have been scaled up or institutionalized. NGOs document the lessons learnt, but mostly circulate them in report form, decontextualized and transformed into a new blueprint of best practices. Opportunities for diffusing such home-grown solutions and consolidating local innovations are lacking. Institutional connections between these experiments and centres of higher-level policy making are virtually non-existent.

The way forward 19

Picture Contact BV / Alamy

Kunduz Province, Afghanistan, February 2011

Looking at development and state building as complex processes of transformation has major implications for the way development programmes are designed. They can no longer be designed on the basis of what worked elsewhere. The impetus of the new development model will encourage the active search for localized solutions, rooted in realities on the ground. The emphasis will lie on encouraging local autonomy for actors within the social systems to develop contextualized solutions, and initiate local institutional change.

The challenge then lies in blending international experience and technical capacity with domestic capacities and insights into local realities. This would make it possible to harness local energy and put it to use. Operationally this means creating a learning infrastructure that can feed a process of continuous learning, adaptation and fine-tuning of emerging policy solutions.

One of the risks with this approach is that existing institutional structures tend to be relatively resistant to change, and often deeply characterized by isomorphic mimicry. Bureaucracies in many fragile countries are particularly prone to these risks, because of a lack of exposure to more modern management systems.

State-building efforts, supported by the international community, tend to be relatively state-centric. While a capable public sector is essential, public sector reform and capacity building is not likely to lead to adaptive policy solutions. Innovative capacity, and the willingness to devise contextual solutions can equally – if not more than equally – be found in the periphery.

This is perhaps where donors can best leverage their influence. They can ensure that policy solutions are not based on external best practice but work on the basis of what is already there and build on existing capabilities and innovative power. These solutions must encourage variety in solution-seeking and create an enabling environment that generates innovative solutions to adaptive problems and actively promotes solutions that work and discourage those that do not.

A new accountability

Such a new development model will require a different accountability framework, which will exert the right pressure on local systems to perform, and pull them out of a state of isomorphic mimicry.

Whenever possible, clear quantifiable outcome measures need to become the cornerstone of the accountability framework, not outputs. It is self-evident that any development programme will need to plan its activities, inputs and outputs in light of its targets. These programmes cannot simply set to work with an open-ended approach. However, in current practice, these targets quickly become carved in stone, which precludes adaptation to changing conditions and blocks new insights into the dynamics of the social system from emerging.

The proposed new development model does not consider inputs, outputs and activities as targets, but simply as a best guess made at a given point in time. Inputs and outputs will still be tracked, but not used as the primary basis of accountability. Development programmes need to use a structured approach that regularly re-evaluates their strategies and adjusts their planned inputs and outputs accordingly.

In addition to quantitative outcome measures, an accountability framework must broaden its scope to encompass unexpected outcomes. These will probably occur in situations where change is likely to be non-linear and unpredictable. Monitoring and evaluation techniques must scan the system widely, and pro-actively detect unexpected surprises.

Lastly, based on the argument that development is essentially about institutional change, we also need to think about how we can gauge behavioural change, for which a number of innovative methodologies have been developed. 20

These three types of data – quantitative outcome measures, unexpected outcomes and behavioural change – are all inputs into the learning infrastructure. A structured process of continuous assessment of progress, a re-evaluation of underlying assumptions, and ongoing adaptation to the realities on the ground is essential to maintain the flexibility required in an unpredictable context. Ultimately, the programme must not merely be held accountable for how well it meets outcome measures. It should also be held accountable for how well it interprets the data and whether it responds accordingly – in other words how well does it learn?

Using these mechanisms, policy or programme design becomes less rigid and more adaptable to the unexpected and the uncertain, and therefore more strongly grounded in reality. Flexibility will become the name of the game. Not as a wild card or an open invitation for corruption, but in a planned, structured and transparent manner.

Out of the comfort zone

Development organizations have yet to come to terms with the inherent complexity of institutional change, and adapt their practices accordingly. This requires a change of mindset that moves away from the notion that imported institutional models will automatically function in different social contexts, and towards more flexible and open-ended approaches. At the foundation of this alternative way of thinking is an increased humility and recognition that many answers cannot be known beforehand, and that uncertainty is the name of the game. Development cannot be planned in advance, and in particular not from an ivory tower in one of the OECD capitals.

This may seem like a daunting proposition as it takes us out of our comfort zone – namely thinking that we are the experts and have the answers, and that all we need to do is implement them. This so-called certainty, which in all likelihood has been a false certainty in the first place, has to be left behind.

Indeed, we have to recognize that we have been setting ourselves up for failure all along. Perhaps we can slowly replace this discomfort with a renewed sense of optimism about the opportunities this way of thinking opens up to us. We should let ourselves be surprised by the force of variety and the unexpected. Perhaps we can become inspired again, by working with the grain rather than against it.


Earl, S., Carden, F., and Smutylo, T. (2001) Outcome Mapping: Building Learning and Reflection into Development Programs. International Development Research Centre.

Gamble, J. (2008) Developmental Evaluation: A Primer. The J.W. McConnell Family Foundation.

Pritchett, L., and de Weijer, F. (2010) Fragile States: Stuck in a Capability Trap? Background Paper to the World Development Report 2011. World Bank.

Roduner, D., and Schlappi, W. (2008) Logical Framework Approach and Outcome Mapping: A Constructive Attempt of Synthesis. Agridea.

Photo credit main picture: Kai Pfaffenbach / Reuters


  • 1.

    At the average rate of improvement in bureaucratic quality a typical fragile state would take 116 years just to get to the level of a country like Kenya. See Pritchett, L., and de Weijer, F. (2010) Fragile States: Stuck in a Capability Trap? Background Paper to the World Development Report 2011. World Bank.

  • 2.

    See Jim Woodhill's 2008 article 'Shaping Behavior' in issue 10 of The Broker.

  • 3.

    See Morgan, P. (2005) The Idea and Practice of Systems Thinking and Their Relevance for Capacity Development. European Centre for Development Policy Management.

  • 4.

    See also Rodrik, D. (2008) Second-best institutions. American Economic Review 98:2, pp. 100-104.

  • 5.

    See Dan Rodrik's weblog for examples from Hanoi and St. Petersburg.

  • 6.

    These notions of uncertainty, unpredictability and the non-linearity of change are the main elements of complexity theory, a way of thinking increasingly applied to complex social problems in the world. It is increasingly recognized that reductionist, linear cause-and-effect analytical models are not well suited for these types of problems.

    For an interesting overview, see Alan Fowler's 2008 article ‘Connecting the Dots’ in issue 7 of The Broker, as well as other sources such as Axelrod, R. and Cohen, M. D. (2000) Harnessing Complexity. Basic Books; Cilliers, P. (1998) Complexity and Postmodernism. Routledge; Mowles, C., Stacey, R., and Griffin, D. (2008) What contributions can insights from the complexity sciences make to the theory and practice of development management? Journal of International Development 20:6, pp. 804-820; and Ramalingam, B., Jones, H., Reba, T. and Young, J. (2008) Exploring the Science of Complexity: Ideas and Implications for Development and Humanitarian Efforts. Overseas Development Institute.

  • 7.

    See the 'Foreign policy as a complex system' blog, The Broker.

  • 8.

    On the importance of 'good struggles' for political and legal reform, see, for instance, Adler, D., Sage, C., and Woolcock, M. (2009) Interim Institutions and the Development Process: Opening Spaces for Reform in Cambodia and Indonesia. Working Paper No. 86, Brooks World Poverty Institute, University of Manchester; and North, D. C., Wallis, J. J. and Weingast, B. R. (2009) Violence and Social Orders. Cambridge University Press.

  • 9.


  • 10.

    See presentation Isomorphic Mimicry: What and How?, Lant Pritchett, 2010.

  • 11.

    See the introduction in Pritchett, L., and de Weijer, F. (2010).

  • 12.

    If procedures and rules were followed to the letter, the work would actually slow down, a principle utilized in work-to-rule actions.

  • 13.


  • 14.


  • 15.

    This line of thinking builds on work on 'good enough governance', see Grindle, M. (2004) Good enough governance: Poverty reduction and reform in developing countries. Governance 17:4, pp. 525–548; and on work on 'just enough governance', see Levy, D., and Fukuyama, F. (2010) Development Strategies: Integrating Governance and Growth. World Bank Policy Research Working Series No. 5196.

  • 16.

    For more on capacity building, check out the 'Capacity on the brink of maturity?' blog on The Broker's website.

  • 17.


  • 18.

    There is a wealth of literature on ‘social messes’. See, for example, Ackoff, R. L. (1994) Systems thinking and thinking systems. System Dynamics Review 10:2-3, pp. 175-188 on ‘wicked problems’; Rittel, H. (1973) Dilemmas in a general theory of planning. Policy Sciences 4, pp. 155-169, Conklin, J. (2006) Wicked Problems and Social Complexity. In J. Conklin, Dialogue Mapping: Building Shared Understanding of Wicked Problems. Wiley, and Heifetz, R. (1994) Leadership without Easy Answers. Belknap Press of Harvard University Press, on ‘adaptive problems’. These works all describe complex public policy problems involving non-linearity and unpredictability. Systems thinking has also been applied to organizational learning and change management, notably by Argyris, C. (1995) Action science and organizational learning. Journal of Managerial Psychology 10:6, pp. 20-27; Senge, P. (1990) The Fifth Discipline : The Art and Practice of the Learning Organization. Doubleday/Currency; Conklin, J. (2006) Wicked Problems and Social Complexity. In J. Conklin, Dialogue Mapping: Building Shared Understanding of Wicked Problems. Wiley; and Marquardt, M. (2000) Action learning and leadership. The Learning Organization 7:5, pp. 233-241.

  • 19.

    This section is based on the idea of ‘the middle way’ as first presented in Pritchett, L. and de Weijer, F. (2010) Fragile States: Stuck in a Capability Trap? Background paper to the World Development Report 2011. World Bank. In terms of practical implementation it draws from de Weijer, F. (forthcoming) Tackling Adaptive Problems in Afghanistan. CID Fellow Paper series, Harvard Kennedy School.

  • 20.

    See for instance Earl, S., Carden, F., and Smutylo, T. (2001) Outcome Mapping: Building Learning and Reflection into Development Programs. International Development Research Centre; Gamble, J. (2008) Developmental Evaluation: A Primer. The J.W. McConnell Family Foundation; and Roduner, D., and Schlappi, W. (2008) Logical Framework Approach and Outcome Mapping: A Constructive Attempt of Synthesis. Agridea.