One of the first things we all learn as development rookies is that you cannot simply transplant institutions, systems or ideas from elsewhere. We are told that solutions have to be organic, locally-developed, country-owned and relevant to the context. But why and when is this true?
Part of the answer is suggested in the writings of Matt Andrews, Michael Woolcock, Lant Pritchett, Justin Sandefur, me, and others (see the reading list at the bottom of this blog post). For at least some problems, there is something useful about the ‘the struggle’ - that is, the need for a community to identify its challenges and grapple iteratively with the solutions. If the process of adaptation and iteration is necessary, then solutions parachuted in from outside will not succeed. Furthermore, efforts to bypass the struggle might actually be unhelpful.
Yet successful institutions in different countries often look similar to each other. For example, a good postal service looks pretty much the same everywhere. Good finance ministries resemble each other. So why should each country have to reinvent the wheel? Can they not bypass the costly and time-consuming process of struggling to create these institutions, and simply import good practice from beacons of success at home, or from good examples abroad, so taking success to scale?
This issue came under the spotlight at a recent meeting at CGD’s Washington Office to discuss the planned role of Global Development Innovation Ventures (GDIV), a joint venture of USAID, DFID and Omidyar Network which aims “to spark innovation and scale successes”. Is there something inconsistent between GDIV’s mandate to help countries take proven success to scale and the need – in at least some cases – for countries to grapple with their own challenges themselves?
Perhaps if we understand why the struggle might be important, we can describe better whether and how progress can be achieved more quickly and with less pain, or at least understand which kinds of development success could plausibly be taken to scale.
With the help of big thinkers Lant Pritchett and Michael Woolcock the CGD meeting was able to explore this. We discussed four reasons why the struggle might be important:
The struggle shapes solutions to the context. Solutions might look similar but actually include subtle, perhaps barely discernible, differences which adapt those solutions to their environment and without which they cannot function properly. (This is the ‘external validity’ problem). Perhaps these subtle differences which are vital to the success of the solution are brought about by the process of local problem-solving.
People take time to learn. If Roger Federer showed you how to serve a tennis ball that would not mean you could immediately serve like a professional. Just because you have an MBA doesn’t mean you can run a company. Other than for simple tasks, most of us have to learn by doing. Establishing habits may require repetition and practice, both for individuals and for organisations.
The struggle confers legitimacy. Michael Woolcock points out that a careful lawyer could have drafted the Good Friday Agreement (which brought peace to Northern Ireland) in a few hours: so why did there have to be so much bloodshed and anguish? Why were all-night negotiations needed to get an agreement? Perhaps the process of compromising – of give and take, of testing limits and building trust – is a pre-requisite for all parties to accept the compromise as the best available.
Systems co-evolve. Individual institutions do not operate in a vacuum. Each organisation is in a process of evolution, shaped by an external environment which includes other institutions which are themselves evolving. This process of co-evolution brings about the self-organising complexity typical of a complex adaptive system. Particular organisations cannot jump ahead of this if the environment they need to authorise and support them is not also evolving.
These four reasons why the struggle could be important raise an obvious question about the role of development cooperation. Typically aid aims in some way to diminish the struggle, or ideally to bypass it altogether. But if the struggle is necessary, at least some of the time, then we should think twice about whether and when it makes sense to try to minimize it.
For what kinds of problem are such struggles likely to be necessary? Lant Pritchett and Michael Woolcock suggest a spectrum of complexity and implementation-intensiveness. Simple, purely logistical interventions might perhaps be replicable. But more complex problems, such as those which depend on the emergence of legitimate systems and institutions, or which require continuing compliance and behaviour change, probably cannot be replicated without some sort of struggle.
There might be disagreement about whether a particular intervention can be replicated without a struggle. For example, distributing bednets appears to be a logistical challenge which, though complicated, can be solved by sharing best practice and good logistical management. Does that mean a successful model in one country can be rolled out elsewhere? If so, this will at minimum require some effort to build support, finance and legitimacy for the programme in each country (this may not be strictly speaking a struggle, but it may not be straightforward). Beyond that, a bednet programme will succeed if people understand why they might want to sleep under bednets, and adapt their behaviour and habits; if the power relations are such that men allow women and children to use the nets; if systems are put in place to distribute new bednets, perhaps through some combination of state provision and private markets; and if old bednets are regularly retreated with insecticide. Now we have moved from a logistical exercise to the realm of developing legitimate and effective institutions to provide continuing services, and the need for sustained changes in behaviour, power and trust. Can these changes be brought about without some kind of struggle?
The need for a struggle, at least sometimes, may have four implications for development cooperation, including for the GDIV programme.
First, the goal of ‘taking proven interventions to scale’ or ‘replicating success’, bypassing or minimising the struggle, may be appropriate to a relatively small set of interventions.
Second, it is possible that some of what donors do to try to accelerate development may instead slow it down by crowding out the necessary struggle. For example, aid could pay for basic services in the short term, while in the long run undermining the social contract that would emerge from the struggle for control of domestic revenues. Donor financing of civil society may lead to challenge to authority in the short run, but it may also crowd out a more legitimate dialogue rooted in local concerns. Donor support for businesses – for example creating jobs by backing firms – may crowd out the innovative, hungry firms on which the long term success of the economy depends.
Third, where it is not possible to replicate success directly, it may be possible to support systems to enable them evolve more rapidly and more surely towards the desired goals. For example, providing circumstances in which people can ‘fail safe’ may encourage more innovation. Better use of data and rigorous evaluation, and greater transparency and accountability, can encourage more effective selection. Donor funding which encourages and rewards local problem-solving, without imposing solutions from outside, may accelerate the struggle and make it less painful. This is part of the rationale for CGD’s proposals for Development Impact Bonds and Cash on Delivery Aid. The Problem Driven Iterative Adaptation approach suggested by Andrews, Pritchett and Woolcock is an effort to describe how aid can support countries’ own efforts to solve problems. Can some kind of technical assistance accelerate the struggle without replacing it? (Even professional tennis players have a coach.)
Fourth, aid can help people while they are struggling. This support may not directly accelerate development – perhaps other than by giving people more space to fail safe – but it could help them live more comfortably while development is taking place. (Of course, it follows from the above that it is important to provide this help in ways that do not crowd out the struggle.)
We would welcome views and comments on this. Is there a significant set of development policies which whose demonstrated success elsewhere suggests that they could be replicated and scaled up elsewhere with little or no struggle? And for the others, where the struggle cannot be bypassed, what are the smart ways that donors can support countries to make progress without crowding out? CGD is involved with a number of suggestions along these lines, including Cash on Delivery Aid, Development Impact Bonds, and Problem Driven Iterative Adaptation: are these good answers to this problem and how else might donors go about supporting countries engaged in the struggle?
Some further reading (compiled for the meeting by Molly Kinder):
Lant Pritchett & Justin Sandefur, "Context Matters: Why External Validity Claims and Development Practice Don't Mix" (+ blog post)
Tessa Bold, Mwangi Kimenyi, Germano Mwabu, Alice Ng’ang’a, and Justin Sandefur, "Scaling Up What Works: Experimental Evidence on External Validity in Kenyan Education" (+ blog post)
Michael Woolcock, "Using case studies to explore the external validity of 'complex' development interventions" (pdf)
Lant Pritchett, Jeffrey Hammer and Salimah Samji, "It's All About MeE: Using Structured Experimental Learning to Crawl the Design Space." (+ blog post)
Owen Barder, “The Implications of Complexity for Development” (narrated presentation) or "Complexity, Adaptation and Results" (blog post)
Lant Pritchett, “Impact Evaluation as a Learning Tool for Development Effectiveness” – presentation and video