Craig Valters of ODI is consistently incisive on Theories of Change, cutting through the flannel surrounding one of the
I don’t care, I really like it
aid business’ favourite new(ish) fuzzwords to identify what is genuinely significant. His new, crisply written paper is a must read for anyone interested in how change happens, doing development differently, or the results agenda. Some excerpts:
‘The development industry is unbalanced in a number of ways. Approaches to accountability are narrow, time-consuming and unrealistic, and this works to displace genuine attempts to learn and adapt. Programmes are often developed in a top-down way rather than being a result of locally led endeavours. Various, rather static, evidence artefacts are produced, but they fail to stimulate learning that can lead to improved programming. Social change processes are often understood in a linear way, when we know things rarely unfold as planned.
Many donors and practitioners recognise these imbalances and try to do something about it. But the political incentives, deep bureaucratic cultures and power dynamics in the aid industry often sustain the status quo. These are longstanding problems that so far have successfully resisted change, despite years of critique. The Theory of Change approach advocated in this paper, in its own small way, seeks to shift the centre of gravity. The agenda put forward is ‘radically reformist’: sensitive to critiques of development thinking and practice but with the conviction that much can be done to make the endeavour more effective.’
Valters distils four key principles for a ‘Theory of Change Approach’:
‘Principle 1: Focus on process
Conventional programme management tools tend to ignore ‘process elements’, treating projects as ‘closed, controllable and unchanging systems’. Theories of Change can help challenge this – first by drawing attention to the oft-forgotten assumptions linking project activities and outcomes but second by encouraging a broader ‘learning process’ approach that is flexible and adaptive. One common problem with Theories of Change is that they are seen primarily as a product; a formal document to be completed at the start of a project and then to sit gathering dust on a shelf. Of course, writing Theories of Change down is important, but the process of uncovering and critically appraising assumptions will need to be ongoing precisely because, in the initial analysis, many assumptions are likely to be remain uncovered. Equally, as programmes unfold, more information will likely emerge to confirm or challenge assumptions in different contexts. The overall aim here is to avoid the production of static ‘evidence’ documents that fail to be integrated into programme strategies. A Theory of Change can then be used a way to record learning and adjusting; no documentation should be erased (as can happen with logframes).
Principle 2: Prioritise learning
While learning and accountability are not necessarily in tension, ‘official policies that profess the importance of learning are often contradicted by bureaucratic protocols and accounting systems which demand proof of results against pre-set targets’. For a reflective and adaptive approach to become mainstream in Theory of Change approaches – and indeed in development more broadly – understandings of what accountability and learning mean need to shift substantially. In many other industries, from business to football, managers are praised for adapting to changing circumstances; in development this is currently not the case. Having accountability for learning could be a promising route: there is no reason why, for example, programmes could not be held accountable for how much has been learnt over time, how they have adapted to new information and why this adaption has been important for improved development outcomes.
Principle 3: Be locally led
One of the dangers with a Theory of Change approach is that it remains a top-down process, imposed by a narrow group within organisations or programmes and/or excluding the input and views of beneficiaries. The guidance suggests a wide range of stakeholders be consulted but this often remains vague. On one level, if a small number of people develop the ‘theory’ itself, it is unlikely to represent broader organisational thinking on strategic, programmatic or intervention goals. This is particularly true if the views of implementing partners – those who are closest to the programme and with often better understanding of shifting local contexts – are excluded. On another level, there is now a well-acknowledged need to move beyond normative posturing around the need to gain the feedback of ‘beneficiaries’ on to explicit and systematic application of that feedback throughout monitoring, learning and evaluation processes. As such, the ‘beneficiaries’ of the programme need also to be consulted, at the start and throughout the Theory of Change process.
Principle 4: Think compass not map
For some leading Theory of Change advocates, the process involves developing a ‘roadmap to get you from here to 
there’. However, this way of thinking can recreate the fallacies in logframes, such as assumptions of linearity. Currently, many Theories of Change have been developed based on a moment of clear perspective in which ‘context’ is understood just enough to enact a grand design for a programme (often because the Theories of Change ignore the above principles of focusing on the process, prioritising learning and being locally led). Far more useful than a ‘roadmap’ is the idea of a ‘compass for helping us find our way through the fog of complex systems, discovering a path as we go along’. This is important since Theory of Change approaches must acknowledge that ‘social contexts and processes are always in flux, with emergent issues, unforeseen risks and surprises arising throughout’. This suggests a need for a considerable degree of modesty about we know about development processes.’
I’ve been worried for some time that Theories of Change will just degenerate into another box to tick – a logframe on steroids. But Craig’s paper shows how that can be avoided. If you can, read it all – it’s only 13 pages and it totally nails it. Or you can read his blogged summary.