If using ‘Theories of Change’ cannot transform the way you operate, why bother?

In a new (and commendably short) paper, Craig Valters advocates ‘modest radicalism’ in the use of Theories of Change (ToC) as an approach to improving reflection and learning in the development sector. In this paper, Craig reflects on the role of the ToC in the context of the ‘results agenda’ and suggests four principles that could help development organisations develop knowledge and improve practice: Focus on processes; Prioritise learning; Be locally-led; and ‘Think compass, not map’. Do read the full paper!

In this post, I share some additional thoughts on the use of ToCs and how they might be improved. I start with two problems in the way we do things.

• In development, failures are hard to detect: Often, organisations that fail find ways to mask failure – by either refusing to acknowledge failure, finding external factors, or moving on to a different desk officer/donor/location. So within the aid industry, we have a peculiar situation where it is real hard to fail – or at least, it is hard to know when a project has failed.
• It’s harder still to ensure that projects that fail face significant consequences of failure: Organisations that implemented the failed projects should be required to make significant changes to key aspects of design or management.

Learning lessons and being accountable to stakeholders (including, but not just donors) is at the heart of both of these issues above. In the wide world outside the aid industry, projects and enterprises are set up and when they fail, those that are invested in these projects/ideas are forced to innovate, or reinvent. The margins of failure are finer and value-for-money considerations are built in to the system (as opposed to being introduced from the outside as tools for accountability and reporting). Now I certainly don’t mean that doing development is like selling cola – but it’s worth reflecting on some fundamental institutional attributes that can offer lessons. It’s a bit like the “‘why not cash?’ challenge” that conventional humanitarian aid is faced with – an uncomfortable question that signals that we can no longer tolerate ‘business as usual’.

So the challenge I have is not a modest one: with ToCs, if you cannot transform the way development organisations operate, why bother?

A ToC is supposed to set out clearly, assumptions behind how change happens (sadly, in my head, the term also summons complicated images with too many boxes, and arrows going in multiple directions). It is almost impossible to have a ‘simple’ ToC, because the world we operate in, is inevitably uncertain and complex. And as both a tool and an approach, as Craig points out, ToCs should assist in organisational learning and accountability.

But there are serious obstacles on the path towards this ideal. In his paper, Craig points to the ‘time-consuming and misleading results agenda’ and refers to Chambers’ ‘accountability v/s accountancy’ culture that’s sweeping through the donor world. These are pressing issues, and allow me point out a couple of additional ones.

An example is DFID’s Smart Rules, introduced last year, ostensibly in response to a call for greater flexibility and innovation from its country offices. It is meant to be a broad operating framework focused on programmes, intended to support DFID staff in improving programme delivery. Out of the 37 mandatory rules listed, I saw no more than 5 rules that refer directly to the design and implementation of programmes, and on managing the complexity of the environments in which they operate. The rest were (possibly justifiably) quite focused on what one may call the ‘administration’ of projects. So while this may have been ‘handing over the stick’ to DFID staff lower down, I am not sure it sets out a clear vision to reduce the upward accountability (often defined in a narrow ‘accounting’ sense) that the bureaucratic aid system perpetuates.

In addition, I also worry about the fate of new ideas in development. For instance, the four principles in Craig’s paper emphasises on attributes that are not too far removed from what David Korten once saw in a ‘learning organisation’. According to Korten, learning organisations are those that “(a) embrace error; (b) plan with the people; and (c) link knowledge building with action”. This is one reminder of how the learning agenda has been part of the conversation in various forms – and then have either remained on the periphery, or have been brutally mainstreamed. It is the latter that I find more interesting, and brings to mind most prominently, the fate of tools such as the ‘logframe’ and approaches such as ‘participation’. Once mainstreamed, these concepts are implemented in a manner that make them unrecognisable to those that initially developed them. But this is indeed a serious dilemma – the trade-off between a ‘pure, but narrowly adopted radical concept’, and a ‘widely-applied, but contaminated standard’. The same easily holds for how much of the monitoring and evaluation data is treated in average organisations – there is no sense that a follow-up is required, and when faced with external pressure (from a donor, for instance), a cursory effort – to satisfy – follows.

Theories of Change are therefore on tenuous ground. On one hand, they are increasingly being used alongside logframes, as a ‘qualitative’ complement; but on the other hand, they are also being seen as a static ‘project design phase requirement’ that pass through without rigorous scrutiny. This brings me back to my original question – if we cannot make ToCs truly transformative, why bother?

So in turn then, I have two very modest proposals on how we can make a start (don’t ask me to explain what my theory of change behind these suggestions for making ‘Theory of Change’ work is)

• Donors should invest more in creating and updating ToCs – formative work in project design is vital, so is space for review and reflection that have tangible outcomes. This means setting aside funds and time (sometimes, timelines are a bigger constraint than funds) for formative research and recognising the contribution of this phase towards the final outcomes of the project.
• Invest in training and building capacity on understanding and using ToCs. Two decades back, when logframes appeared on the scene, there were a million trainings on the “science” of LFAs/logical frameworks. In my graduate programme in India, logframes were part of the curriculum and we were tested on them. Notwithstanding the actual utility of logframes, the point I am making is that if ToCs are to be used more widely and intelligently, development sector personnel need training and skilling in their development and use.

As projects and programmes are designed and implemented, one should set out clearly, what success looks like. A Theory of (good) Change will do just that, along with spelling out the complex nature of the path involved. Being bold about what success looks like in a particular context will also force one to be explicit about what constitutes ‘failure’. With this, it should be possible to continuously review and reflect on progress, and transparently tell between success and failure. It follows then, that projects that fail should prompt a re-evaluation of strategies and options and lead to concrete changes. As Craig says in his paper and I agree, “This may not sound too radical to those outside the industry, but within it, this is an important and pressing need”.

***

If you are interested in the subject, do also read these posts by Heather, Duncan Green and Jennifer Ambrose