Fragility v Conflict – can you help with a new 2×2 please?

November 17, 2016

It’s International Men’s Day tomorrow – here’s why it’s a bad idea

November 17, 2016

Some highlights from the first 30 book launches for How Change Happens

November 17, 2016
empty image
empty image

I’m about six weeks into launching How Change Happens, and am having a great (if knackering) time. Highlights soauthor-sad-flirting far include a Kurdish/Dutch guitar combo warming up the crowd in Nijmegen, conversations with an Islamic finance entrepreneur trying to do financial inclusion in South Wales, a great group of women managing a community-run service station on the M5 motorway and a network of ‘social leaders’ in the North-East of England who were completely outside the aid and development bubble, but totally got what the book is about. Plus being asked ‘do you believe in the perfectability of the human species’ in Q&A at a predictably erudite launch at Blackwells bookshop in Oxford…..

Impressions so far? A lot of ‘practitioners’ are finding the book reassuring – it seems to back up their instinctive (and widespread) concerns about the direction the aid business has taken and how to put it right.

Some of the most thought provoking events have involved discussants – academics who read the book, then come back at you with probing criticisms and questions. It’s a bit gruelling having these kind of public vivas, but very productive. Three examples:

mushtaq-khanSOAS’ Mushtaq Khan argued that it’s all very well talking about positive deviance and looking for successful outliers, but what if there isn’t anything very positive going on? Aren’t these approaches essentially incremental, rather than transformational ? We also need to think about how aid or other interventions can catalyse big structural leaps.

That got me thinking about an idea that I toyed with introducing in the book, but then abandoned as just too geeky. Fitness landscapes. These are used in evolutionary biology to visualize the relationship between genetic make-up and successful reproduction/survival. The more successful you are, the higher you are up a peak, with the different peaks corresponding to different species – two peaks close together could be two species of Darwin’s species, two far apart could be a finch and an elephant. (Feel free to tell me that I’ve got this completely wrong, btw – I’m expecting it).

Mushtaq was arguing that in development, positive deviance looks at how to move higher up on a given peak, but

Fitness Landscape

Fitness Landscape

doesn’t capture the jumps between peaks achieved through structural change. I’m not sure that’s entirely true – positive deviance would also detect new peaks that have emerged spontaneously, for example a new business model or form of immunity to disease – but I take his point that overall, you are looking for positive change within the existing system, rather than major transformation of the system itself. In a follow up email exchange, he elaborated further: ‘The point about positive deviance is not just about missing distant peaks but also nearby peaks that haven’t sprouted yet. There must be a thought process for the potential leader to think through the chances of success and the implications for the poor of some new initiative. Positive deviance isn’t a substitute for that analysis. Positive deviance does however tell us about opportunities that shouldn’t be missed.’

The other point from Mushtaq that hit home was on power. The book argues that power is an all pervasive ‘developmental force field’ and that one key task for any activist is to render power visible in all its many guises (hidden, informal, visible etc). But it also takes what Mushtaq saw as a fairly crude line that redistributing power is the underlying purpose of development – in any given situation, an activist should seek to equalize that distribution by regulating/curbing the activities of those with too much power, and helping those with too little to organize and be heard. Given that I had just been slagging off conventional aid projects as linear approaches to complex systems, wasn’t that a very linear approach to a complex system, asked Mushtaq? Ouch.

Examples he cited to show that moving from less to more equal distribution of power is not always the best choice: what if empowering one group disempowers another? Or when rulers use nominally empowering measures such as decentralization to weaken their opponents in the capital? Or if a short term empowerment (the Arab Spring in Egypt or Syria, say) leads to a backlash and overall massively negative results for poor people? And given that the book also talks about the importance of leadership, what level of hierarchy (i.e. power imbalance) is optimal? Ouch again. Anyone care to chip in?

naila-kabeerLSE’s Naila Kabeer helped in a different way – she identified an underlying theme of the book, which I’ve been using ever since:  whenever you see what looks like a fixed, monolithic institution (state, political party, legal system, NGO, corporation, social norm), you should ask yourself how is it evolving, and what is its internal make-up? Then you can try to identify potential allies in bringing about change.

Another enjoyable discussion was in Nijmegen (after the guitarists). Paul Hoebink asked why there is so little discussion in the book on the legitimacy of different kinds of change agents taking certain actions. His view was that membership organizations such as trade unions are inherently more paul-hoebinklegitimate than non-membership organizations like Oxfam, but that neither is perfect. Legitimacy is an overall outcome both from accountability to different stakeholders (members in rich countries, and – for development NGOs – poor communities in developing countries) and taking part in a broad range of coalitions. I would like to read more on this – any suggestions for reading on the nature of legitimacy of non-state actors such as NGOs?


  1. Re ” Positive deviance isn’t a substitute for that analysis. Positive deviance does however tell us about opportunities that shouldn’t be missed.’ Good to see some attention to this. There are some interesting implications here for the application of predictive analytics (which I think the Positive Deviance website has shown an interest in- and which I am encouraging Evaluation people to explore. That is we need to find a predictive model that is good at predicting the _absence_ of an outcome we are interested in. Usually we are trying to predict its presence. Then focus on the False Positives i.e. those cases where the model attributes predicted absence of the outcome but in practice the outcome was present. These cases qualify, on first glance, as Positive Deviants. They are the cases where it would be well worthwhile doing a within-case investigation in order to find out how they managed to succeed against the odds. One example I have found is in a re-analysis of data on women representation in parliament in Africa, available in a paper by Krook. The absence of quotas for women in parliament is a sufficient condition for low levels of women’s participation in parliament. It predicts 13 of the 14 countries with such low levels. The one exception is Lesotho, where there are no quotas but there are high levels of participation of women in parliament. This is an example of a “positive deviance” case that would be worth investigating

    Krook, M.L., 2010. Women’s representation in parliament: A qualitative comparative analysis. Political Studies 58, 886–908.

  2. Regarding Paul Hoebink’s comment on legitimacy, you may want to look at. the book edited by Lisa Jordan and Peter van Tuijl : NGO Accountability, Politics, Principles and Innovations, Earthscan, 2006, especially the chapter by Enrique Peruzzotti : Civil Society, Representation and Accountability: Restating Current Debates on the Representativeness and Accountability of Civic Associations.

  3. The discussion of fitness landscapes and related ideas in Black Box Thinking (Syed) and Messy (Harford) suggests that perhaps it’s not too geeky an idea. Harford even discusses ‘simulated annealing’, an algorithm for finding the highest point in a search space without getting stuck on a local peak by ‘hill-climbing’ (

    Even if we had a clear picture of the entire fitness landscape, though (i.e. the performance of every available alternative), or knew the best approach to exploring it, the way the landscape changes is also critical, as you point out. And since our ability to predict such developments is fundamentally limited, in a complex world we can never be assured that we’re taking the best possible approach to solving a problem.

    There are clearly relatively better and worse methods for dealing with complexity, but when exploring them (as you have done so brilliantly on this blog and in How Change Happens) we mustn’t forget that, as Paul Ormerod puts it in Positive Linking: “The silver bullet of twenty-first-century networked reality is simple: it is to realise that there is no silver bullet.”

  4. Thanks for sharing some of the feedback from your book tour Duncan. Very helpful.

    I’m particularly interested in the comments about legitimacy, which seems to me to be a very close cousin of trust, and closely related to inclusive and effective governance, in the US, the UK and elsewhere. (As an aside: I know colleagues working with OGP are exploring the literature about whether openness e.g. open governance, fosters trust – answers on a postcard to the Paris Global Summit).

    With the World Development Report for 2017 – focused on governance (and politics) coming out soon – it might be a good time to revisit the excellent World Development Report for 2011 which was focused on institutions and legitimacy. And, as Karel mentions there was quite a bit of work done on NGOs’ and other actors’ legitimacy some years back. Some of it is referred to in here

  5. Thanks for these thoughts to supplement your excellent book Duncan.

    Re Fitness landscapes: my understanding (from more of a virology perspective) is that the different peaks represent different genotypes (of a species), rather than different species necessarily. There is an interesting literature about how organisms can explore the ‘sequence space’ available to them – thanks for taking me back to my undergrad in cell and systems biology… I like the idea of social innovations/technologies being shared as horizontal gene transfer within a population (say on bacterial plasmids). Something I’ll mull over..

Leave a comment

Translate »