[Originally posted on the ESID blog.]

These days I am reading psychologist Daniel Kahneman‘s book Thinking, Fast and Slow (2012), in which he outlines two aspects of our brains which determine how we process information, associate ideas and solve problems. Kahneman speaks of two systems: System 1, which is quick, intuitive, and effortless; andSystem 2, which is slow, analytical, and costly.

The first chunk of the book is devoted to the interaction between these two systems, and in particular how System 1 is prone to bias by jumping to unwarranted conclusions on the basis of what’s familiar or sounds right, even without us consciously realising what we are doing; System 2 can then jump in to check our intuitions against facts and avoid logical mistakes, but doing so requires willpower and freedom from disruptive stimuli (we all have a limited budget for effort, be it mental, emotional or physical).

As I read the book, I started wondering whether the proponents of political analysis in aid agencies could learn something from the interaction between these two systems in our brains.

“Development” is good, “politics” is bad, “governance” is hard

I was particularly struck by the role of associative thinking, in which we tend to make quick subconscious links between ideas, physical reactions and emotions. For instance, Kahneman refers to experiments in which participants tended to walk more slowly when primed to think about words associated with old age, even when they reported not having thought about age at all. Or an experiment in which people primed to think about money were likely to sit farther from others, be less helpful to those in need, and behave in a more individualistic manner. Could this kind of associative bias account for the almost visceral resistance to the idea of politics entering development?

We like to argue that the distinction between “political” and “technical” in development is a fallacy. But is that the real barrier to politically smart aid? In fact I would say that most aid practitioners consider what they are doing “development”, whereas we suggest they do “politics”.

Development is a nice word, and it conjures positive associations, whereas politics elicits at best mixed reactions. Go into a thesaurus and look for synonyms to “development”: you will find words like “growth”, “advancement”, “evolution”, “expansion”, “improvement”, “progress”. These are all positive ideas that sound good and right. Now look for the term “politics”: you get words like “affairs of state” and “government”, but also “power struggle”, “opportunism” or “backroom”. These are either ideas that themselves require unpacking (what is and what isn’t an affair of state), or that have outright negative connotations. The term “development” is easily processed by our intuitive System 1, whereas  “politics” has either bad connotations for System 1 or else it requires the costly analytical engagement of System 2.

Consider other popular terms that “politics” is clashing with in the aid business: “results” (“consequence”, “effect”), “value-for-money” (“efficiency”, “benefit”), and the double whammy that is “science of delivery” (“knowledge”, “skill”; “distribution”, “rescue”). No wonder the “Doing Development Differently” movement is doing great, while the “Thinking and Working Politically” community struggles: “doing” sounds much better than “thinking”, let alone “working”; and “development” is a much more positive term than “politics”.

What about using “political economy” or “governance” instead? Both terms may avoid the associative biases of System 1 thinking, as they seem neutral enough when compared to “politics”. So neutral, in fact, that I doubt anyone can intuitively engage with them at all: these are technical terms belonging to very limited fields of expertise, and even those familiar with them hardly ever agree on a definition. “Governance” in particular is an incredibly demanding term for us, requiring System 2 to operate at full throttle: whose governance, governance for what, governance for whom, and so on.

The politics of System 1

The big irony behind all of this is that in fact System 2 thinking, while costly, is much less prone to errors than System 1: saying “it’s complicated” may in fact be the first step in the path to better interpretations of reality and, through that, to better aid programming. But not everybody has the time or cognitive budget to engage in System 2 thinking on a daily basis. In fact, the professional pressures that aid practitioners are subject to (“disburse, disburse, disburse”, and its twin sibling, “comply, comply, comply”) may well be too taxing for them to approach politics with any kind of cognitive ease.

So it is really no wonder that practitioners default to well established associations, especially when mere repetition imbues ideas with an intuitive form of truth. It does not really matter that conventional aid is often ineffective, as long as practitioners remain exposed to a world of “logical frameworks”, “theories of change”, “results chains” and “value-for-money” (as opposed to “illogical messes”, “beliefs of stasis”, “disconnected random events” and “wasteful spending”). Sheer cognitive attrition can make these ideas ring true and, moreover, feel right.

What can advocates of political analysis make of this? Well, for starters they may need to drop the word “politics” when trying to persuade outsiders. “Context”, “dynamics” and “relationships” are all potential alternatives that conjure much more positive associations. “Analysis” too may be a tricky word, as it invokes “study” and “examination” (“boring”), and some in DFID have tried the word “diagnostic”, which seems equally demanding in a cognitive sense. Perhaps something along the lines of “engagement” could do the trick: who would want to feel “disengaged” or “out of touch”? Out with the “problem-driven political-economy analysis”, in with the “development engagement toolkit”!

A second implication for advocates may be how to design and present frameworks and models for political analysis in such a way as to facilitate System 1 thinking instead of asking people to fire up their lazy System 2. This probably requires “human readable” terms instead of the jargon that we are so fond of, but also developing analytical steps that mirror our intuitive thought processes instead of the instruction manual for a household appliance. In other words, let political analysis experts use their System 2 to produce reliable models, but then translate them into toolkits that a non-expert’s System 1 can easily grasp and process.

This may be just my particular System 1 falling for a cognitive trap, but I feel – notice the wording – that there is much that we can learn from Daniel Kahneman and other psychologists working on intuitive vs rational thinking. And the beauty of it is that if I repeat this idea enough times, you too will probably start to feel that it is true.