So, the argument is as follows:

1. Evidence-based policy discourse that focuses on ‘what works, where and for whom’ presupposes the existence of evidence about ‘what is, what has been and what is likely to be’.
2. In some key policy domains – e.g. in relations to significant parts of the agricultural sector in large parts of sub-Saharan Africa (our STEPS work on the poultry sector in Ghana is a case in point) – the available evidence about ‘what is, what has been and what is likely to be’ is extremely weak (if not misleading or erroneous).
3. Without evidence about ‘what is, what has been and what is likely to be’, it is pretty much impossible to move to a meaningful consideration of ‘what works, where and for whom’.

And the question is: what is the alternative approach to policy where the lack of evidence on ‘what is, what has been and what is likely to be’ makes it impossible to move to a reasoned consideration of ‘what works, where and for whom’?

Let’s switch for a moment from policy making to navigation. For the mariner, navigation is essentially about: knowing where you are; avoiding danger and hazards; and arriving (eventually) where you want to go. Modern navigation – with GPS (that very precisely locates the ship), electronic chart plotters, tide tables, accurate weather forecasting, etc – has revolutionised the nature of marine transportation. If you know where you are and where you want to go, take all the relevant factors (tides, wind, weather, hazards etc) into consideration through the appropriate algorithm, and steer a true course, you will have a very high probability of (1) not going aground and (2) arriving where and when predicted.

In this way, modern navigation has some important commonalities with the ideal of evidence-based policy. Perhaps most importantly, it is the knowledge about what is, what has been and what is likely to be that both underpins and links precision navigation and evidence-based policy. It is of course true that neither modern navigation nor evidence-based policy making is immune to gales, freak events or changing political weather, which may well require a change or adaptation to course, speed and/or tactics.

But for centuries mariners navigated without the benefit of GPS, radar or accurate charts. Can we learn something from these early navigators that could inform approaches to policy making where the lack of evidence on ‘what is, what has been and what is likely to be’ makes it impossible to move to a systematic analysis of ‘what works, where and for whom’?

Older approaches to navigation were much more approximate. Instead of steering a course to the nearest degree, a general direction of travel toward a desired destination was maintained using stars or other celestial bodies to provide a rough-and-ready steering guide. ‘What is’ (e.g. knowing where you are) was established by observation and simple calculations of angles, speeds and distances. Evidence about ‘What has been’ drew on sailors’ knowledge and experience, while evidence about ’What will be’ drew on their ability to predict upcoming weather conditions, interactions with tides and so on. All of this information was partial so the critical element was maintaining the general direction of travel in the face of uncertainty and constantly changing conditions, as opposed to predicting precisely the time or location of landfall.

The policy maker who is faced with a lack of evidence on ‘what is, what has been and what is likely to be’ is in a position that is more akin to the traditional than the modern marine navigator. Perhaps the most important thing for policy makers in this position is to establish and maintain a general direction of travel. This could be specified broadly, and could easily accommodate fuzzy boundaries. For example, notions like ‘food security’ or ‘sustainable production’ might provide a ‘good enough’ specification of the general direction of travel.

In addition, important hazards and courses to be avoided could also be identified, and (provisional) success defined as NOT taking a number of specific actions rather than arriving at a precise destination. At the level of policy theory, this means identifying a number of necessary conditions for success. Even without the modern technology, celestial navigators knew that there were numerous ways they could reach the desired location (being successful): but none of these paths entailed running aground, losing their crew or going back to the point of departure. Running into bad weather was not necessarily the end of the world: as long as the crew were safe and the ship was not lost or damaged beyond repair, there was always the possibility of getting back on course.

If we follow the celestial navigation metaphor, the job of the policy maker is to assess, using some trusted ‘guiding stars’, various policy options and interventions as to whether or not they are likely to help maintain movement in the general direction of travel (in addition to whether they are affordable, politically acceptable etc etc). It would also be helpful to identify a set of undesirable actions (such as running aground) that must NOT be taken, because NOT taking them is necessary for success.

This seems like a very different proposition than that commonly portrayed as evidence-based policy making. Critically, the focus has moved up a level to the choice of the general direction of travel and the guiding star. The falsely reassuring idea that it is all about objective, technical (‘evidence-informed’) choice amongst competing policy approaches and instruments is no longer at centre stage.

(Images: Compass Study from calsidyrose on Flickr; HMS Beagle in the Galapagos by John Chancellor from nunataak on Flickr)