When is a policy product a public good? (Or, don't let VFM crush your double loop dreams)

This is a question that we have recently posed ourselves through some work with one of our development partners.

The analyses we are doing are grounded in specific policy processes, typically at the national or sub national levels and are intended to nudge those processes towards more informed decisions about how to reduce poverty and promote human development, and therefore contribute to the realisation of those ultimate goals. The policy products we are producing are things like public policy briefs, in person briefings, analysis papers, and knowledge sharing roundtables and other knowledge sharing platforms.

Public goods are things that the market does not provide which are non-excludable (everyone can consume them) and non-rivalrous (consumption by one individual does not have an impact on the consumption by others) and have positive externalities (the more they are used, assuming they are of good quality).
Can context-specific policy products be public goods? Well, they can pass the technical definition of public goods—if they are not behind a paywall then they are non-excludable and they are non-rivalrous in that they can be accessed by unlimited numbers of readers without diminishing anyone’s individual consumption.

But can these policy products pass a more meaningful definition of public good? That is, a public good that is useful outside of its context. For example, take a piece of analysis that, say, suggests norm setting around violence against women and girls in country x is shaped, in large part, by the attitudes towards alcoholism and therefore programmes to treat and prevent alcoholism. This analysis is available and non rivalrous but, assuming it is of good quality, how useful is it in other contexts? The internal validity of the analysis and conclusions might be high (this is a credible conclusion in this context at this time), but what of the external validity (does this have any relevance for other contexts and times?). Perhaps in other cultures it is the design of urban spaces that emerges as a contributing factor to violence against women and girls, or poor protection in schools. 

So how do we make specific policy analysis outputs attain more features of a useful public good? One way is by surveying multiple experiences in several contexts, looking for commonalities, differences and patterns.  This is helped if the analyses surveyed pass a quality standard and are surveyed in a transparent and systematic way (and there are many transparent and systematic ways of doing this). Vital to this is a balanced interpretation of the evidence assembled to determine which factors are identified most frequently and which experiences carry more weight?

But most important, I think, is to determine whether the results we assemble cause us to fundamentally rethink our general assumptions about why violence against women and girls happens and what public policy can do about it. Does the assembled evidence, say, point to a complete rethink about violence against women? Is it more about norm-setting while the manifestations are context specific? If it is about norm-setting, that is the widely useful public good and the context specific knowledge about how to reset those norms and how to deal with their current fallout, while a public good, is the new locally valid policy product.  

For many in the illegal drugs control movement a similar ah-ha moment happened when they began to realise that different types of drug control was having as much of a negative effect on development as the drugs themselves. It was time to rethink the term “drug control”. Or when the reproductive health community began to realise that sex was always portrayed as a negative risk factor rather than something that could be a positive force for better reproductive health. Or when the HIV/AIDS community recognised how fundamental inequalities of all kinds were to the spread of HIV.

I suppose this is a kind of double loop learning. Single loop learning involves trying different approaches based on a given set of assumptions about the nature of the problem. Double loop learning questions the assumptions more and then uses that analysis to be more creative in generating potential solutions. Study after study shows that most organisations don’t have time for single, let alone double loop learning (never mind triple loop). 

Under the intense VFM pressures of today’s fiscal climate, all of us are being pressured into more and more single loop learning. We should resist. Policy and analysis will be the better for it--as will VFM.