This blog was originally posted on the Working with the grain: Integrating governance and growth website and has been cross-posted with the permission of the blog author, Brian Levy (@Brianlevy387).

New measurements confirm that governance systems matter for learning outcomes. But knowing, in a statistically robust way that systems matter is one thing. Understanding how they matter, and what are implications for action, is another.

A useful blog by Marla Spivack of the Center for Global Development and the global RISE (Research on Improving Systems of Education) programme reviews the new econometric measures–which were used by Gabrielle Wills, Debra Shepherd and Janeli Kotze to assess the impact of governance systems on learning outcomes  in  chapter 6 of the recent book, The Politics and Governance of Basic Education: A tale of two South African provinces (Oxford University Press, 2018). A broader goal of the 2018 book (with which I was centrally involved) was to anchor specific empirical findings within a comprehensive,  multilevel (national, provincial and school) analysis of how politics, institutions and governance interact.

The multi-level research points to three ways to move from general recognition of the role of governance towards better education sector practice. First (and of particular relevance to the work of many economists and educationalists in the sector):

• Move beyond a narrowly technocratic pre-occupation with education ‘production functions’—with exploring in a narrowly economistic way  how specific inputs influence learning outcomes.  

Here’s how Spivack describes the limitations of the ‘production function’ approach: “There are numerous component parts of an education system that can either promote or impinge on student outcomes…. RISE calls these the ‘proximate determinants’ of education outcomes. Vast academic and policy literatures exist examining the proximate determinants of learning…. Questions like ‘what is the effect of teacher training on learning?’ what is the effect of missing textbooks on learning?’ and ‘what is the effect of a new pedagogical approach on learning?’ all follow this formula…… Their effects on children’s outcomes differ across contexts. RISE is interested in understanding the features of systems that mediates these varied effects”. To put it differently, it is not so much the ‘independent’ magnitude of the effects on learning of  proximate determinants, but how they interact with governance systems that is key for understanding their effects on learning outcomes.

This leads to a second proposition:

• Governance influences learning outcomes via three distinct channels: the technical efficiency with which inputs are deployed; whether inputs are used for their intended purpose; and the evocation of agency—the commitment and motivation of those involved in the education endeavor. 

Having embraced governance as crucial, those with a technocratic bias may be inclined to fall into the technocratic trap of focusing only on the first and second channels –  with an implied presumption that better top-down, process-compliant  hierarchical systems are sufficient to improve educational outcomes. But,  as per the above, governance functions are multi-dimensional – and include the ‘agency’ channel (on which more below).  As the book explored in depth (see HERE), different locales vary in the strengths and weaknesses of each of these governance dimensions. These divergent patterns  explain why,  as Spivack’s blog summarizes, “the Western Cape effect is not always positive – the WC’s education system is stronger than Botswana, Gauteng and the Eastern Cape, and weaker than Mauritius and Nairobi and the Central Region of Kenya”.

Indeed, as chapter 10 of the 2018 book explored (and as I summarized HERE), a crucial difference between the Western Cape and Kenya appears to be in the evocation of agency—the commitment of parents, teachers, communities, public officials to do what it takes to improve learning outcomes. Within South Africa itself, school-level case studies of governance dynamics and learning outcomes in both the Western and Eastern Cape provinces (chapters 8 and 9 of the book)  further revealed, at a micro-level, the powerful role played by agency on the part of school leaders, teachers, parents and communities in accounting for variations in school-level learning outcomes. To those with a technocratic bent, ‘agency’ might seem to be an especially ‘soft’ causal driver–but don’t underestimate its power. (Its evocation is, for example, central to the power of the PDIA–problem-driven iterative adaptation–approach to change, of which RISE’s Lant Pritchett, plus Matt Andrews and Michael Woolcock are leading champions.)

My final proposition concerns an additional hazard of using comparative research findings  for policy prescription. The extended comparison between the Western and Eastern Cape provinces in chapters 4, 5, 6 and 7 of the book reveals stark differences between the performance of the Western and Eastern Cape bureaucracies. Given this contrast,  a seemingly straightforward implication is to focus on “fixing the bureaucracy” as a way to improve learning outcomes.  This would be a mistaken use of the research findings:

• Context matters—politically and institutionally-driven incentives and constraints shape what governance entry points for improving learning outcomes are feasible in specific settings. 

A central feature of the 2018 book was a systematic juxtaposition of  hierarchical and horizontal approaches to education sector governance. The intent was less to explore whether one is ‘better’ than the other, but to enrich the menu of options for improving learning outcomes.  As chapter 7 of the book detailed (and as I summarized HERE) variations in the performance of the Western and Eastern Cape education bureaucracies are rooted in profound differences in the socio-economic, political and institutional context of the two provinces. Given these structural constraints it would be misguided to focus narrowly on “fixing the bureaucracy” as a way to improve learning outcomes in the Eastern Cape province. A broader range of options (including an investment in strengthening learning-oriented parental and community participation) might usefully also be considered as part of the ‘governance improvement’ mix.

So (speaking as much to myself  and my current enthusiasm for the ‘evocation of agency’ as to others reading this piece), let’s avoid the lure of easy answers. As per RISE’s mission, taking governance seriously  is an important step forward in efforts to improve learning outcomes. But there is much to be learned about how governance matters—with bold, persistent, learning-oriented experimentation at least as important as a further round of studies.