The move into the sustainable development era in 2015 with the agreement of global goals has in tandem also seen growing academic interest and practitioner action research into the concept of flexible and adaptive programming.   The ODIs Adapting development: improving services to the poor clearly makes the case that ‘more of the same’ will not be sufficient to meet the SDG, that we must pursue innovative and politically smart ways to tackle the most intractable problems.

The education sector is a case in point. The stocktaking around progress on MDG’s 2 & 3  (gender equity  and completion of primary  schooling)  has repeatedly stressed the lesson that there was an over-emphasis on expansion towards these universal access targets: build more schools, deliver more units of teacher training, books and materials.  Tanzania was a prime example of this with seemingly good ‘linear’ progress in the 2000s, that more recently has been hollowed out as the quality of service plummeted and the low motivation/performance of teachers has significantly eroded progress.    It’s interesting to note the inverse relationship on Tanzania’s GDP growth and access to water provides another service delivery example on the infographic below.

Oxfam’s Duncan Green is one of the leading proponents of these adaptive approaches and frequently writes and posts on the Thinking and Working Politically and Doing Development Differently initiatives. He has noted the pathway within DFID that has been taken with a less rigid Smart Rules approach to programme cycle management that aim to empower staff and by extension contracted programme teams to flex and adapt the much larger initiatives financed with UKAID.

It therefore seemed an opportune moment at the mid-term review of a major DFID education programme EQUIP-Tanzania to reflect on how it has been able to practice adaptive programming.  As is increasingly the case with DFID investments, the programme is large, with around £50 million budget. Numerically it targets 2.3 million enrolled primary students and 47,446 in 4,439 schools across 47 districts: that’s over a quarter of the nation. It has operated since 2013 in support of better learning outcomes at basic education level. It seeks to develop, implement and demonstrate approaches to strengthen the quality of education in seven regions, ready for national scale-up.  The approved design therefore did inherently give scope for experimentation and adaptation within the paradigm of demonstrable and scalable approaches.

Following a broadly positive review bursting with enthusiasm from stakeholders as diverse as parents, community teaching assistants and central curriculum developers, I posed the question to the Management Agent team, so what has worked and to what extent have you been able to adapt, survive and prosper?

I attach the articulate response here which uses as a starting point the importance of a data driven and rigorous process to understand what the underlying problem is and what the beneficiaries  want (e.g. teachers, students) and how the rapid use of data and quick feedback loops can drive better adaptive delivery.  Such an approach is new to Tanzania within the education system and to some extent, the sub-national governance structures.  The programme came up with 5 key examples of adaptation from the first two years of implementation,  that I’ve tried to summarise below.

1. Move to Decentralised Financial Execution Local government officials requesting support was insufficient for real buy-in and risk project style dependency.  £23m of DFID finance was rerouted via the central government Treasury to reach the LGA, empowering and changing the dynamics.
2. Inclusion of Sub-National Strategic Leaders Within Tanzania’s political economy, the Regional Administrative Secretary (RAS) had been overlooked, the programme adapted to support the RAS and his team provide leadership for the activities, including co-locating regional offices.
3. From a Hierarchical Cascade to a Partnership Training Model Integrate collaborative training, for example of head teachers, local supervisors and inspectors to avoid bureaucratic blame games and mistrust
4. Focus on the data needs of Schools and Districts National school census data collection did little to support local decision making or comparative analysis of performance.  A School Information System is rolling out in 2016 and will be digitized with the use of 5,500 tablets in the hands of heads and local supervisors.
5. Adaptation of Teacher In-Service Training Mass professional training is inherently complex with a context of geographically scattered, unionised and demotivated workforce.   Through regular feedback and reflection, weaknesses of the modality in relation to clear training objectives were learnt and incremental modifications made to improve the programme’s delivery.  For example mixed delivery methods and the inclusion of additional teachers at zero cost in situ at schools.  This was seen to be much more effective than alternative approaches currently operational in Tanzania.

In the choice of examples presented I was struck by a focus on understanding the local government political economy, which is clearly highly context specific to Tanzania, but also the driver for a real understanding of public education in a huge, mostly rural environment where central decrees and guidelines have relatively little real authority.

It also emphasises that the real challenge in this instance of education reform was not of new pedagogic and/or technology interventions but in effect is about understanding and engaging education sector governance.

Like all true adaptors the programme has also managed to fail relatively fast and move on.    Providing expenses to teachers via mobile money worked in principle but was unpopular with recipients and struggled with a mismatch of formal id cards to SIMs.   A technically sophisticated national data collection system fell prey to vested interest and in-fighting, but is evolving to support local data needs through the tablet based school approach highlighted above.

These latter examples of technology and data systems being some of the hardest interventions to make work were driven home by a recent email flurry with my DFID peers around the globe.  We all identify teacher motivation and attendance monitoring as a key problem/solution and a World Bank example circulated on the use of Smartphones in Haiti that foundered on complexity and lack of compliance.   Within an hour similar examples came, (although I would stress with varying degrees of success), from colleagues in Rwanda, South Sudan and Pakistan.

Certainly the will to learn and adapt is present, but the laws of jungle still remind us that only the fittest survive!

Ian Attfield Senior Education Adviser, DFID Tanzania

You can follow Ian on twitter – @Ian Attfield