As I reread Duncan Green’s blog about an interesting conversation we had over a coffee few weeks ago in Brixton (How might a systems approach change the way aid supports the knowledge sector in Indonesia?), it occurred to me that I could add something.
‘Systems’, he writes, ‘evolve through the endless churn of variation, selection and amplification. Variation = rate of mutations – new species, new companies; Selection = some are fit for the landscape, others are rubbish and die out; Amplification = the fit ones expand or proliferate.’
When it comes to development interventions, in order to contribute to a change or changes in a system, the various stakeholders have to ‘see’ where the web of variation, selection and amplification is weakest and focus on those weaknesses.
Meadows writes about the key components that make a system: elements, interconnections, and purpose.
- The elements of a system are easy to spot. They are visible and tangible. In a knowledge sector, for example, elements include universities, policy research institutes, policy analysis units, ministries, local governments, civil servants, researchers, data scientists, etc.
- Interconnections are the relationships that hold the elements together. To spot these interconnections requires some research and analysis to understand why elements are linked as they are. Interconnections are often the results of information flows and in a knowledge sector there can be a government department linking with think tanks to procure an analysis about the state of the economy or the quality of education and health services. The government can then use (or not use) this information to inform funding decisions.
- The purpose is the most difficult component of a system to spot. The purpose of the systems can be articulated in written documents such as regulations or bylaws, but it is better deduced from the behavior of organizations and individuals. For example, a written commitment to use more research-based evidence in policymaking may or may not be followed by concrete requests from ministries for research results and analysis to think tanks and policy research organizations.
To generate and/or contribute to changes in a system, it is important to understand how systems evolve (variation, selection and amplification) but also to have a clear map of the components of the systems trying to go beyond the mapping of the elements and changes to the elements of the system (e.g., new research organizations, trained civil servants, etc.) which usually, as Meadows reminds us, have the least impact on the system as a whole but are easier to measure and report on.