Wednesday, May 8, 2019

Thoughts on system transformation

Some thoughts in the context of the WDR 2004 and systems transformation

In physics, the second law of thermodynamics states that the entropy of a system always increases. Any system perturbation sets in motion a dynamic whose transmission trajectory and consequence is unpredictable. This is so even for physical systems where particles are not self-acting.

Now consider the example of a human system where agents are clearly self-acting. A social system is a tenuously balanced one with countless and overlapping relationships involving human beings with a wide variety of preferences. This balance has emerged organically through repeat interactions. Further, this is no balance in the static sense, but a dynamic one. It is impossible to predict the outcome of any perturbation (a reform) to this system, beyond its immediate aftermath. Even when well-intentioned and planned, the emergent unintended consequences can be several and often very damaging.

The only response in such a situation is to get the minimum viable product (MVP) out and engage actively with the system, especially in the initial period, to respond quickly to emergent trends and iteratively improve the intervention's design accordingly.

This assumes relevance because it goes against the conventional wisdom on planning a program end-to-end. If we accept the aforesaid reality, then it becomes impossible to plan the details of execution beyond the MVP.

In fact, the only thing that can be planned is to be ready with enabling requirements at different levels to respond quickly and effectively to emergent situations. And given the importance of iterative adaptation, mechanisms to collect data and make available granular-enough actionable information (or feedback) relevant to each level of the bureaucratic system becomes central.

Here is an illustrative plan for a reform idea in terms of the four principal-agent relationships. One, engage continuously to keep ripening the conditions so that the change being sought can be a political good. Two, design an incentive compatible compact between the government and organisational providers (its own departments and non-government providers). Three, design an MVP implementation plan that is sufficiently flexible to allow for local adaptation and has a feedback monitoring system that enables iterative refinement of the plan. Four, create enabling conditions that make frontline service providers accountable to the end-users.

Apart from the second relationship, all others are likely to be dynamic, especially in the initial periods of a program implementation. The first and last would require ongoing political engagement and community mobilisation respectively, with the need for opportunistic engagement being critical in the former. The third would be the realm of iterative adaptation of the implementation plan.

Prudence dictates that the level of flexibility and iteration be calibrated to be consistent with the state's capacity to be able to respond effectively and engage as required.

Programs like the GST or IBC or Ayshman Bharat or Rythu Bandhu scheme or RERA in India are all good examples of reforms which could benefit from such an approach to implementation. 

No comments: