Introduction to minimum economic recovery standards 2nd edition
Richard v20121031
1. Simplifying complexity
through systems thinking
Panel: Measuring impact in market
systems (November 7th 2012)
Richard Hummelbrunner
ÖAR Regionalberatung
Graz, Austria
2. Systemic or Systematic?
Systemic
focus on the whole
and the parts
Three core dimensions:
Interrelationships
Perspectives
Boundaries
Systematic =
+
focus on the parts,
step-by-step
4. Consequences for monitoring
• Regard interventions as social systems
Unit of observation: intervention and context
Observe relevant contextual factors (scanning) during
implementation, in particular relevant actions of others
Look beyond intended routes and effects, avoid tunnel
view, capture broader range of effects (irrespective of
intentions)
• Different approach towards deviations from plan
Do not per se regard as negative (‘correction reflex’)
Do not treat as isolated phenomena, but connect with
intervention logic
Information to understand the internal dynamics and self-
organising forces at work within target social system
5. Linear or ‚circular‘ logic models ?
Inputs Outputs Results
Needs /
Problems Mechanisms Impact
Issues
Context
Complexity challenges our everyday thinking, which functions along ‘ linear ’ rules. Human cognition tends to think in simple cause-effect patterns, to ignore what cannot be directly observed or is not in line with expectations. Most frequent reaction: Ignore complexity - or reduce it mentally (simplify) in order to cope with it . Comparison of four types of simplification, their appropriateness for dealing with complexity - and alternative forms of simplification inspired by Systems Thinking.
Systematic: Simplifying by taking things apart, focus on details Appropriate for non-living (mechanic) systems, long tradition in natural sciences Systemic: Simplifying with a view of the whole, maintaining the big picture Appropriate for living (social) systems Many different schools / approaches in Systems Thinking, but three over-arching systems concepts: Useful for structuring the whole and its parts, Possibility to zoom in and out, reduce and increase complexity Systems can be perceived as 1. actual real world entities – e.g. ecosystems, financial systems, market systems ( thinking about systems ); and/or conceptual constructs for inquiry into real world entities e.g. models, maps, metaphors ( systems thinking )
Simplistic reduction through ‘mechanistic’ thinking or ‘M e ntal trivialisation’ (Heinz v. Foerster): Social systems regarded as trivial machines, linear input-output models where the same input will lead to identical output, regardless of circumstance. Strips social systems of their most important qualities (internal dynamics, self-organisation). Non-trivial simplification through systems thinking: Social systems and their context mutually influence each other, they adapt and co-evolve. In social systems achievement of effects does not follow a linear logic, they can react differently to the same input – and their behaviour can neither be explained by inputs, internal states or context alone, but results from their interaction. Successful work with/in social systems requires to take account of their specificities and the limits with respect to information, influence and control: Changes in social systems are essentially self-organised, based on internal mechanisms of regulation and stabilisation. Adaptation to external inputs follows internal logic and cannot be influenced externally in a direct, mechanistic sense. Differences from original states are inherent to assure the stability of social systems. Systems Thinking provides concepts and methods for dealing more appropriately with social systems.
If an intervention is regarded as social system, changes in short term targets or actions are often necessary for the achievement of long-term objectives. M&E conceived as negative feed-back loop (deviation needs correction in opposite direction) is inappropriate, should be in line with - and strengthen - internal regulatory mechanisms. Requires a different attitude towards differences, bearing in mind their regulatory function. If this is not taken into account, M&E risks counteracting internal mechanisms, which might result in misleading and even counter-productive conclusions: E.g. insisting on the implementation of original plans despite relevant changes in the operative context might have counter-intentional effects and ultimately result in failing to achieve objectives. Evaluations should progress from adaptive, single loop learning (‘d o ing things right’) to generative double loop learning, which opens the view to other alternatives (‘d o ing the right things’). But this requires influencing the internal state of social systems, notably their cognition and behaviour patterns.
In international development interventions are still predominantly based on li near models, assuming mechanistic progression of effects (irrespective of actors involved or contextual conditions). Can be expanded into a ‘c i rcular’ logic model, by adding two components that are interrelated with the elements of the original logic model, but also connected with each other: relevant operational context that can influence implementation (e.g. socioeconomic development, legal or administrative framework, interests of implementing partners and beneficiaries). intervention’s mechanisms (e.g. activities, criteria and conditions, decision maki ng ) that influence the relations between the elements. Interventions can be structured and linked to their context in a recursive logic: Changes of context due to impacts has potential effects on mechanisms, which in turn can affect the relations between elements... Consequences for M&E: Achievement of effects not seen isolated, but takes intervention characteristics and relevant context conditions into account Allows more refined/detailed answers: Effects are the consequence of specific mechanisms carried out under specific context conditions (‘wh at works for whom – and under which circumstance’?).
Dominant planning approaches in international development (e.g. log-frames, result chains) assume single intervention logic (sometimes synthesized from different stakeholder-views in a participatory process). During implementation different logics will be at play, which often are only implicit, cannot always be reconciled and might be contradictory. They result from different perspectives: Stakeholders or Stakes. Systems field provides methods and/or a language for conveying ideas between stakeholders, overcoming differences, improve mutual understanding, achieve consensus or create new insights or options Perspectives define framing of an intervention: relevance of elements, actors, inter-relations, boundaries. Systems approaches are not ‘h o listic but reductionist,facili tate a critical reflection of boundary choice (including related power issues) Changing perspectives or boundaries is helpful in identifying unplanned effects / phenomena. There are no unintended effects! Perspectives can emerge through analysis of exceptions, contradictions, surprises or puzzles in data. Can provide useful clues (e.g. relevant changes, new challenges, innovative or ‘i n formal’ ways of handling situations), which can help to improve implementation.