Topic 2:Aggregation

While aggregation into a single index is always useful for comparative purpose, I was wondering if such comparison is at all useful in SD context. I say this because while aggregation would invariably be at the nation-state or country level, SS is much more complex in terms of intra-country and inter-country or inter-regional coupling effects. These effects probably do not add up or subtract in a linear fashion.

Taking the example of institutions, while absence of institutions is a problem, too many institutions also create institutional conflict and lethargy. This happens as one moves from local to global or global to local scales for the same event. This "nestedness" is a problem in aggregation effort and may result in totally inaccurate results / predictions.

Similarly, redistribution aspects do not lend itself to anlysis due to pressure to aggregate. This is also related to substitutability issues - not just between different capital stocks, but also between different geographic regions. For a forest cover lost in the Himalayas, can afforestation in the Andes or elsewhere be a perfect substitute in terms of regional and global impact?

We had an interesting conversation in the Cambridge group at the end of class yesterday about the relative advantages of a single aggregated indicator vs. the amoeba diagrams that were presented by the UNAM discussants. We discussed the fact that individuals are very uncomfortable with making decisions based on aggregated indicators, but that the amoeba diagrams may also problematic, in that there is no clear-cut way to compare them. This left me wondering if it might be feasible to combine the two approaches through an iterative, participatory process of indicator definition.

The simplest way to compare outcomes based on the amoeba diagrams would be to simply compare the area of the "web." However, this would basically be identical to aggregating the information into a single indicator score, with the relative distance between the dimensions filling in for the "weights" on a typical aggregated indicator. In this case, amoeba diagrams have only a slight advantage over an indicator in the perceived ease of interpretation by individual decision-makers. However, if people are uncomfortable making decisions based on an indicator score they would likely hesitate to agree to such a decision rule as well.

Perhaps a way to deal with this is to show representative groups of stakeholders and technical experts various outcomes (in the form of amoeba diagrams) and ask them to rank them in order of preference. This will help us identify any non-linearities in the preferences. For example, if amoeba diagrams with smaller areas are still generally preferred to larger webs in which one of the dimensions drops below a certain threshold, this would indicate a non-linearity that wouldn't be incorporated into a simple weighted average indicator. We could then use that information to improve the calculation of the indicator.

Another example (aside from reservation thresholds on certain dimensions of sustainability) could be an aversion to skewed amoeba diagrams. For example, if an amoeba with a smaller area is preferred to a larger web in which a single indicator is responsible for the difference, this would indicate that people have a preference for balanced development along all the dimensions. This could then be incorporated into the calculation of the indicator.

Perhaps by including decision-makers in this iterative process of indicator revision, we could simultaneously improve the ability of the indicator to capture complex dimensions of sustainability and increase the willingness of decision-makers to use the indicator in a meaningful way.