[Next] [Previous] [Up] [Top] [Contents]

5 Some future directions for economic modelling

5.1 Applying our model of modelling to ourselves

Finally let us apply the above model of modelling to ourselves as modellers with bounded rationality and inadequate information. Further let us suppose that our current economic models gave us an unacceptable level of error when applied to situations of greater environmental and hence behavioural complexity. What avenues might this suggest?

  1. Accept the level of error as inevitable.

  2. Trade in specificity for accuracy, i.e. accept vaguer models of the same type to decrease error.

  3. Trade simplicity for accuracy, i.e. expend more effort in search for more complex models of the same type.

  4. Change the modelling language, i.e. consider new types of models.

These are roughly ordered in terms of increasing optimism. (1) represents a position of extreme pessimism, and is probably only held by a few holists. (2) is a classical defensive posture for the less specific your models are the less refutable they are. (3) represents an optimistic approach - it assumes that the type of model is fine but just needs refining (e.g. by adding a "missing" variable). (4) represents the most radical response, appropriate when the existing modelling language can not even describe the behaviour to be modelled. It is perhaps this option that is usually what is indicated by authors when the "complexity" banner is unfurled.

From Complexity to Agent Modelling and Back Again - Bruce Edmonds - 15 MAY 97
[Next] [Previous] [Up] [Top] [Contents]

Generated with CERN WebMaker