lmbett
commented
7 years ago The Gini is intuitive (now that we are used to it) and I use it too. But I think that the bridge between distributions at different scales or times is best approached via comparing distributions and this is the job of information theory.
Henry Thiel did the same for "Gini" by proposing an information theoretic index of inequality with aggregation properties similar to what I introduced. This is almost as popular as Gini because of its properties under decomposition or aggregation. The Gini is not easily relatable across scales.
But in the end they all just reflect the underlying distributions.
In your paper, it is clearly identified one salient aspect of the complex system, i.e. disparity. Not only the overall disparity is measure, the locally disparities are also measure.
But the question is, why do you think and where do you think this separation of average disparity and local disparities could add to literature? And why this approach is better than calculate the statistics both in general and locally using traditional measure?
Take income equality as an example, why do you think that your approach is better than calculate Gini index both at neighborhood level and large level such as city?